pediatric population reference value distributions for ... · pediatric population reference value...

Pediatric Population Reference Value Distributions for Cancer Biomarkers: A CALIPER Study of Healthy

Community Children

by

Victoria Bevilacqua

A thesis submitted in conformity with the requirements for the degree of Masters of Science

Laboratory Medicine and Pathobiology University of Toronto

© Copyright by Victoria Bevilacqua 2014

ii

Pediatric Population Reference Value Distributions for Cancer

Biomarkers: A CALIPER Study of Healthy Community Children

Victoria Bevilacqua

Masters of Science

Laboratory Medicine and Pathobiology University of Toronto

2014

Abstract

As part of CALIPER program, a national research initiative aimed at closing the gaps in pediatric

reference intervals, I sought to develop a database of covariate-stratified reference intervals in

children for 11 circulating tumor markers in accordance with CLSI C28-A3 guidelines. Healthy

children from birth to 18 years were recruited to participate in CALIPER and serum samples

from 400-700 subjects were analyzed on the Abbott Architect ci4100™. Significant fluctuations

in biomarker concentrations by age and/or gender were observed in 10 of 11 biomarkers. Age

partitioning was required for CA 15-3, CA 125, CA 19-9, CEA, SCC, ProGRP, Total & Free

PSA, HE4 and AFP, and gender partitioning was required for CA 125, CA 19-9, Total & Free

PSA. The establishment of these reference intervals will aid in harnessing the full potential of

tumor markers in a pediatric population and in research aimed at determining the clinical value

of these markers.

iii

Acknowledgments

I would like to express my gratitude to my supervisor Dr. Khosrow Adeli for his years of

guidance and support throughout this experience. I have learned far more than I ever imagined I

would when I initially began work on the CALIPER project and I will carry with me those skills

and lessons for the rest of my career. I would also like to acknowledge Dr. Cole and Dr. Yousef

– thank you for your advice throughout this process.

There are many hard working and talented CALIPER team members who have made incredible

contributions to the project over the years especially Jennifer Clarke, Dr. David Colantonio,

Ashley Cohen, Caitlin Wilkinson, Yunqi Chen, Megan Smith, Olivia Virag and Man Khun Chan.

Without your hard work and support this study would not be possible. I would also like to give

special thanks to Dr. Dana Bailey for her mentorship – I learned so much from you in the few

short months we worked together and the lessons were invaluable. I also have to acknowledge

Sarah Delaney for both her contributions to CALIPER and her unwavering support and

friendship throughout this process. In addition, I am very grateful to the laboratory technicians

who were instrumental in completing the analytical portion of this study especially Tony

Khommanivong who worked many weekends to help complete this project.

Additionally, I would like to acknowledge our funding partners the Canadian Institutes of Health

& Research (CIHR) as well as Abbott Diagnostics.

I would like to thank my family and friends for their constant support and encouragement. I

would like to especially thank my Mom - I am certain that I would not have made it through

without you.

Most importantly, I would like to express my deepest gratitude and respect for the thousands of

brave children and teens who recognized the importance of helping their peers struggling with

illness. This study would not have been possible without your contribution.

iv

Contributions

Sample collection for this project was completed by the CALIPER Project’s Hamilton and

Toronto outreach teams made up of myself, Olivia Virag, Sarah Delaney, Megan Smith and a

large group of volunteers from McMaster Children’s Hospital and the Hospital for Sick Children.

Sample selection and location was completed by myself and sample testing was completed by

certified technicians, Tony Khommanivong and Barbara Hawkins, at the Hospital for Sick

Children. Data entry of the sample testing results was completed by myself

Finally, statistical analysis was completed by myself with help from Yunqi Chen using an R

program created by Ashley Cohen and Yunqi Chen.

Finally, the principal investigator, Dr. Khosrow Adeli, supervised all stages of the study.

v

Table of Contents

ACKNOWLEDGMENTS. III!

CONTRIBUTIONS. IV!

TABLE.OF.CONTENTS. V!

LIST.OF.TABLES. IX!

LIST.OF.FIGURES. X!

LIST.OF.ABBREVIATIONS. XI!

1! INTRODUCTION. 1!1.1! REFERENCE.INTERVALS. 1!1.2! THE.CALIPER.PROJECT. 2!1.3! CANCER.BIOMARKERS. 3!1.4! ALPHA>FETOPROTEIN.(AFP). 4!1.4.1! BIOCHEMICAL!PROPERTIES! 4!1.4.2! CLINICAL!APPLICATIONS!IN!CANCER! 5!1.4.3! CLINICAL!APPLICATIONS!IN!OTHER!CONDITIONS! 5!1.4.4! AVAILABLE!REFERENCE!INTERVALS! 6!1.5! ANTI>THYROGLOBULIN.(ANTI>TG). 6!1.5.1! BIOCHEMICAL!PROPERTIES! 6!1.5.2! CLINICAL!APPLICATIONS!IN!CANCER! 6!1.5.3! CLINICAL!APPLICATIONS!IN!OTHER!CONDITIONS! 7!1.5.4! AVAILABLE!REFERENCE!INTERVALS! 7!1.6! CANCER.ANTIGEN.15>3.(CA.15>3). 7!1.6.1! BIOCHEMICAL!PROPERTIES! 7!1.6.2! CLINICAL!APPLICATIONS!IN!CANCER! 8!1.6.3! CLINICAL!APPLICATIONS!IN!OTHER!CONDITIONS! 8!1.6.4! AVAILABLE!REFERENCE!INTERVALS! 9!1.7! CANCER.ANTIGEN.19>9.(CA.19>9). 9!

vi

1.7.1! BIOCHEMICAL!PROPERTIES! 9!1.7.2! CLINICAL!APPLICATIONS!IN!CANCER! 9!1.7.3! CLINICAL!APPLICATIONS!IN!OTHER!CONDITIONS! 10!1.7.4! AVAILABLE!REFERENCE!INTERVALS! 10!1.8! CANCER.ANTIGEN.125.(CA.125). 10!1.8.1! BIOCHEMICAL!PROPERTIES! 10!1.8.2! CLINICAL!APPLICATIONS!IN!CANCER! 11!1.8.3! CLINICAL!APPLICATIONS!IN!OTHER!CONDITIONS! 12!1.8.4! AVAILABLE!REFERENCE!INTERVALS! 12!1.9! CARCINOEMBRYONIC.ANTIGEN.(CEA). 12!1.9.1! BIOCHEMICAL!PROPERTIES! 12!1.9.2! CLINICAL!APPLICATIONS!IN!CANCER! 13!1.9.3! CLINICAL!APPLICATIONS!IN!OTHER!CONDITIONS! 13!1.9.4! AVAILABLE!REFERENCE!INTERVALS! 14!1.10! HUMAN.EMBRYONIC.PROTEIN.4.(HE4). 14!1.10.1! BIOCHEMICAL!PROPERTIES! 14!1.10.2! CLINICAL!APPLICATIONS!IN!CANCER! 14!1.10.3! CLINICAL!APPLICATIONS!IN!OTHER!CONDITIONS! 15!1.10.4! AVAILABLE!REFERENCE!INTERVALS! 15!1.11! PRO>GASTRIN>RELEASING.PEPTIDE.(PROGRP). 16!1.11.1! BIOCHEMICAL!PROPERTIES! 16!1.11.2! CLINICAL!APPLICATIONS!IN!CANCER! 17!1.11.3! CLINICAL!APPLICATIONS!IN!OTHER!CONDITIONS! 17!1.11.4! AVAILABLE!REFERENCE!INTERVALS! 17!1.12! FREE.AND.TOTAL.PROSTATE.SPECIFIC.ANTIGEN.(PSA). 17!1.12.1! BIOCHEMICAL!PROPERTIES! 17!1.12.2! CLINICAL!APPLICATIONS!IN!CANCER! 18!1.12.3! CLINICAL!APPLICATIONS!IN!OTHER!CONDITIONS! 19!1.12.4! AVAILABLE!REFERENCE!INTERVALS! 20!1.13! SQUAMOUS.CELL.CARCINOMA.ANTIGEN.(SCC). 20!1.13.1! BIOCHEMICAL!PROPERTIES! 20!1.13.2! CLINICAL!APPLICATIONS!IN!CANCER! 21!1.13.3! CLINICAL!APPLICATIONS!IN!OTHER!CONDITIONS! 22!1.13.4! AVAILABLE!REFERENCE!INTERVALS! 22!

vii

1.14! SENSITIVITY.AND.SPECIFICITY.OF.CANCER.BIOMARKER.ASSAYS. 22!1.15! RATIONALE.FOR.DEVELOPMENT.OF.CANCER.BIOMARKER.REFERENCE.INTERVALS. 23!1.16! RATIONALE.FOR.DEVELOPMENT.OF.CANCER.BIOMARKER.REFERENCE.INTERVALS.IN.A.PEDIATRIC.

POPULATION. 24!1.17! HYPOTHESIS.&.OBJECTIVES. 25!

2! MATERIALS.AND.METHODS. 29!

2.1! PARTICIPANT.RECRUITMENT.AND.SAMPLE.ACQUISITION. 29!2.1.1! PARTICIPANT!RECRUITMENT! 29!2.1.2! SAMPLE!ACQUISITION! 29!2.1.3! SAMPLE!PROCESSING! 30!2.2! PARTICIPANT.SELECTION. 30!2.3! SAMPLE.ANALYSIS. 31!2.3.1! INSTRUMENT!AND!ASSAYS! 31!2.3.2! PRINCIPLES!OF!PROCEDURE! 31!2.3.3! CALIBRATION!AND!QUALITY!CONTROL! 34!2.4! STATISTICAL.ANALYSIS.AND.DETERMINATION.OF.REFERENCE.INTERVALS. 36!

3! RESULTS. 41!3.1! OVERALL.FINDINGS. 41!3.2! TYPICAL.ONCOFETAL.ANTIGENS. 41!3.3! ATYPICAL.ONCOFETAL.ANTIGENS. 41!3.4! SEX>SPECIFIC.PARTITIONING. 42!

4! DISCUSSION. 56!4.1! KEY.OBSERVATIONS. 56!4.2! TYPICAL.ONCOFETAL.ANTIGENS. 56!4.3! ATYPICAL.ONCOFETAL.ANTIGENS. 58!4.4! SEX.DIFFERENCES. 59!4.5! COMPARISON.TO.PREVIOUSLY.ESTABLISHED.PEDIATRIC.REFERENCE.INTERVALS. 61!4.5.1! KEY!OBSERVATIONS! 61!4.5.2! TYPICAL!ONCOFETAL!ANTIGENS! 61!4.5.3! ATYPICAL!ONCOFETAL!ANTIGENS! 62!4.5.4! ANALYTES!REQUIRING!SEXDSPECIFIC!PARTITIONS! 62!

viii

4.5.5! ANTIDTHYROGLOBULIN! 63!4.6! COMPARISON.TO.PREVIOUSLY.ESTABLISHED.ADULT.REFERENCE.INTERVALS. 63!4.7! STUDY.LIMITATIONS. 66!4.7.1! SINGLE!PLATFORM!STUDY! 66!4.7.2! STORAGE!EFFECTS! 66!4.7.3! CONFIDENCE!INTERVAL!SIZE! 67!4.7.4! NEONATAL!SAMPLE!SIZE!RESTRAINTS! 67!4.8! FUTURE.DIRECTIONS. 67!4.8.1! ESTABLISHING!REFERENCE!INTERVALS!FOR!ADDITIONAL!CANCER!BIOMARKERS! 67!4.8.2! EXAMINING!ETHNIC!DIFFERENCES! 68!4.8.3! TRANSFERRING!REFERENCE!INTERVALS!TO!ADDITIONAL!PLATFORMS! 70!4.8.4! ESTABLISHING!BIOLOGICAL!VARIATION!PARAMETERS! 70!4.9! CONCLUSIONS. 71!

REFERENCES. 72!

List of Tables

Table 1. Clinical applications and available pediatric reference intervals for 11 key

circulating tumor biomarkers ………….…………………………….......27

Table 2. Analytical Specificity and Sensitivity of 11 Cancer Marker Assays on the

Abbott Architect ci4100 ……………………………………...………….28

Table 3. Number of Samples Tested For Each Analyte…………………………...32

Table 4. List of Antibodies Used in Abbott Architect ci4100 Assays ………........33

Table 5. Analytical Performance of Immunoassays on the ARCHITECT

ci4100…………………………………………………………………….35

Table 6. Age- and Sex-Stratified Reference Intervals For 11 Serum Immunoassay

Cancer Biomarker Analytes Analyzed on Abbott Architect ci4100…......43

Table 7. Adult Reference Intervals and Decision Limits………………………….65

x

List of Figures

Figure 1. CALIPER study data-analysis protocol based on CLSI guidelines

document C28-A3……………………………………………………......40

Figure 2. Age-dependent scatter plots by sex of Anti-Thyroglobulin (Anti-Tg). …44

Figure 3. Age-dependent scatter plots by sex of Alpha-Fetoprotein (AFP). ………45

Figure 4. Age-dependent scatter plots by sex of Cancer Antigen 19-9 (CA19-9)....46

Figure 5. Age-dependent scatter plots by sex of Carcinoembryonic Antigen (CEA).

……………………………………………………………………………47

Figure 6. Age-dependent scatter plots by sex of Pro-Gastrin-Releasing Peptide

(ProGRP). ………………………………………………………………..48

Figure 7. Age-dependent scatter plots by sex of Human Embryonic Protein 4

(HE4)...................... ………………………………………………….......49

Figure 8. Age-dependent scatter plots by sex of Squamous Cell Carcinoma Antigen

(SCC). ………………………………………………...………………....50

Figure 9. Age-dependent scatter plots by sex of Cancer Antigen 15-3 (CA15-3).

………………………………………………………………………… ...51

Figure 10. Age-dependent scatter plots by sex of Cancer Antigen 125 (CA125).

…………………………………………………………………………....52

Figure 11. Age-dependent scatter plots by sex of Cancer Antigen 125 (CA125) and

Cancer Antigen 15-3 (CA15-3) during the neonatal period. ....………....53

Figure 12. Age-dependent scatter plots by sex of Free Prostate Specific Antigen

(PSA).. ………………………………...……………...………………....54

Figure 13. Age-dependent scatter plots by sex of Total Prostate Specific Antigen

(PSA). …………………………………...………………….…………....55

List of Abbreviations

Alpha-fetoprotein AFP

Anti-thyroglobulin Anti-Tg

Basic fibroblast growth factor bFGF

Canadian Laboratory Initiative in Pediatric Reference Intervals CALIPER

Canadian Institutes of Health Research CIHR

Cancer Antigen 15-3 CA 15-3

Cancer Antigen 19-9 CA 19-9

Cancer Antigen 125 CA 125

Carcinoembryonic antigen CEA

Clinical Laboratory Standards Institute CLSI

Computerized Tomography CT

Coefficient of variation CV

Differentiated Thyroid Cancer DTC

External Quality Assessment EQA

Hepatoblastomas HB

Hepatocellular carcinomas HCC

Human choriogonadotropin hCG

Human Embryonic Protein 4 HE4

Human kallikrein-related peptidase 3 hK3

xii

Homovanillic acid HVA

IGF binding protein IGFBP

Insulin-like growth factor IGF

Magnetic Resonance MR

National Academy of Clinical Biochemistry NACB

Natural killer NK

Parathyroid hormone PTH

Pro-Gastrin Releasing Peptide ProGRP

Prostate specific antigen PSA

Radioimmunoassay RIA

Reference change values RCV

Risk of malignancy algorithm ROMA

Secretory leucocyte proteinase inhibitor SLPI

Serine proteinase inhibitor Serpin

Serum separator tubes SSTTM

Squamous cell carcinoma antigen SCC

Src family kinases SFKs

Thyroglobulin Tg

Transforming growth factor- β TGF-β

Tumor necrosis factor α TNFα

xiii

Vascular endothelial growth factor VEGF

Vanillylmandelic acid VMA

Whey-Acidic Protein WAP

1

Chapter 1 Introduction

1 Introduction

1.1 Reference Intervals

In order to properly interpret laboratory test results, it is crucial that appropriately established

reference intervals be available. Reference ranges encompass the central 95% of the distribution

of test results in a healthy population (1) and, although the need for these comparison values is

well understood, the establishment of these ranges is often a challenging and involved process.

As such, it has been difficult for the field to keep pace with novel biomarkers emerging

constantly and technological advances progressing at increasingly rapid rates.

Currently, the Clinical Laboratory Standards Institute (CLSI) recommends recruitment of at least

120 healthy reference individuals in order to achieve an acceptable level of statistical confidence

to establish the 2.5 and 97.5th percentiles in addition to two sided 90% confidence intervals (2).

This is a challenging, time consuming and expensive undertaking in any population, indeed, the

establishment of reference intervals in a pediatric population is particularly difficult. In addition

to the common challenges faced when establishing reference intervals in adult populations, the

pediatric community presents a set of unique complications. Specifically, recruitment of large

numbers of healthy children across the entire pediatric age range is a not an easy task. Informed

consent must be obtained from both the parent or guardian of the participant and assent must be

obtained from the participant. Therefore, researchers are tasked with the challenge of educating

both adults and children on the importance of this type of research in a way that is clear and

accessible for both groups. In addition to the challenges involved in the informed consent and

recruitment process, levels of many biomarkers fluctuate with the continuous physiological

changes that occur throughout childhood. Development and growth also profoundly influence

reference intervals for many of the disease biomarkers measured in the laboratory and, as such,

the need for age and gender partitioning becomes all the more crucial and results in the need for

additional sample collection (3-5). It is, therefore, not surprising that many of the pediatric

reference interval studies that have been conducted to date have failed to meet the statistical

standards set out by CLSI (6). Additionally, those studies that do meet statistical standards often

2

use hospital patients instead of sampling from a healthy population; this is a less than ideal

approach discussed further in Section 4.5.

1.2 The CALIPER Project

The CALIPER (Canadian Laboratory Initiative in Pediatric Reference Intervals) program is a

national research initiative established by a Canadian team of investigators, aimed at closing the

gaps that currently exist in pediatric reference intervals by establishing a comprehensive database

of reference values in Canadian children stratified by age, gender and ethnicity. To date, the

CALIPER project has published reference intervals for over 60 assays. In addition, through

funding from the Canadian Institutes of Health Research (CIHR), the CALIPER project has

recruited more than 8600 study participants from birth to 18 years of age. These samples are

currently stored at -80 °C in the CALIPER biobank located at The Hospital for Sick Children,

Toronto, Canada.

The first major CALIPER study, published in 2012, established age- and sex-specific reference

intervals for 40 common serum biochemical markers, protein biomarkers, lipids and enzymes

(4). This initial study was carried out using Abbott Diagnostics clinical chemistry assays,

however, subsequently, a transference study was carried out in order to establish reference

intervals on four other major analytical platforms (Beckman, Ortho, Roche & Siemens) (7). In

2013, two additional CALIPER studies were completed, which determined age- and sex-specific

reference intervals for 14 endocrine and biochemical markers as well as 7 fertility hormones (3,

5). The CALIPER project has continued to be productive in 2014. This year, to date, CALIPER

has also published age- and sex-specific pediatric reference intervals for Vitamin A and Vitamin

E, as well as within-day biological variation parameters for 38 biochemical markers (8, 9). In

addition, reference interval data for a number of genetic and metabolic markers, as well as

additional chemistry and immunoassays on the Abbott Architect ci4100 and Beckman AU

systems have been presented at the American Association for Clinical Chemistry and Canadian

Society for Clinical Chemists 2014 Annual Meetings, and manuscripts are currently in

preparation.

3

1.3 Cancer Biomarkers

Despite the great strides that the CALIPER project has taken to close the gaps in the field of

pediatric reference intervals, there are a number of areas where information is still lacking.

Notably, there are considerable gaps in pediatric reference intervals for cancer biomarkers.

Pediatric cancer diagnoses account for approximately 1% of new cancer cases in North America

(10), however, this number ought not mask the impact of this disease on the pediatric population.

Cancer is the most common cause of death from disease in children and teens and is the second

overall leading cause of death in children under the age of 15 (11).

Recently, cancer biomarkers have become an important tool in the fight against this disease.

Table 1 provides a summary of the clinical applications for 11 key cancer biomarkers in both

pediatric and adult cancers. In addition, the biochemical properties, clinical applications,

sensitivity and specificity of each of the 11 markers examined in this study are discussed in

Sections 1.5 – 1.14. Despite the great potential for these markers to aid in cancer treatment

efforts, information on normal levels of these markers in the pediatric population is sorely

lacking. Table 1 provides a summary of the pediatric reference intervals currently available for

these 11 key cancer biomarkers. The gaps are evident with little to no information available for

7 of the 11 markers of interest.

A biomarker is defined as “a characteristic that is objectively measured and evaluated as an

indicator of normal biological processes, pathological processes or pharmacological responses to

a therapeutic intervention” (12). This concept implies that an observation or a measurement can

be used as an indicator that a certain biological process or disease is present (13). Ideally, a

biomarker test should be relatively easy to perform, minimally invasive and inexpensive (14).

Clinically, biomarkers have a number of potential applications. Firstly, biomarkers can be used

as a diagnostic tool to help identify patients with specific conditions and/or diseases. Biomarkers

can also be used to help in the assessment of disease prognosis. Finally, biomarkers can play an

important role in both determining which treatment or intervention to use, and in monitoring a

patient’s response to a specific intervention (12).

A large body of ongoing research is actively examining the ways in which various biomarkers

can be used clinically in different cancers and populations. It is clear that there are limitations to

how and when markers can be used. Specifically, research efforts aimed at identifying

4

biomarkers for early detection and screening have been largely unsuccessful. It has been

difficult to identify markers where the benefits of screening outweigh the risk and cost. In order

to justify the use of a biomarker for screening purposes, it is important that the benefit to patient

survival rates outweigh the risk of overdiagnosis and unnecessary treatment costs (15). For

example, although prostate specific antigen (PSA) is currently being used in many countries for

prostate cancer screening, there remains a great deal of controversy regarding whether or not this

is actually a useful practice [Section 1.12].

In addition to the uncertainty associated with using cancer markers for screening, the potential

usefulness of these markers in patient diagnosis also appears to be limited (13). In order for a

marker to be useful in early diagnosis it must meet a number of criteria that are often difficult to

fulfill. For example, the molecule must be released into circulation by small asymptomatic

tumors at levels that are substantially higher than the levels found in normal, healthy patients

(13).

Despite the limitations related to how and when cancer biomarkers can be used in patient

diagnosis and screening efforts, there remains great potential for these cancer biomarkers to aid

in monitoring and follow-up assessments of cancer patients as described in the following

sections [Sections 1.4 – 1.13].

1.4 Alpha-Fetoprotein (AFP)

1.4.1 Biochemical Properties

AFP is a glycoprotein with a molecular weight of 61,000 to 75,000 Da. This glycoprotein is

produced in the developing fetus by the yolk sac and the fetal liver until 13 weeks post-

conception when the yolk sac degenerates and the fetal liver becomes the primary site of

synthesis. At birth, AFP synthesis ceases and levels fall to adult concentrations (< 10 ng/mL)

(16).

The biological role of AFP has not been well characterized, although, due to its biochemical

similarity to albumin, it has been hypothesized that AFP may be a carrier protein (16). AFP may

act as a carrier for estrogen and bilirubin, it may play a role in stimulating growth, and, like

many other oncofetal antigens, it may also play a role in immunosuppression (17-19).

5

1.4.2 Clinical Applications in Cancer

In general, there are very few guidelines for the use of cancer biomarkers in the pediatric

population and very few large-scale studies have examined the diagnostic, prognostic or

monitorial applications of these markers. However, some studies have begun to explore

applications for these markers in pediatric populations and, in many cases, the preliminary

evidence is promising.

In terms of the potential function of AFP in the context of pediatric neoplasms, there are several

relevant applications. Specifically, there is evidence for the utility of AFP measurement in

patients with germ cell tumors, hepatoblastomas (HB) and hepatocellular carcinomas (HCC).

Hepatic tumors are not particularly common in the pediatric population, and account for

approximately 5% of total neoplasms in the fetal and neonatal populations (20). However,

children who are affected with biliary atresia, infantile cholestasis, glycogen-storage diseases,

and a wide array of cirrhotic diseases of the liver are predisposed to developing HCC. In

addition, neuroblastomas, leukemia, renal tumors, yolk sac tumors, rhabdomyosarcomas and

rhabdoid tumors have all been shown to metastasize to the liver (20). HB is the most common

type of liver tumor in children, comprising approximately 50 – 60 % of all liver neoplasms.

Abnormally low as well as abnormally high levels (≤ 100 ng/mL or ≥ 1 000 000 ng/mL) of AFP

have been associated with poor prognosis in pediatric HB patients (21).

In addition, serum AFP can be used as a diagnostic aid in yolk sac tumors (22), and as an

independent prognostic factor for pediatric patients with germ cell tumors (23). Notably, a long-

term follow-up study identified CA 125, along with AFP, as potential tumor markers that can be

used to monitor patients with sacrococcygeal teratomas, which have a tendency to recur (24).

Finally, AFP is also elevated in patients with HCC, however, not to the same extent as the

elevations observed in HB (25).

1.4.3 Clinical Applications in Other Conditions

Increased levels of AFP are associated with a number of other pathological and

physiological conditions. Namely, high levels of AFP are found in a number of liver diseases

including neonatal hepatitis, biliary atresia, Indian childhood cirrhosis and tyrosinemia (26, 27).

Physiologically, AFP is also elevated during pregnancy and is used as a screening tool for a

6

number of different conditions like neural tube defects, abdominal wall defects, nephrosis, ataxia

telangiectasia and Down Syndrome (16).

1.4.4 Available Reference Intervals

Age-stratified reference intervals are especially crucial for AFP as its levels fluctuate throughout

early development. One study established reference intervals for AFP from the neonatal period

to 2 years of age (22). In addition, CALIPER conducted a full study establishing age- and sex-

specific reference intervals for AFP from birth to 18 years of age for an earlier generation Abbott

AFP immunoassay (3).

1.5 Anti-Thyroglobulin (Anti-Tg)


Anti-thyroid antibodies are autoantibodies that target the thyroid and are involved in thyroid

tissue destruction. As its name suggests, anti-thyroglobulin (Anti-Tg) specifically targets the

thyroglobulin (Tg) molecule. Tg is a protein produced in the thyroid follicle and an important

component of thyroid hormone synthesis (28).

Like many of the other markers examined here, the majority of the research available on Anti-Tg

relates to its clinical uses (discussed below). There is limited knowledge on the biological action

of these autoantibodies aside from their known thyroid damaging potential (29). It is thought that

Anti-Tg binds Tg with high affinity and cleaves the glycoprotein at several different peptide

bonds (29).


The use of Tg in the monitoring of thyroid diseases, including thyroid cancer, is controversial for

many reasons. Although there is evidence that Tg measurement can help in monitoring of

patients with thyroid cancer post-surgery, there are a variety of analytical issues surrounding

measurement of both Tg as well as Anti-Tg. The first challenge is that Anti-Tg can interfere with

radio and immunoassays attempting to measure levels of Tg (30). However, if Anti-Tg is used

instead of Tg, there are issues with poor concordance between different assays (31). This second

issue in particular, highlights the need to establish reference values specific to the platform used

in order to appropriately monitor patients with cancers and other thyroid diseases.

7

Despite the challenges associated with Tg and Anti-Tg measurements, there is evidence that

Anti-Tg may serve as a useful marker for patients with Differentiated Thyroid Cancer (DTC).

Specifically, it appears that persistently high levels, de novo appearance, or a rising trend in Tg

antibodies in the post-operative period, are all risk factors for persistent or recurrent disease (32).


Anti-Tg is also known to be a hallmark of autoimmune thyroid disease and can be useful in

diagnosing conditions like Hashimoto’s thyroiditis, postpartum thyroiditis and Graves’ Disease.

In addition, it has been reported that anti-Tg levels are elevated in children with transient

congenital hypothyroidism (33). Finally, thyroid autoimmunity is also very common in women

of reproductive age and is associated with negative outcomes in pregnancy (34).


The 5th and 95th percentiles for children aged birth to 20 years have been published recently. The

results were arbitrarily divided into 8 different age groups. It is also important to note that

approximately 87% of the participants used in this study were hospital patients (35).

1.6 Cancer Antigen 15-3 (CA 15-3)


CA 15-3 is a member of the Mucin family of glycoproteins (36) and has been shown to play an

important role in cancer progression. It has a high molecular weight and is present on the apical

surface of many secretory epithelial cells, and is also highly expressed in various types of

carcinomas (36). There is evidence that CA 15-3 is not merely a marker of tumor growth but

may also play a role in tumour progression. Specifically, one study demonstrated that CA 15-3

knockout mice have significantly slower tumour growth than wild type mice (36). Based on the

findings of this study by Spicer et al., it is not clear how CA 15-3 promotes tumour growth,

however, there are many hypotheses regarding this question that have been explored by other

groups.

Similar to CA 125, it is hypothesized that the large extra-cellular domain of CA 15-3 may play a

role in regulating cell adhesion, in turn, contributing to metastasis in conditions of aberrant

expression (36). Specifically, CA 15-3, which is normally expressed on the apical side of the

8

cell, can be found on the basolateral cell surface of cancerous cells. This loss of polarity may

disrupt normal cell-cell and cell-substratum interactions and contribute to metastasis (36).

There is also evidence that CA 15-3 interacts with the actin cytoskeleton, which may indicate a

function related to cell motility. In addition, similar to many oncofetal antigens, there is

evidence that CA 15-3 protects cancer cells from both recognition as well as destruction by the

host immune system (37).


CA 15-3 is most commonly used as a marker of breast cancer (38, 39). Specifically, this marker

has been used for surveillance of patients following diagnosis with breast cancer. However, there

is some controversy over the usefulness of CA 15-3 for detecting recurrences in patients in

remission, and different guidelines disagree on whether or not CA 15-3 should be used in this

context (40, 41). However, most guidelines recommend the use of CA 15-3 to monitor patients

with advanced disease in conjunction with other monitoring, like clinical examination and

imaging (40).

To date, guidelines for the use of tumor markers have only particularly recommended the use of

CA 15-3 in already diagnosed breast cancer patients as outlined above. However, more recently,

some studies have found correlations between elevated levels of CA 15-3 and other types of

cancers including: pancreatic, uterine, gastric and lung carcinomas (42). Further research on

these correlations will be required in order to establish whether or not CA 15-3 can be used

clinically in the diagnosis, prognosis or monitoring of patients with these types of cancers.


As noted above, there are limitations to the use of CA 15-3 in patients with early or localized

breast cancer. This is because often times CA 15-3 levels in these patients are similar to either

healthy women or women with benign breast disease (42). In addition to this, other benign

diseases such as chronic hepatitis, liver cirrhosis, sarcoidosis, hypothyroidism and megablastic

anemia may also result in increased levels of CA 15-3 in the circulation (43-46).

9


To my knowledge, no studies examining the levels of CA 15-3 in a healthy pediatric population

exist.

1.7 Cancer Antigen 19-9 (CA 19-9)


CA 19-9 is a glycolipid with biological functions that remain largely uncharacterized. It is

important to note that approximately 5% of people have a Lewis-null blood type and are,

therefore, unable to produce CA 19-9 (47). This information must be taken into account when

assessing a patient.

Despite the limited knowledge on the biological role of this glycolipid in both normal physiology

and pathophysiology, some information is available. CA 19-9 is primarily synthesized by normal

human pancreatic and biliary ductular cells. In addition, CA 19-9 can also be synthesized by

gastric, colonic, endometrial and salivary epithelia (48, 49). Like many of the other tumor

markers examined, it is hypothesized that CA 19-9 may play a role in cell adhesion. This is

based on the function of CA 19-9 as a ligand for the adhesion molecule E-selectin (50).


CA 19-9 is used primarily as a marker for pancreatic cancer (51), however, additional

applications for CA 19-9 are being explored. For example, CA 19-9 can be used together with

carcinoembryonic antigen (CEA) for both diagnosis and follow-up of colorectal and

gastrointestinal cancers. In addition, CA 19-9 has been detected in biliary and hepatocellular

cancers (22, 52). Like many of the other markers examined, the utility of CA 19-9 for screening

and diagnosis of pancreatic cancer is limited. Instead, it is recommended that CA 19-9 be used

to monitor patients who have already been diagnosed with pancreatic cancer.

In terms of the application of CA 19-9 to the pediatric population specifically, there is a paucity

of information available. However, one study examined levels of CA 19-9 in pediatric patients

with germ cell tumors and discovered that, in some cases, levels are elevated in patients with

immature teratomas and malignant germ cell tumors. This was in contrast to patients with

mature teratomas where CA 19-9 levels were consistently normal (24).

10


In addition to the CA 19-9 elevations observed in the various types of cancers described above,

there are also a number of other benign conditions that are associated with high levels of CA 19-

9. Specifically, pancreatitis, biliary disease and cirrhosis have all been correlated with elevated

levels of CA 19-9 (47).


Reference values have been established for CA 19-9, but only from birth to 1.5 years of age (53).

1.8 Cancer Antigen 125 (CA 125)


CA 125 is a member of the Mucin family of glycoproteins (36) and has been shown to play an

important role in cancer progression. CA 125 has been found in seminal fluid, the fallopian tube

and the endometrium (54). In comparison to many of the other markers examined here that have

been primarily studied in a clinical context, the biological function of CA 125 has been studied

in more detail. Like other mucins, CA 125 spans the cell membrane and contains a large

extracellular domain, an intra-membrane domain and a short cytoplasmic tail (55).

Importantly, this cytoplasmic tail contains a putative phosphorylation site that is important to its

role in cancer progression (54). Specifically, it was recently demonstrated that the cytoplasmic

tail is phosphorylated by Src family kinases (SFKs) which are known to be negative regulators of

E-cadherin mediated cell to cell adhesion (55). It is also known that interaction of CA 125 with

β-catenin inhibits the ability of β-catenin to bind to E-cadherin, and this disruption has been

associated with increased metastatic potential (55). Therefore, it is possible that the cytoplasmic

domain of CA 125 may disrupt cell adhesion via its interaction with β-catenin, which is mediated

by SFK phosphorylation of the CA 125 cytoplasmic domain.

In addition, it is hypothesized that the large extra-cellular domain of CA 125 may play a role in

cell adhesion and can contribute to metastasis in conditions of aberrant expression (36).

Specifically, it has been demonstrated that endometrial cells can interact with mesothelial cells

via binding of CA 125 to mesothelin (56). Therefore, there are a variety of mechanisms at play

11

which may alter cell-cell interactions and, in turn, contribute to metastasis and cancer

progression.

CA 125 has also been implicated in increased cell proliferation and cell motility. Knockdown of

CA 125 induced cell density-dependent growth arrest and inhibited tumor growth both in in vitro

and in in vivo models. In addition, overexpression of CA 125 decreased cell-density-dependent

growth arrest and increased tumor growth in vivo and in vitro (57). CA 125 overexpressing cells

also demonstrated increased cell motility and invasiveness in vitro (57).

Finally, CA 125 has also been implicated in chemoresistance and immunoresistance (54). The

immunoresistant function of CA 125 is not surprising given its oncofetal nature. Specifically, CA

125 has been shown to be capable of preventing synapses between natural killer (NK) cells and

ovarian cancer cells, which limits the ability of the host’s immune system to kill cancerous cells

(58).


CA 125 is used primarily in cancers affecting women; notably, breast and ovarian cancers. CA

125 is one of the few markers that is recommend for use in early detection and screening efforts.

Specifically, the National Academy of Clinical Biochemistry (NACB) recommends that CA 125

testing be used in addition to transvaginal ultrasound to screen high risk patients (40). However,

CA 125 is more commonly used in combination with human embryonic protein 4 (HE4) to aid in

the diagnosis of patients with pelvic masses, in the monitoring of therapy in ovarian cancer

patients undergoing treatment, and in the monitoring of ovarian cancer patients in remission for

potential recurrences. CA 125 can also be used to help determine the prognosis of ovarian

cancer patients (40). CA 125 is one of the few markers that may be useful in screening,

diagnosis, prognostic determination and monitoring.

Despite the wide use of CA 125 in breast and ovarian cancers, new roles for this marker are

being assessed as other forms of cancer have also been correlated with elevated levels of CA

125. Of particular importance to the pediatric population is a long-term follow-up study that

identified CA 125, along with AFP, as a potential tumor marker to be used in the monitoring of

patients with sacrococcygeal teratomas, which have a tendency to recur (24). In addition,

12

elevated levels of CA 125 have been observed in children with leukemia and lymphoma when

compared to a healthy pediatric population (59, 60).


There are a number of other conditions, both physiological and pathological, that are also

associated with elevations in CA 125. Specifically in relation to physiological conditions,

elevations in CA 125 levels have been related to pregnancy as well as specific phases of the

menstrual cycle (61-63). With regards to the relationship between CA 125 and benign

pathological conditions, elevations in CA 125 have been related to the presence of fibroids,

ovarian cysts, pelvic inflammation, cirrhosis, ascites, pleural and pericardial effusions and

endometriosis (64-66).


Reference values have been established for CA 125, but only from birth to 1.5 years of age (53).

1.9 Carcinoembryonic Antigen (CEA)


CEA is a complex, highly glycosylated glycoprotein belonging to the immunoglobulin

superfamily and it is highly expressed during fetal development. In adult tissues, CEA is more

limited in its expression but is present in epithelial and goblet cells in the colon, mucous neck

cells, pyloric mucous cells in the stomach, squamous epithelial cells of the tongue, esophagus

and cervix, secretory epithelial cells in sweat glands and epithelial cells of the prostate (67).

CEA is typically located on the apical side of cells in the colon, however, malignant colon cancer

cells may have no basal lamina and no polarity and CEA is, therefore, distributed around the cell

surface. It is believed that this contributes to the increased levels of CEA found in serum of

cancer patients (67).

CEA is thought to play a role in cell adhesion and can act as both a homophilic or a heterophilic

cell adhesion molecule (67). CEA/CD44 knockdown cell lines showed a reduced capacity to

tether to purified L- and E-selectin when compared to wild-type controls (68). It is thought that

the role of CEA in disrupting cell adhesion can contribute to cancer progression. The ability of

13

colorectal cell lines to grow in nude mouse spleen and liver models correlates positively with the

levels of CEA production (69). In addition, like many of the other tumor markers examined here,

CEA seems to contribute to cancer cell resistance to the action of NK cell mediated cell-lysis

(70, 71).


As a tumor marker, CEA is used post-surgery to monitor individuals diagnosed with colon

cancer (67). Primarily, CEA is used to monitor patients who are in remission for potential

recurrences of the cancer. Due to the fact that poorly differentiated and early stage tumors tend

to produce less CEA, it is not useful as an early screening tool for colorectal cancer or for the

diagnosis of patients suspected to have the disease (72).

In addition to the primary role of CEA as a monitoring tool for patients with colon cancer, a

number of other malignancies are also associated with elevated levels of CEA including: breast,

lung, gastric, pancreatic, bladder, medullary thyroid, head and neck, cervical and hepatic cancers,

as well as lymphoma and melanoma (73, 74).

In terms of the role of CEA in pediatric cancers, the research is fairly limited. A study examining

the usefulness of CEA in colorectal carcinomas occurring in the pediatric population revealed

that CEA does not appear to be effective at detecting recurrent or progressive disease, which

contrasts to findings in the adult population (75). However, it should be noted that these studies

did not have access to information on reference values for CEA in the pediatric population and,

therefore, used adult cut-offs in their studies. It is possible, that with access to rigorously

established age- and sex-specific pediatric reference intervals, the outcome of this study would

have been different.


It is important for clinicians to consider the fact that a number of other conditions and patient

characteristics are also associated with elevated CEA levels. Perhaps most importantly, CEA

levels appear to be higher in individuals who smoke cigarettes. In addition, peptic ulcer disease,

inflammatory bowel disease, pancreatitis, hypothyroidism, biliary obstruction and cirrhosis are

also known to be associated with higher levels of CEA (74).

14


To my knowledge, no studies examining levels of CEA in a healthy pediatric population exist.

1.10 Human Embryonic Protein 4 (HE4)


HE4 is a member of the Whey-Acidic Protein (WAP) family (76) and has been shown to play an

important role in cancer progression. HE4 was originally thought to be expressed exclusively in

the epididymis, however, recently it has been shown that HE4 is expressed in a variety of other

normal human tissues including the respiratory tract and nasopharynx, in addition to various

types of tumors (77).

Although the cellular role of HE4 has not been well characterized, WAP proteins are typically

small secretory proteins that are often involved in cell growth and differentiation (77). Two of

the most well-characterized WAP proteins, secretory leucocyte proteinase inhibitor (SLPI) and

elafin, are known to have anti-proteinase activity, which is important in immune defense where

proteolytic enzymes are released by inflammatory cells during disease (78).

In relation to HE4 specifically, it was initially suggested that the main function of this protein

was as an anti-proteinase in the process of sperm maturation (79), however, no studies have

provided direct evidence for this. There is, however, early evidence that HE4 may play a role in

carcinogenesis and cancer progression. Notably, a recent study demonstrated that HE4

overexpression enhanced malignant phenotypes both in cell culture and in a mouse model (76).

Specifically, HE4 overexpression in cell lines resulted in increased cell proliferation,

invasiveness and anchorage independent growth. Additionally, in mice xenograft models, HE4

overexpressing cells grew larger tumors than other cell lines (76).


HE4 is a recently discovered marker used primarily in the context of ovarian cancer. Although it

is still being heavily researched, there is evidence that this marker may be a useful complement

to CA 125 in the management of ovarian cancer. Despite the utility of CA 125 in monitoring

ovarian cancer patients, there are a number of unmet clinical expectations regarding its use,

which stem specifically from the variety of benign conditions associated with elevations in CA

15

125 and from the inability to use CA 125 to detect early stage disease. Although HE4 and CA

125 have similar sensitivity and specificity, HE4 is not elevated in benign gynecological

conditions as often as CA 125 (80). Additionally, it was demonstrated that HE4 is better for

detecting early-stage tumors than CA 125 (81). There is also a significant body of literature that

has investigated the ways in which combined use of CA 125 and HE4 can be helpful in

diagnosis, prognosis and monitoring of ovarian cancer. There is some evidence that combined

measurements of CA 125 and HE4 can be useful in the diagnosis of women presenting with

pelvic masses (82, 83). As such, an algorithm known as the risk of malignancy algorithm

(ROMA) was developed to combine patients’ CA 125 and HE4 measurements along with

menopausal status to aid with distinguishing malignant from benign pelvic masses (84). In

addition to its role in diagnosis, it is also recommended that HE4 be used to monitor ovarian

cancer patients undergoing treatment (40).

While HE4 is most commonly used in combination with CA 125 to aid in the diagnosis and

monitoring of patients with ovarian cancer (38, 39), new roles for this marker are currently being

assessed as various other cancers have been correlated with elevated levels of HE4. For

example, the role of HE4 as a prognostic factor in lung adenocarcinomas has also been explored

(39, 77). Elevated levels of HE4 have also been observed in gastrointestinal and renal cancers

(38).


In addition to the role of HE4 as a tumor marker, it is also an important biomarker in a variety of

other benign conditions including: chronic kidney disease, renal failure and kidney fibrosis (85-

87). In addition, pregnancy, menopausal status and smoking can also impact the levels of HE4

in women (88-90).


To my knowledge, no studies examining levels of HE4 in a healthy pediatric population exist.

16

1.11 Pro-Gastrin-Releasing Peptide (ProGRP)


Pro-Gastrin Releasing Peptide (ProGRP) is a proenzyme synthesized from a pre-propeptide,

preproGRP. Subsequently, the C-terminal fragment of ProGRP is cleaved to produce Gastrin-

releasing peptide (GRP), which was initially identified for its role in stimulating the release of

gastrin (91). However, after discovering very high levels of ProGRP in human fetal and neonatal

lung, it has been hypothesized that GRP may play a role in the growth and development of both

the lung and the alimentary tract during the fetal and neonatal period (92).

The majority of the literature in this area focuses on the biological activity of amidated forms of

GRP. To date, GRP has been implicated in a variety of physiological as well as pathological

processes. Specifically, GRP has been implicated in exocrine and endocrine secretion, smooth

muscle contraction, pain transmission, blood pressure and body temperature regulation, satiety,

and behavior (93, 94). In addition, GRP has been linked to cancer progression via its mitogenic

and morphogenic proangiogenic action (95, 96).

The literature on precursor peptides like ProGRP is more sparse, however, it is now widely

accepted that propeptides like ProGRP can also be biologically active. Specifically, it has been

shown that propeptides can act on the same receptors as their peptide counterparts as well as on

novel receptors (91). In addition, evidence that ProGRP is found in high levels in the serum of

lung cancer patients has peaked interest in the biological function of this propeptide in cancer

(97). A recent study demonstrated that recombinant and synthetic ProGRP peptides were capable

of stimulating proliferation and migration in colorectal cancer cell lines. In addition, shRNA

knockdown of ProGRP resulted in decreased levels of phosphorylated ERK1/2. Although the

specific pathway involved in the action of ProGRP has not been elucidated, this is a significant

finding given the importance of kinases in many human cancers (97). Finally, it has also been

shown that in vivo injection of ProGRP-derived peptides resulted in stimulation of mitosis in the

colons of normal mice (97). Thus, like many of the other oncofetal antigens identified so far, it

would appear that ProGRP is not only important to fetal growth and development but, also, to

certain processes that are important to cancer progression.

17


Given the role of ProGRP in the development of the lung, it is not surprising that, as a tumor

marker, ProGRP is most often used in the context of lung cancer. A number of other tumor

markers have been investigated for their potential use in lung cancer including CEA and

squamous cell carcinoma antigen (SCC). However, primarily due to the fact that none of these

markers are specific to lung cancer [See Section 1.9 and 1.13], these markers have not been ideal

for use in lung cancer patients. ProGRP on the other hand, does not appear to be elevated in other

types of cancer with the exclusion of some tumors of neuroendocrine origin (98, 99).

Specifically, ProGRP, like many of the other markers examined here, is thought to be most

useful in monitoring patients during treatment and for predicting relapses in patients who are in

remission (100).

Although ProGRP does not currently have any pediatric-specific applications and lung cancer is

rare in children it does sometimes occur (101). As such, it is useful to have properly established

reference intervals in the pediatric population for this biomarker.


A study examining ProGRP levels in normal patients and patients with benign disease

discovered that approximately 10% of patients with benign disease had abnormal levels of

ProGRP. The highest proportion of these patients were those with renal failure (102). An

additional study showed similar results (99). Therefore, it is important for clinicians to consider

kidney function when interpreting ProGRP test results.


To date, one study has examined the ProGRP levels in the serum of relatively healthy patients

from birth to 16 years of age but due to a small sample size, reference intervals were not

established (92).

1.12 Free and Total Prostate Specific Antigen (PSA)


Prostate specific antigen (PSA), also known as human kallikrein-related peptidase 3 (hk3), is an

androgen-regulated serine protease and part of the kallikrein family of proteases (103).

18

Originally, it was thought to be uniquely secreted by epithelial cells in the prostate into male

ejaculate in order to liquefy semen and break down cervical mucous. However, additional

production sites for PSA have been discovered, namely, low levels of PSA have been discovered

in the endometrium, brain and breast tissue (103, 104).

Although little is known about the biological or pathological role of this marker, there is

accumulating evidence that PSA has functions in addition to its role in the liquefication of semen

and breakdown of cervical mucous. A small number of studies have been conducted that have

provided some information on the biochemical properties and functions of this tumor marker in

cancer progression. Notably, these studies have suggested that PSA plays a role in cell

proliferation in both physiological and pathological contexts through interaction with the insulin-

like growth factor (IGF) axis. Specifically, it is thought that PSA can hydrolyze IGF binding

protein (IGFBP) to increase IGF-1 bio-availability (105-107). There is also evidence that PSA

can activate a number of mitogenic proteins like transforming growth factor- β (TGF-β) and can

also degrade the extracellular matrix; therefore, PSA may contribute to both cell proliferation

and the spread of cancer cells to secondary sites (108, 109).

PSA also has anti-angiogenic properties through inhibition of basic fibroblast growth factor

(bFGF) and vascular endothelial growth factor (VEGF) (110, 111). This is a somewhat

counterintuitive property for a tumor marker as their biological functions tend to be linked to

activities that promote tumor growth and cancer progression. However, it is thought that this

anti-angiogenic property of PSA may be linked to the low cell proliferation and slow growth

observed in prostate cancer (112).


Free and Total PSA are used primarily in the context of prostate cancer (40). With family

history being a major risk factor for the development of prostate cancer, early detection and

screening programs have become increasingly important (113). Although PSA is frequently used

as a marker for screening and early detection of prostate cancer, there remains a great deal of

controversy surrounding whether or not this is a useful practice. One large scale European study

compared a control group to a PSA-screened group of men. Although death rates from prostate

cancer were 20% lower in the PSA-screened group, the screening also resulted in over-detection

19

and over-treatment (114). Another large study found no difference between the screened and

controlled group of men (115).

As is the case with many of the other markers examined here, there are also limitations to the

use of PSA as a diagnostic tool (109) and most guidelines recommend that PSA be used

primarily for staging and prognosis, as well as for surveillance and monitoring of patients

undergoing therapy and to monitor patients in remission (40).

Recent studies have begun to explore the use of free/total PSA ratio measurements in screening

of high-risk individuals (even if asymptomatic) (116). However, the evidence remains fairly

sparse and more work is required to assess the usefulness of the free/total PSA ratio in screening

efforts. Some guidelines, however, have recommended the use of free/total PSA ratio

measurements to differentiate benign prostatic disease from prostate cancer, and also as a follow-

up for patients with negative biopsies or in patients with increased biopsy risk (40).

There is also some emerging interest on the potential role of PSA as a breast cancer marker.

Studies have demonstrated that elevated levels of PSA were associated with breast cancer in both

pre- and post-menopausal women (117).

Although the current clinical applications for PSA apply primarily to cancers that typically arise

in the adult population, the potential applications for this marker are expanding and there is a

paucity of information on normal levels in children and teens. Therefore, as research on the

potential applications for PSA continues, it is useful to have an understanding of how analyte

levels fluctuate in the pediatric population.


PSA levels can also be altered in a number of different conditions and physiological states. For

example, PSA concentrations appear to be higher in menstruating women, particularly, during

the mid to late follicular phase. These fluctuations seem to reflect changes in serum progesterone

levels (118). The potential for PSA to be used in the diagnosis of polycystic ovary syndrome is

also currently being investigated, as elevated PSA levels correlate with hirsutism and the

hyperandrogenism state in women (119).

20

In addition, PSA levels also appear to be elevated in prostatitis, benign prostatic hypertrophy and

prostatic trauma. It is also important for clinicians to consider that levels of PSA are known to

be elevated after ejaculation (120, 121).


A study by Randell et al. calculated 95th percentile total PSA in children aged birth to 18 years of

age using the IMMULITE analyzer (Diagnostics Products Corp.). However, it is important to

note that this study was carried out using children who had been hospitalized for special care or

surgery (122).

1.13 Squamous Cell Carcinoma Antigen (SCC)


The biological role of SCC has not been well characterized, however, SCC is known to belong to

the serine proteinase inhibitor (serpin) family of proteins. The SCC antigen is typically found

inside squamous cells and tends to be released outside the cells in squamous cell carcinomas

(123).

Like many of the other cancer markers examined here, the majority of the literature has focused

on the potential clinical uses of the SCC antigen, however, there is some information currently

available on the role of the SCC antigen in both physiological and pathological contexts. In

normal tissue, SCC tends to be found on the apical side of squamous epithelium. SCC can also

be detected in human fetal skin at approximately 16 weeks gestation which suggests a potential

role in epidermal maturation (124).

There is also some evidence that SCC is capable of contributing to the promotion of cancer

progression. Specifically, transduction of SCC-negative tumor cells with the SCC antigen

resulted in significantly increased tumor survival compared to non-transduced cells. In addition,

these transduced cells were resistant to apoptosis mediated by NK cells (125). Additional studies

have shown that SCC can suppress apoptotic cell death by NK cells and promote cell invasion

(126). The mechanism by which SCC inhibits apoptosis remains unclear. However, it has been

noted that incubation of squamous cells with tumor necrosis factor α (TNFα) increased the

production of SCC (127). Therefore, it is possible that SCC antigen production occurs as a

21

response to TNFα-induced apoptosis. Although there remains a paucity of information on the

biological role of SCC there is early evidence that SCC is not merely a byproduct of tumor

metabolism but, rather, may play an important role in enhancing tumor progression in addition to

its physiological role in epidermal maturation (125-127).


SCC was initially thought to be useful primarily in the context of cervical cancer (128).

Specifically, high levels of SCC antigen in the early stages of squamous cell carcinomas

correlate with recurrence rates, lymph node metastasis, lymphovascular space invasion, deep

stromal invasion and larger tumor size (129). However, there is controversy surrounding whether

or not the SCC antigen, on its own, is an independent risk factor for recurrence (125, 129, 130).

Serial measurements of SCC may also be useful for monitoring of cervical cancer patients

undergoing treatment and those in remission. The role of SCC in the latter, however, remains

controversial (131), primarily because SCC follow-up measurements in patients with early-stage

cervical cancer do not appear to improve clinical outcome. In addition, SCC can become falsely

elevated in patients without recurrence (132). It is possible that SCC measurement in addition to

other techniques like CT scanning or MR imaging may be more useful to monitor patients with

later stage cervical cancer (133).

Like many of the other markers examined here, although elevations in SCC levels can sometimes

precede symptoms, there is no evidence from large scale, prospective studies that SCC

measurements are useful for early detection or screening programs (134).

Although SCC has primarily been used in the context of cervical cancer, like many of the other

markers examined here, as our knowledge of this tumor marker increases, its potential

applications as a marker for other types of cancer becomes more apparent. SCC is elevated in

various other cancers including skin, head and neck, esophageal and lung cancers (128). As such,

there may be additional utility for SCC as a biomarker in a variety of cancers. Perhaps most

important, is the potential role it may play in skin cancer which has become increasingly

prevalent in all age groups, including the pediatric age range due, in part, to the popularity of

artificial tanning (135).

22


A study conducted by Molina et al. examined levels of SCC in patients with malignant and

benign diseases. Approximately 30% of patients with benign pathologies had elevated levels of

SCC. Of the conditions examined, the highest proportion of elevated SCC levels were found in

patients with renal failure and patients with lung diseases and head and neck diseases (136). The

authors discovered that exclusion of patients in renal failure or with creatinine values > 133

µmol/L improved the specificity of SCC (136). These results highlight the importance of testing

renal function when using SCC as a biomarker to monitor progress during cancer treatments. In

addition, elevated levels of SCC have also been associated with a number of skin diseases

including eczema, pemphigus, erythoderma epidermitis and psoriasis (128).


To my knowledge, no studies examining levels of SCC in a healthy pediatric population exist.

1.14 Sensitivity and Specificity of Cancer Biomarker Assays

The sensitivity and specificity of an assay may concern either its analytical properties or its

diagnostic/clinical properties. The analytical sensitivity of an assay refers to the smallest amount

of a substance that can be detected by the assay, and the analytical specificity of an assay refers

to how well an assay can detect the substance of interest rather than other interfering substances.

On the other hand, the diagnostic sensitivity of an assay refers to the percentage of patients with

a given disorder who are correctly diagnosed as positive for the condition by the assay, and

diagnostic specificity refers to the percentage of patients without a given condition who are

correctly diagnosed as negative for the condition (137). The analytical sensitivity and specificity

of the 11 assays examined in the present study are listed in Table 2.

The diagnostic/clinical sensitivity and specificity of the markers are difficult parameters to

estimate. This is largely due to the heterogeneity between studies conducted to assess the

clinical utility of these markers. For instance, in many cases the cut-offs used in different studies

are not the same; this can alter the estimates of specificity and sensitivity considerably. In

addition, the characteristics of the patients examined can be different between studies and even

within a given study; for example, patients may have varying stages of cancer, or males and

females may be used interchangeably. These studies also examine the use of markers in different

23

types of cancers and may examine diagnostic, prognostic or monitorial potential of the markers.

With all of these factors at play, combined with the substantial gaps in knowledge with regards

to reference values, it is not surprising that it is incredibly difficult to estimate the clinical

sensitivity and specificity of these markers.

1.15 Rationale for Development of Cancer Biomarker Reference Intervals

The previous sections have illustrated the many potential clinical applications of the markers

examined in the present study. In addition, the many substantial gaps in information regarding

the levels of these markers in the healthy population have been highlighted. In order to take full

advantage of the potential of these biomarkers to aid in predicting patient outcomes and

monitoring treatment progress, it is important to have age- and sex-stratified reference range

values for comparison. The importance of age- and sex-stratified reference intervals for tumor

biomarkers has been previously noted in the adult population (138). Specifically, Gutierrez et al.

discovered that reference intervals for CA 125 that were stratified by sex and menopausal status

(in females) had increased prognostic value compared to reference intervals that did not take into

account these factors (138). It is reasonable to assume that this may also be the case in the

pediatric population where pivotal developmental stages akin to menopause (i.e. puberty) are

likely to impact the levels of these markers.

In addition to the clinical importance of reference intervals, an understanding of how levels of

key biomarkers fluctuate in the normal population is also important from a research perspective.

Studies have illustrated the ways in which gaps in pediatric reference intervals can hinder our

ability to assess the prognostic value of a biomarker. For example, Baranzelli et al. carried out

an assessment of various prognostic factors in children with malignant germ cell tumors (23).

Among the factors assessed, was AFP level; using multivariate analysis, the authors determined

that AFP was the most important prognostic variable. However, the authors also noted that

elevated AFP levels during infancy and early childhood precluded accurate statistical analysis

and, therefore, children under the age of 1 were excluded from the analysis (23). Properly

established, age-stratified reference intervals encompassing the entire pediatric range may

simplify statistical analysis for these types of studies and improve the ability of researchers to

assess how and when tumor markers ought to be used clinically. An additional study noted

24

similar issues in measuring AFP levels in the first 18 months of life (24). Specifically, while

seeking to assess the value of AFP, CA 125 and CA 19-9 in follow-up monitoring of patients

with sacrococcygeal teratomas, the authors observed that one third of patients who did not have a

recurrence still had at least one AFP measurement that fell above the reference range at some

point during follow-up. All of these patients were 18 months of age or younger. It was suggested

that this may be due to the wide reference range during the first year of life (24, 139). It is

possible that this wide reference range indicates a need for additional age-partitioning within this

age group. As Gutierrez et al. demonstrated for CA 125, the usefulness of AFP as a marker for

follow-up monitoring may also be enhanced with appropriate age-partitioning (138).

1.16 Rationale for Development of Cancer Biomarker Reference Intervals in a Pediatric Population

With the majority of the literature on tumor biomarkers focusing on the adult population, it is

important to note the reasons for establishing reference values in a pediatric population

specifically. There are several reasons why this aim is crucial.

Firstly, it is important to note that there is a paucity of information on the levels of these markers

in a healthy pediatric population. As highlighted in Table 1, there is little to no information

available for 7 out of the 11 markers examined in this study. The findings of this study will be

important in closing crucial gaps in our basic knowledge of these 11 biomarkers.

Closing the gaps in the basic knowledge surrounding levels of these markers in a healthy,

pediatric population is crucial as there is early evidence for potential applications in pediatric

cancers for several of these markers. Notably, CA 19-9, CA 125 and AFP to various types of

pediatric cancers including yolk sac tumors and teratomas, leukemia and lymphoma Specifically,

elevated levels of AFP have been associated with yolk sac tumors, immature teratomas and

malignant germ cell tumors. In addition, elevations in CA 19-9 and CA 125 measurements have

also been observed in patients with immature teratomas and malignant germ cell tumors (53).

CA125 my also have additional applications in both leukemia and lymphoma. Two separate

studies have demonstrated that a high percentage of children with leukemia (50%) and

lymphoma (61%) have elevated levels of CA 125 (59, 60).

25

In addition, it is also important to note that elevations of these markers are not only associated

with cancer but a variety of other conditions described in Sections 1.4 to 1.13. In order for these

markers to be used to aid in the diagnosis, prognosis and monitoring of patients with these non-

cancerous conditions, it will be important for physicians to have access to robustly established

reference intervals based on a truly healthy pediatric population. Therefore, these reference

intervals will have application beyond cancer for a variety of conditions that can arise in the

pediatric population.

Secondly, tumor biomarkers are not simply byproducts of tumor metabolism. Rather, many of

these markers have been shown to play an important role in cancer progression. For example,

recent studies seeking to characterize the cellular function of HE4 have demonstrated that HE4

overexpression enhances malignant phenotypes both in cell culture and in a mouse model (76).

Similarly, the role of CA 125 in cancer progression has been explored extensively and linked to

metastasis, motility, proliferation and chemoresistance (54). It is, therefore, reasonable to

speculate that these markers may also play a role in the progression of pediatric cancers. Indeed,

as described in Sections 1.4 to 1.13 it is common to see the application of these tumor markers

expanding from one to several different types of cancers as research progresses.

Finally, it is important to note that disease incidence can change over time and, thus, cancers that

have, in the past, typically arisen in the adult population can begin to arise in the pediatric

population. An excellent example of this is melanoma. Likely due to the advent of indoor

tanning, the incidence of melanoma in the pediatric population, particularly in female teenagers,

has been steadily increasing since 1973 (140). Thus, tumor markers like SCC may now begin to

play an important role in pediatric as well as adult cancers.

1.17 Hypothesis & Objectives

The present study sought to close the gaps that currently exist in reference values for 11 key

cancer biomarkers. In previous CALIPER studies, it has been observed that the vast majority of

markers require some partitioning by age and/or sex due to the significant developmental

changes that occur throughout the pediatric age range (3-5, 9). Therefore, I hypothesized that

circulating concentrations of tumor biomarkers would be significantly influenced by child sex

and growth (age). The major objectives of my research were:

26

1. Establish population reference values for key circulating tumor biomarkers.

2. Understand how key covariates (age & sex) effect tumor biomarker levels.

3. Develop a comprehensive database of covariate-stratified reference intervals for cancer

biomarkers.

These reference intervals will not only be useful in practice by contributing to a more accurate

assessment of tumor markers in cancer diagnosis and treatment, but also in research efforts

aimed at determining the prognostic, diagnostic and monitorial value of tumor markers in various

types of cancers and populations.

27

Table&1.&Clinical&applications&and&available&pediatric&reference&intervals&for&11&key&circulating&tumor&biomarkers.!MARKER& CLINICAL&RELEVANCE&

PEDIATRIC&CANCER&CLINICAL&RELEVANCE&ADULT&CANCER*&

AVAILABLE&PEDIATRIC&REFERENCE&INTERVALS&

AlphaKFetoprotein&(AFP)&

&

Monitoring!of!sacroccygeal!teratomas!(24)!Prognosis!of!germ!cell!tumors!(23)!

Diagnostic!aid!in!yolk!sac!tumors!(22)!

Diagnosis!and!prognosis!of!hepatocellular!carcinoma!patients!(141)!

Reference!Intervals!established!by!CALIPER!using!previous!generation!Abbott!Diagnostics!AFP!Assay!(3)!

AntiKThyroglobulin&(AntiKTg)&

! Monitoring!of!thyroid!disease,!including!thyroid!cancer!(30)!

95th!percentile!for!individuals!aged!birth!–!20!years!reported!(35)!

Cancer&Antigen&15K3&(CA15K3)&

&

! Surveillance!of!breast!cancer!patients!following!diagnosis!(141)!

None!reported!

Cancer&Antigen&19K9&(CA&19K9)&

&&

Elevated!in!some!malignant!germ!cell!tumors!and!immature!teratomas!(53)!

Monitoring!of!pancreatic!cancer!(141)!Used,!in!combination!with!CEA,!in!

diagnosis!and!followPup!of!colorectal!and!gastrointestinal!cancer!patients!(22)!

Reference!intervals!available!for!children!birth!–!1.5!years!(53)!

Cancer&Antigen&125&(CA&125)&

&&

Potential!to!aid!in!monitoring!of!patients!with!sacroccygeal!teratomas!(24)!

Elevated!levels!observed!in!children!with!NonPHodgkins!Lymphoma!and!leukemia!

(60)!

Used,!in!combination!with!HE4,!to!aid!in!diagnosis!and!monitoring!of!!ovarian!

cancer!patients!(76)!

Reference!intervals!available!for!children!birth!–!1.5!years!(53)!

Carcinoembryonic&Antigen&(CEA)&

&

! Monitoring!of!colon!cancer!patients!postPtreatment!(141)!

None!reported!

Human&Epididymis&Protein&4&(HE4)&

&

! Used,!in!combination!with!CA!125,!to!aid!in!diagnosis!and!monitoring!of!!ovarian!

cancer!patients!(76)!Potential!prognostic!factor!in!lung!cancer!

(142)!

None!reported!

ProKGastrinKReleasing&Peptide&

(ProGRP)&

! Monitoring!tool!for!lung!cancer!patients!undergoing!treatment!and!in!remission!

(142)!

None!reported!

Squamous&Cell&Carcinoma&Antigen&

(SCC)&

Elevated!in!various!skin,!esophageal,!lung,!head!and!neck!cancers!(128)!

Can!help!determine!prognosis!for!cervical!cancer!patients!(128)!

!

None!reported!

Free&&&Total&Prostate&Specific&Antigen&(Free&PSA)&

&

! Used!in!prostate!cancer!screening!efforts,!prognosis!and!monitoring!(141)!

!95th!percentile!of!Total!PSA!for!individuals!aged!birth!–!18!years!

reported!(122)!

*Relates!to!cancers!that!typically!occur!in!the!adult!population,!although!they!may!still!occur!in!pediatric!patients.

28

Table&2.&Analytical&Specificity&and&Sensitivity&of&11&Cancer&Marker&Assays&on&the&Abbott&Architect&ci4100.!MARKER& Analytical&

Sensitivity&Analytical&Specificity&(%)&

AlphaKFetoprotein(AFP)&& ≤!1.0!ng/mL! ±!10!

AntiKThyroglobulin&(AntiKTg)&& ≤!1.0!IU/mL! ≤!15!

Cancer&Antigen&15K3&(CA&15K3)& ≤!1.0!U/mL! ≤!15!

Cancer&Antigen&19K9&(CA&19K9)& ≤!2.00!U/mL! ±!12!

Cancer&Antigen&125&&(CA&125)& ≤!1.0!U/mL! ≤!12!

Carcinoembryonic&Antigen&(CEA)& ≤!0.5!ng/mL! ≤!10!

Human&Epididymis&Protein&4&(HE4)& ≤!15!pmol/L! ±!15!

ProKGastrinKReleasing&Peptide&(ProGRP)& ≤!3!pg/mL! ≤!10!

Squamous&Cell&Carcinoma&Antigen&(SCC)& ≤!0.1!ng/mL! ≤!10!

Free&Prostate&Specific&Antigen&(Free&PSA)& ≤!0.008!ng/mL! ≤!10!

Total&Prostate&Specific&Antigen&(Total&PSA)& ≤!0.008!ng/mL! ≤!10!

29

Chapter 2 Materials and Methods

2 Materials and Methods

2.1 Participant Recruitment and Sample Acquisition

2.1.1 Participant Recruitment

This study was approved by the Research Ethics Board at the Hospital for Sick Children,

Toronto, Canada as well as the Research Ethics Board at McMaster Children’s Hospital in

Hamilton, Canada. Healthy children and adolescents from birth to 18 years of age were recruited

from the greater Toronto and Hamilton areas from schools, community centers and sports

leagues. Project Coordinators at the Hospital for Sick Children and McMaster Children’s

Hospital contacted program coordinators and other staff members at these community groups

and schools to share information about the CALIPER Project and establish recruitment

partnerships with these locations.

Assembly presentations and information sessions were set up with locations willing to partner

with the CALIPER project by allowing the CALIPER staff and volunteer team to set up a clinic

at their site. The assembly and information sessions informed both the parents as well as the

children and teens about the importance of reference intervals, the potential risks and benefits of

participation in CALIPER and the details of the participation process. In addition, prospective

participants were given the opportunity to ask questions about the CALIPER project and the

participation process. During assembly presentations and information sessions, parents and

children were given both questionnaires and consent forms to fill out should they choose to

participate in the project.

2.1.2 Sample Acquisition

Subsequently, CALIPER staff and volunteers set up clinics at these various community locations

in order to collect blood samples. Willing participants returned completed questionnaires and

signed consent forms which were verified by staff and volunteers. Measurements including

height, weight and waist were taken for each participant. In some cases, participants were asked

30

to self-report on Tanner stage, a measure of sexual maturity (143). Next, participants donated a

blood sample; participants aged birth - < 1 year donated 5 mL of blood, participants aged 1 - <

11 years donated 10 mL of blood and participants aged 11 - < 19 years donated 15 mL of blood.

Finally, participants were compensated $10 for their participation in CALIPER, given two hours

of community service and the choice of a t-shirt, teddy bear or storybook.

2.1.3 Sample Processing

Following clinics, samples were transported back to the Hospital for Sick Children or

McMaster’s Children’s Hospital for processing. Samples collected in serum separator tubes

(SSTTM; BD) were centrifuged at 4000 rpm for 8 minutes within 30 minutes to 4h following

collection. The separated serum was then aliquoted and stored at -80 °C (CALIPER biobank)

until testing at the Hospital for Sick Children.

2.2 Participant Selection

Participant demographic and health information was collected using the self-report

questionnaires and subsequently entered into a Microsoft Access Database hosted on the

Hospital for Sick Children server. Based on the information provided in these questionnaires,

participants were then selected for the study. Firstly, participants were excluded based on three

criteria: history of chronic illness or metabolic disease, acute illness within the two weeks

preceding donation, regular use of prescribed medication or use of prescribed medication within

the two weeks preceding donation. Next, the study population was selected randomly using the

remaining healthy participants, and ensuring a balanced age and sex distribution. Furthermore,

efforts were made to ensure that the ethnic composition of the study participants reflected that of

the province of Ontario as reported by 2006 Canadian census data (144).

In order to ensure a sufficiently large sample size for participants less than 5 years of age, in

addition to CALIPER samples, samples from apparently healthy/metabolically stable children

from the maternity wards of Women’s College Hospital and Mt. Sinai Hospital in Toronto,

Canada, and select outpatient clinics at the Hospital for Sick Children were used. These clinics

included: Allied Health Clinic, General Surgery, Medical Day Care, Ambulatory Surgical,

Orthopaedics, Otolaryngology, Adolescent Medicine, Plastic Surgery and Dentistry. Test results

31

from blood work ordered by these clinics were examined before selection and patients with

abnormal results were not included in the study.

Finally, selected samples were located in the biobank and stored at -80 °C until testing. Samples

that were removed from the biobank were tracked manually and the Microsoft Access Database

was updated to reflect the changes in sample availability after testing.

2.3 Sample Analysis

2.3.1 Instrument and Assays

Serum samples for selected participants were analyzed on the Abbott ARCHITECT ci4100

system for each of the 11 key cancer biomarkers: Alpha-fetoprotein (AFP), Cancer Antigen 15-3

(CA 15-3), Cancer Antigen 125 (CA 125), Cancer Antigen 19-9 (CA 19-9), Carcinoembryonic

Antigen (CEA), Human Epididymis Protein 4 (HE4), Pro-Gastrin-Releasing Peptide (ProGRP),

Anti-Thyroglobulin (Anti-Tg), Squamous Cell Carcinoma Antigen (SCC), Total and Free

Prostate Specific Antigen (PSA). See Table 3 for a summary of the number of samples tested for

each analyte. Samples were analyzed in batches over a 10-month period.

2.3.2 Principles of Procedure

The morning of testing, samples were thawed in at 5 °C. Subsequently, the required amount of

sample was loaded into a cuvette and onto the Abbott ARCHITECT ci4100 and scheduled tests

were ordered. The Abbott ARCHITECT ci4100 uses a chemiluminescence immunoassay method

with a two-step immunoassay procedure. Specifically, a primary antibody coated with

paramagnetic microparticles was combined with the serum sample. Subsequently, after washing

with phosphate buffered saline solution, an acridinium-labeled antibody conjugate was added.

After a second wash, pre-trigger solution containing hydrogen peroxide and trigger solution

containing sodium hydroxide were added to the sample. This resulted in a chemiluminscent

reaction which was measured as relative light units (RLUs). The RLU detection correlates

positively with the amount of analyte in the sample. Table 4 provides a summary of the

antibodies used for each of the analytes examined in this study.

32

Table 3. Number of Samples Tested for Each Analyte.

Analyte Samples Tested

AFP 474

Anti-Tg 402

CA 15-3 403

CA 19-9 474

CA 125 689

CEA 709

Free PSA 625

HE4 625

ProGRP 630

SCC 431

Total PSA 700

33

Table 4. List of Antibodies Used in Abbott Architect ci4100 assays*

Assay Microparticles Conjugate

Alpha-Fetoprotein (AFP) Anti-AFP (mouse, monoclonal) Anti-AFP (mouse, monoclonal)

Anti-Thyroglobulin (Anti-

Tg)

Human Thyroglobulin Anti-human IgG (mouse,

monoclonal)

Cancer Antigen 15-3 (CA

15-3)

115D8 (mouse, monoclonal) DF3 (mouse, monoclonal)

Cancer Antigen 19-9 (CA

19-9)

1116-NS-19-9 (mouse,

monoclonal)

116-NS-19-9 (mouse, monoclonal)

Cancer Antigen 125 (CA

125)

Anti-CA 125 (mouse,

monoclonal)

Anti- CA 125 (mouse, monoclonal)

Carcinoembryonic Antigen

(CEA)

Anti-CEA (mouse, monoclonal) Anti-CEA (mouse, monoclonal)

Human Embryonic Protein

4 (HE4)

Anti-HE4 (mouse, monoclonal) Anti-HE4 (mouse, monoclonal)

Pro-Gastrin-Releasing

Peptide (ProGRP)

Anti-ProGRP (mouse,

monoclonal)

Anti-ProGRP (mouse, monoclonal)

Squamous Cell Carcinoma

Antigen (SCC)

Antibody to SCC Ag (mouse,

monoclonal)

Antibody to SCC Ag (mouse,

monoclonal)

Free Prostate Specific

Antigen (PSA)

Anti-Free PSA (mouse,

monoclonal)

Anti-PSA (mouse, monoclonal)

Total Prostate Specific

Antigen (PSA)

Anti-PSA (mouse, monoclonal) Anti-PSA (mouse, monoclonal)

*Antibodies used were those that were included in Abbott Diagnostics’ ARCHITECT reagent

kits

34

2.3.3 Calibration and Quality Control

Analytical methods were controlled according to manufacturer’s instructions by preventive

maintenance, function checks, calibration and quality control. Specifically, calibration was

carried out after initial test set-up on the ARCHITECT ci4100 and each time a reagent kit with a

new lot number was used.

In addition to calibration, quality control procedures were carried out for each assay prior to

testing, for each day that sample testing was conducted. Specifically, a single sample of each

control level for a given assay was tested. Results of this testing were examined in order to

verify that the values were within the recommended range for each control level. Performance

characteristics of each assay are summarized in Table 5. Finally, all samples tested underwent

automated interference analysis for hemolysis, icterus and turbidity.

35

ConventionalUnits

Quality Control 1Mean

SDCV (%)


SDCV (%)


SDCV (%)

Linearity Reference Material(Standardization) Reference Method

Alpha-fetoprotein ng/mL19.510.804.1

198.58.064.0

938.3638.78

4.10.91 - 2487.76 WHO 1st International Standard

72/225 for AFP NA

Anti-TG IU/mL0.550.0611.7

146.023.342.3

----- 3.0 - 1000.0WHO 1st InternationalReference Preparation

65/093NA

CA 125 U/mL37.62.15.5

283.92.15.5

611.627.24.5

1.0 - 1000.0Referenced to a

Standard Prepared ByFujirebio Diagnostics, Inc

NA

CA 15-3 U/mL36.02.98.0

247.716.06.5

----- 0.5 - 800.0Referenced to a


NA

CA 19-9 U/mL37.492.637.0

136.797.745.7

722.5454.55

7.62 - 1200

Referenced to aStandard Prepared By

Fujirebio Diagnostics, IncNA

CEA ng/mL4.960.285.7

19.461.065.5

100.754.794.8

0.5 - 1500.0 WHO 1st International Standard73/601 for CEA NA

HE4 pmol/L49.32.85.7

167.811.36.7

667.355.08.2

20.0 - 1500.0Referenced to a


NA

ProGRP pg/mL40.882.927.1

152.3615.3510.1

2452.54121.24

4.93 - 5000 Abbott Internal

Reference Standard NA

Free PSA ng/mL0.4000.012

3.0

1.0210.034

3.3

7.0350.299

4.20.008 - 30.00 Stanford 90:10 PSA

Reference Material NA

Total PSA ng/mL0.4380.022

5.0

3.5250.140

4.0

21.9271.136

5.20.05 - 100.00 WHO 1st International

Standard for PSA (90:10) 96/670 NA

SCC ng/mL2.00.15.8

10.30.55.1

50.13.46.7

0.1 - 70.0 Abbott InternalReference Standard NA

SD = Standard deviation.

CV = Coefficient of variation.

NA = Not Applicable

Analytical imprecision data (using two or three levels of quality control materials) for the duration of the study (June to April).

Analytical Imprecision

Table&5:&Analytical)Performance)of)Immunoassays)on)the)ARCHITECT)ci4100

36

2.4 Statistical Analysis and Determination of Reference Intervals

Data were analyzed in accordance with CLSI C28-A3 guidelines as outlined in Figure 1.

Statistical analysis was performed with Excel (Microsoft) and R software. Briefly, scatter and

distribution plots were used to visually inspect the data. Extreme outliers were removed using

visual inspection and suspected partitions were identified. Suspected partitions identified using

visual inspection were subsequently tested using the Harris & Boyd method (145). In order to

warrant partitioning one of the following criteria must be met:

where:

Where x represents the sample mean, n represents the sample size and s represents the standard

deviation.

Importantly, this method examines not only the differences in means between the two subgroups

in question but also the differences in standard deviation.

Due to the highly skewed nature of the data, once suspected partitions were either confirmed or

rejected by the Harris & Boyd method, the data in each individual partition were then

transformed using the Box-Cox transformation method in order to obtain a Gaussian distribution

(146).

The normality of each of the partitions following the Box-Cox transformation was then assessed

using Q-Q plots (147). Outlier removal in normally distributed partitions was carried out using

the Tukey test twice (148). This method involves a calculation of the interquartile range (IQR) as

well as the first and third quartile (Q1 and Q3). Data points that fell above Q3 + (1.5*IQR) or

37

below Q1 – (1.5*IQR) were deemed to be outliers. For normally distributed data, it was expected

that approximately 0.7% of the data would lie outside of this range. However, for skewed data,

the probability could be much higher, therefore, outlier removal for skewed data was carried out

using the adjusted Tukey test twice. In contrast to the regular Tukey test, the adjusted Tukey test

creates fences that are non symmetrical to account for the skewness of the data (149).

Specifically, depending on the direction of skewness, outliers were defined as data points that

fell outside of either of these ranges:

or this range:

Where MC stands for medcouple which is a robust measure of skewness:

Where med refers to the sample median and:

Where medn represents the sample median, xi represents the smallest observation and xj

represents the largest observation.

Finally, upon completion of outlier removal, reference intervals were calculated for partitions

with a sample size ≥ 120 using a non-parametric rank method (2). This is a simple method that

does not require any assumption regarding the distribution of the data. In order to calculate the

reference interval, the reference values were sorted from least to greatest and given a rank value

from 1 to n where n was the number of samples in the partition of interest. Next, the rank of the

values that were deemed the upper and lower limit of the reference intervals (r1 and r2) were

38

calculated. Specifically, to calculate the central 100(1-α)%, r1 ([α/2]th percentile) and r2 ([1 -

α/2]th percentile) were calculated as follows:

The corresponding test results for the rank values of r1 and r2 were used as the upper and lower

limit of the reference interval.

Finally, the 90% confidence interval of the reference limits were calculated using ranked

observations. The ranks of the lower limit ([α/2]th percentile) confidence interval for a 100(1-

c)% confidence interval were defined as a and b where:

On the other hand, the ranks of the upper limit ([1 - α/2]th percentile) confidence interval for a

100(1-c)% confidence interval were determined by y = n – b + 1 and z = n - a + 1. The

corresponding test results for the rank values of a, b were used to delineate the lower limit

confidence interval of the calculated reference interval while the corresponding test results for

the rank values of y and z were used to delineate the upper limit confidence interval of the

calculated reference interval.

In contrast, reference intervals for partitions with a sample size > 40 and < 120 were calculated

using the robust method of Horn and Pesce (150). The robust method uses an iterative process to

estimate the center and spread of a dataset. Specifically, the center of the dataset, denoted as Tbi

was initially estimated to be the median of the reference values. Next, the median absolute

deviation, which is more resistant to outliers than standard deviation, was calculated as follows:

39

This measure was used in order to weight each observation based on its distance from the

median. Subsequently, the weighted observations were used to recalculate Tbi. The initial Tbi

estimate was then compared to the new Tbi estimate based on the weighted observations and this

process was repeated until there was an insignificant change between the old and the updated Tbi

estimate. At this point, the upper and lower limit of the 100(1-α)% reference interval were

calculated as follows:

Finally, percentile bootstrap estimates were used to construct 90% confidence intervals for the

upper and lower limits of the calculated reference intervals (151).

40

!

Inspect(Data!to!ensure!there!are!no!missing!or!incorrect!values!

!

Inspect(Partitions(using!box!plots!and!scatterplots!to!identify!trends!in!age!and!sex!

!

Determine(Partitions((test!partitions!with!the!Harris!&!Boyd!method!

!

Transform(Data(using!Box;Cox!Method!

!

Check(Normality(using!Q;Q!Plot!and!Shapiro–Wilk!test!

!Normal(Data(

remove!outliers!using!Tukey!test!twice!

!

Skewed(Data(remove!outliers!using!

adjusted!Tukey!test!twice!!

Calculate(Reference(Interval!!

N(>(=(120(Calculate!reference!interval!using!non;parametric!rank!

method!!

N(<(120(&(>(40(Calculate!reference!interval!

using!robust!method!!

Calculate(90%(Confidence(Interval!!

Remove(extreme(outliers(using!visual!inspection!!

!

Figure 1. CALIPER study data-analysis protocol based on CLSI guidelines document C28-A3.

41

Chapter 3 Results

3 Results

3.1 Overall Findings

Approximately 400 to 700 pediatric samples from the CALIPER cohort, depending on the

analyte in question, were tested for each of the 11 cancer biomarkers examined and results were

used to calculate age- and sex-specific reference intervals (Table 6). All analytes examined

required some partitioning by age, sex or both, with the exception of Anti-Tg, which showed

steady levels of expression across the entire pediatric age range (Figure 2).

3.2 Typical Oncofetal Antigens

Six of the eleven analytes examined (HE4, ProGRP, SCC, CEA, CA 19-9 & AFP) all showed the

expected pattern for oncofetal antigens. Specifically, high concentrations were observed at birth

followed by a rapid decrease in concentration throughout the first year or first two years of life

(Figure 3 - 8). Several of these analytes also required some partitioning after the first year of

life; namely, HE4 (2 - < 10 years and 10 - < 19 years) and ProGRP (1 - < 12 years and 12 - < 19

years) although it is important to note that fluctuations that occurred in the 1 – < 19 year age

range were not nearly as drastic as those observed during the first year of life.

3.3 Atypical Oncofetal Antigens

Two of the analytes examined, CA 15-3 and CA 125 showed atypical patterns of expression

during the first year of life given their status as oncofetal antigens (Figure 9 - 11). CA 15-3

levels were lower during the first week of life with the upper limit of the reference interval

falling at 24 U/L, followed by an increase in concentration with the upper limit of the 1 week - <

1 year reference interval falling at 33 U/L (Table 6). Similarly, the upper limit of the 0 - < 4

month reference interval for CA 125 fell at 22 U/L whereas the upper limit of the 4 month - < 5

years reference interval fell at 33 U/L (Table 6).

42

3.4 Sex-Specific Partitioning

Three of the markers examined, CA 125, Total PSA and Free PSA, required sex-specific

partitions (Figure 10, 12, 13; Table 6). CA 125 required sex-specific partitions in the teenage

age group. Specifically, the upper limit of the reference interval for females aged 11 - < 19 years

was significantly higher than that of males (39 U/L vs. 28 U/L).

In contrast to CA 125 where only one age partition demonstrated sex differences, in the case of

Total and Free PSA, given the vastly different patterns of expression observed in males and

females, age partitioning for the two sexes was carried out separately. Several notable trends

were observed with respect to these two markers. Firstly, for Total PSA, higher upper limits

were observed in the 0 - < 1 week reference interval compared to the 1 week - < 6 months

reference interval in males (0.047 ng/mL vs. 0.038 ng/mL) (Figure 13). Similarly, higher upper

limits were observed in the 0 - < 1 week reference interval compared to the 1 week - < 1 year age

partition in females (0.039 ng/mL vs. 0.010 ng/mL) (Figure 13). In addition, for both Free &

Total PSA, increases in concentration were observed in males over the age of 12. Both Total and

Free PSA required a unique partition for males age 12 - < 19 years (Figure 12 & 13). In contrast,

to the dynamic changes observed in male and female Total PSA levels and male Free PSA

levels, only one Free PSA reference interval spanning the entire pediatric age range was required

for females (Figure 12).

43

Table&6.&Age+&and&Sex+stratified&reference&intervals&for&11&serum&immunoassay&cancer&biomarker&analytes&analyzed&on&Abbott&Architect&ci4100.a

Analyte Age Lower&Limit95th&

PercentileUpper&Limit

No.&of&Samples

Lower&and&Upper&Limit&Confidence&

IntervalsAge

Lower&Limit

95th&Percentile

Upper&Limit

No.&of&Samples

Lower&and&Upper&Limit&Confidence&

Intervals0"#"<"1"month " >"2000 >"2000 113 0"#"<"1"month " >"2000 >"2000 113

1"#"<"6"months9.8 1158 1359 44

("4.5"#"19.4")""""""""""""""""("1099"#"1655")

1"#"<"6"months9.8 1158 1359 44

("4.5"#"19.4")""""""""""""""""("1099"#"1655")

6"months"#"<"1"year0.4 88.6 103.1 57

("0.2"#"0.7")""""""""""""""""""("85.5"#"121.6")

6"months"#"<"1"year0.4 88.6 103.1 57

("0.2"#"0.7")""""""""""""""""""("85.5"#"121.6")

1"#"<"19"years 0.8 14.0 34.8 252("0.7"#"0.9")""""""""""""""""""

("15.7"#"46.1") 1"#"<"19"years 0.8 14.0 34.8 252("0.7"#"0.9")""""""""""""""""""

("15.7"#"46.1")

Anti+Tg&&&&&&&&&&&&&&&&&&&(IU/mL,&kIU/L)

0"#"<"19""years 0.4 12.6 17.7 387.0 ("0.3"#"0.4")""""""""""""""""""("12.7#"32.4")

0"#"<"19""years 0.4 12.6 17.7 387.0 ("0.3"#"0.4")""""""""""""""""""("12.7#"32.4")

0"#"<"1"week3.4 22 24 49

("1.8"#"4.9")"""""""""""""""""""("21"#"25")

0"#"<"1"week3.4 22 24 49

("1.8"#"4.9")"""""""""""""""""""("21"#"25")

1"week"#"<"1"year4.9 29 33 100

("4.2"#"5.6")"""""""""""""""""""("29"#"35")

1"week"#"<"1"year4.9 29 33 100

("4.2"#"5.6")"""""""""""""""""""("29"#"35")

1"#"<"19"years3.9 18 21 252

("3.6"#"4.6")""""""""""""""""""""("18"#"23")

1"#"<"19"years3.9 18 21 252

("3.6"#"4.6")""""""""""""""""""""("18"#"23")

0"#"<"1"year<"2.0 59 64 122

""""""""""""""""""""""""""""""""""""""("53"#"88")

0"#"<"1"year<"2.0 59 64 122

""""""""""""""""""""""""""""""""""""""("53"#"88")

1"#"<"19"years <"2.0 28 41 251 ("29"#"69") 1"#"<"19"years <"2.0 28 41 251 ("29"#"69")

0"#"<"4"months2.4 19 22 55

("1.7"#"3.0")"""""""""""""""""""("19"#"24")

0"#"<"4"months2.4 19 22 55

("1.7"#"3.0")"""""""""""""""""""("19"#"24")

4"months"#"<"5"years7.7 32 33 190

("7.0"#"9.3")"""""""""""""""""""("32"#"38")

4"months"#"<"5"years7.7 32 33 190

("7.0"#"9.3")"""""""""""""""""""("32"#"38")

5"years"#"<"11"years 4.7 28 30 178("4.4"#"5.8")"""""""""""""""""""("28"#"36") 5"years"#"<"11"years 4.7 28 30 178

("4.4"#"5.8")"""""""""""""""""""("28"#"36")

11&+&<&19&years5.4 27 28 123

(&4.4&+&6.2&)&&&&&&&&&&&&&&&&&&&(&23&+&30&)

11&+&<&19&years5.9 32 39 127

(&5.0&+&6.9&)&&&&&&&&&&&&&&&&&&&(&30&+&41&)

0"#"<"1"wk8.1 55 62 44

("6.2"#"10.8")""""""""""""""""("53.8"#"70.6")

0"#"<"1"wk8.1 55 62 44

("6.2"#"10.8")""""""""""""""""("53.8"#"70.6")

1"wk"#"<"2"years <"0.5 3.8 4.7 136 ("3.7"#"8.3") 1"wk"#"<"2"years <"0.5 3.8 4.7 136 ("3.7"#"8.3")2"#"<"19"years <"0.5 2.3 2.6 516 ("2.5"#""2.8") 2"#"<"19"years <"0.5 2.3 2.6 516 ("2.5"#""2.8")0&+&<&12&years <&0.008 <&0.008 69 0&+&<&19&years <&0.008 0.032 0.097 302 (&0.044&+&&0.194&)

12&&&+&<&19&years <&0.008 0.239 0.279 122 (&0.223&+&0.552&)

0"#"<"1"week159 567 618 44

("137"#"182")""""""""""""""""("549"#"689")

0"#"<"1"week159 567 618 44

("137"#"182")""""""""""""""""("549"#"689")

1"week"#"<"6"months55.7 165 178 68

("51.1"#"61.1")"""""""""""""""("163"#"193")

1"week"#"<"6"months55.7 165 178 68

("51.1"#"61.1")"""""""""""""""("163"#"193")

6"months"#"<"2"years30.9 92.4 98.6 92

("27.00"#""34.9")""""""""""""("92.4"#"105")

6"months"#"<"2"years30.9 92.4 98.6 92

("27.00"#""34.9")""""""""""""("92.4"#"105")

2"#"<"10"years27.3 63.4 69.7 181

("25.7"#"28.7")"""""""""""""""("63.4"#"81.7")

2"#"<"10"years27.3 63.4 69.7 181

("25.7"#"28.7")"""""""""""""""("63.4"#"81.7")

10"#"<"19"years22.5 52.7 61.8 233

("21"#"24.4")""""""""""""""""""("53.0"#"75.4")

10"#"<"19"years22.5 52.7 61.8 233

("21"#"24.4")""""""""""""""""""("53.0"#"75.4")

0"#"<"1"week535 1749 1889 45

("446"#"622")""""""""""""""""("1710"#"2072")

0"#"<"1"week535 1749 1889 45

("446"#"622")""""""""""""""""("1710"#"2072")

1"week"#"<"6"months 57 732 817 70("39"#"79")"""""""""""""""""""""

("733"#"907") 1"week"#"<"6"months 57 732 817 70("39"#"79")"""""""""""""""""""""

("733"#"907")

6"months"#"<"1"year25 180 198 52

("18"#"31")"""""""""""""""""""""("178"#"219")

6"months"#"<"1"year25 180 198 52

("18"#"31")"""""""""""""""""""""("178"#"219")

1"#"<"12"years22 108 129 257

("21"#"24")"""""""""""""""""""""("106"#"134")

1"#"<"12"years22 108 129 257

("21"#"24")"""""""""""""""""""""("106"#"134")

12"#"<"19"years 17 74 83 183("14"#"20")"""""""""""""""""""""("73"#"107") 12"#"<"19"years 17 74 83 183

("14"#"20")"""""""""""""""""""""("73"#"107")

0"#"<"1"week >"70 >"70 44 0"#"<"1"week >"70 >"70 44

1"week"#"<"1"year0.6 12 17 130

("0.6"#"0.7")"""""""""""""""(11"#"19)

1"week"#"<"1"year0.6 12 17 130

("0.6"#"0.7")"""""""""""""""(11"#"19)

1"#"<"19"years 0.4 1.5 1.6 208("0.4","0.5")""""""""""""""""""""("1.5"#"1.6") 1"#"<"19"years 0.4 1.5 1.6 208

("0.4","0.5")""""""""""""""""""""("1.5"#"1.6")

0&+&<&1&week<&0.008 0.041 0.047 50 (0.039&+&0.055)

0&+&<&1&&week<&0.008 0.033 0.039 40 (0.029&+&0.051)

1&week&+&<&6&months <&0.008 0.032 0.038 47 (0.030&+&0.046) 1&week&+&<&1&year <&0.008 0.009 0.010 53 (0.006&+&0.016)6&months&+&<&12&years <&0.008 0.017 0.353 119 (0.287&+&0.454) 1+&<&19&years <&0.008 0.011 0.015 183 (0.011&+&0.020)

12&+&<&19&years <&0.008 0.494 0.566 82 (0.509&+&0.625)a."Bold"values"indicate"a"sex#specific"reference"interval

Male&Reference&Interval Female&Reference&Interval&

AFP&&&&&&&&&&&&&&&&&&&&&&&&&(ng/mL&or&μg/L)

Total&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&PSA&&&&&&&&&&&&&&&&&&&&&&&&&

(ng/mL)

SCC&&&&&&&&&&&&&&&&&&&&&&&&&&(ng/mL&or&μg/L)

ProGRP&&&&&&&&&&&&&&&&&&(pg/mL&or&ng/L)

HE4&&&&&&&&&&&&&&&&&&&&&&&&(pmol/L)

Free&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&PSA&&&&&&&&&&&&&&&&&&&&&&&&&

(ng/mL&or&μg/L)

CEA&&&&&&&&&&&&&&&&&&&&&&&&&(ng/mL&or&μg/L)

CA125&&&&&&&&&&&&&&&&&&&&&&(U/mL&or&kU/L)

CA19+9&&&&&&&&&&&&&&&&&&&&&(U/mL&or&kU/L)

CA15+3&&&&&&&&&&&&&&&&&&&&&(U/mL&or&kU/L)

44

A

B

Figure 2. Age-dependent scatter plots by sex of Anti-Thyroglobulin (Anti-Tg). (A) Regular

scatterplot (B) Scatterplot with outliers highlighted. ! indicates an outlier.

45

A

B

Figure 3. Age-dependent scatter plots by sex of Alpha-Fetoprotein (AFP). (A) Regular

scatterplot (B) Scatterplot with outliers highlighted. Figure insets show data using a more narrow

y-axis scale in order to visualize lower value test results more clearly. ! indicates an outlier.

46

A

B

Figure 4. Age-dependent scatter plots by sex of Cancer Antigen 19-9 (CA 19-9). (A) Regular



47

A

B

Figure 5. Age-dependent scatter plots by sex of Carcinoembryonic Antigen (CEA). (A) Regular



48

A

B

Figure 6. Age-dependent scatter plots by sex of Pro-Gastrin-Releasing Peptide (ProGRP). (A)

Regular scatterplot (B) Scatterplot with outliers highlighted. Figure insets show data using a

more narrow y-axis scale in order to visualize lower value test results more clearly. ! indicates

an outlier.

49

A

B

Figure 7. Age-dependent scatter plots by sex of Human Embryonic Protein 4 (HE4). (A) Regular



50

A

B

Figure 8. Age-dependent scatter plots by sex of Squamous Cell Carcinoma Antigen (SCC). (A)

Regular scatterplot (B) Scatterplot with outliers highlighted. Figure insets show data using a

more narrow y-axis scale in order to visualize lower value test results more clearly. ! indicates

an outlier.

51

A

B

Figure 9. Age-dependent scatter plots by sex of Cancer Antigen 15-3 (CA 15-3). (A) Regular


52

A

B

Figure 10. Age-dependent scatter plots by sex of Cancer Antigen 125 (CA 125). (A) Regular


53

A

B

Figure 11. Age-dependent scatter plots by sex of Cancer Antigen 125 (CA 125) and Cancer

Antigen 15-3 (CA 15-3) during the neonatal period. (A) CA 125, (B) CA 15-3.

54

A

B

Figure 12. Age-dependent scatter plots by sex of Free Prostate Specific Antigen (PSA). (A)

Regular scatterplot (B) Scatterplot with outliers highlighted. ! indicates an outlier.

55

A

B

Figure 13. Age-dependent scatter plots by sex of Total Prostate Specific Antigen (PSA). (A)

Regular scatterplot (B) Scatterplot with outliers highlighted. ! indicates an outlier.

56

Chapter 4 Discussion

4 Discussion

4.1 Key Observations

Cancer biomarkers have become an important tool in the treatment of cancer patients,

particularly in the monitoring of patients during and post-treatment (Table 1). However, there

remains a paucity of information regarding the application of these markers to pediatric cancers.

Specifically, little is known about how levels of these markers fluctuate in a healthy reference

population and, as previous studies have demonstrated (23, 24, 139), the usefulness of these

markers remains limited without rigorously established reference intervals that have been

properly stratified by key covariates like age and sex. By establishing age- and sex-specific

pediatric reference intervals for 11 key biomarkers, the present study has taken an important first

step towards harnessing the full potential of these markers in the pediatric population. As

expected, significant fluctuations of biomarker concentrations by age and/or gender were

observed in 10 of the 11 biomarkers investigated. Age partitioning was required for CA 15-3,

CA 125, CA 19-9, CEA, SCC, ProGRP, Total & Free PSA, HE4 and AFP, while gender

partitioning was also required for CA 125, Total and Free PSA (Table 6).

4.2 Typical Oncofetal Antigens

Oncofetal antigens, also occasionally referred to as carcinofetal antigens, are markers that are

expressed both during malignancy and neonatal development, but tend to be absent or present in

low concentrations in normal adult tissues. This link between cancer and early development is

not surprising, as many of the processes that contribute to cancer progression are also crucial to

early development, namely, cell proliferation and cell differentiation. In addition, there are

similarities between immune defense to the maternal immune system during pregnancy and to

the host immune system during malignancy (63).

57

Several common patterns of expression were observed between the various analytes examined.

Six of the oncofetal antigens examined (AFP, HE4, ProGRP, CEA, CA 19-9 and SCC) exhibited

the expected pattern of expression with high concentrations at birth followed by rapid decreases

to significantly lower levels (Figure 3 - 8). This pattern is in agreement with the fact that many

of these analytes are expressed during fetal development and/or play an important role in early

post-natal development (16, 67, 92).

It is important to note that the serum levels of various biomarkers may not necessarily be solely

reflective of the amount of marker that is produced by organs and/or tumors. Rather, it is

important to also consider the elements that may affect the transfer of analytes into the

circulation (63). For instance, certain analytes may have been found at higher concentrations in

the serum of neonates due to immature hepatic and/or renal function (53, 63). That is, if the liver

and/or kidney normally process these proteins, a lack of efficiency due to the fact that the liver

and/or kidney are not fully mature, may be an additional explanation for the high levels observed

in the serum of neonates.

The factors affecting the movement of markers from organs and other biological fluids into

circulation are important to consider, not only as a potential explanation for the results observed

here, but also, in the development of guidelines and interpretation of test results pertaining to

cancer biomarkers. Both pathological as well as physiological conditions can affect the release of

these markers into circulation. For example, studies have been conducted that examined the

relationship between menstrual cycle phase and tumor marker concentration in serum.

Currently, there is evidence that CA 125, PSA and CA15-3 levels in serum are significantly

influenced by menstrual cycle phase (61, 62, 152). Pregnancy is an additional physiological

condition that may alter the levels of marker found in circulation. During pregnancy, the passage

of oncofetal antigens from the embryo and placenta to maternal circulation is favored (63).

In addition to these physiological conditions, a number of other common pathophysiological

conditions like inflammation, hepatic and renal disease can also impact the amount of the marker

that enters circulation. There are also some studies suggesting that the type of tumor may be an

additional factor regulating how much of the marker in question is able to enter into circulation.

For instance, in the case of CA 125, it has been hypothesized that benign ovarian tumors, which

tend to have intact basement membranes, provide a more effective barrier and thus prevent this

58

marker from entering the circulatory system. In contrast, in malignant tumors, which tend to

have disrupted basement membranes, CA 125 can enter the circulation more readily (153). It is

important that the numerous mitigating factors influencing marker entry into circulation be

considered by physicians when interpreting results and ordering tests. It is also important that

guidelines for tumor marker use address these factors as potential explanations for test results

outside of reference ranges or significant changes in serial measurements.

4.3 Atypical Oncofetal Antigens

Surprisingly, two oncofetal antigens, CA 125 and CA 15-3, did not show the characteristic

pattern of expression observed with the other oncofetal antigens (Figure 9, 10, 11). Both CA 15-

3 and CA 125 had slightly lower concentrations in the early days and months of life,

respectively. Given that both these markers are known to be expressed in fetal and neonatal

tissue, this pattern of expression is unexpected, however, in the case of CA 125, it has been

shown that levels in both maternal serum and amniotic fluid are higher in the early stages of

pregnancy and tend to attenuate by birth (63). This is in contrast with the findings observed for

other oncofetal antigens like SCC, which showed steady increases in concentration in both

maternal serum and amniotic fluid as pregnancy progressed (63). It is hypothesized that the

presence of the developing fetus suppresses the release of CA 125 into the maternal circulation,

potentially because the growth of the gestational sac obstructs the usual drainage route of CA

125 from the gestational sac into the peritoneal cavity and maternal circulation (153, 154). It is,

therefore, possible that CA 125 may play a role in fetal development at the early stages of

pregnancy and may again become important to neonatal development after the first week of life.

A study examining the levels of CA 125 in healthy infants up to 1.5 years of age showed a

similar pattern to the one observed here, with low levels at birth followed by increases in CA 125

concentration in the 0.1 – 0.5 year age range. This spike was followed by decrease in CA 125

concentration in the 0.5 – 1.5 year age range (53).

In contrast to the fluctuating levels of CA 125 during pregnancy, the concentration of CA 15-3 in

maternal serum and amniotic fluid did not differ significantly over the gestational course (63). In

addition to this observation, other studies have questioned that status of CA 15-3 as an oncofetal

antigen. While some studies have found differences in CA 15-3 concentrations between

pregnant and non-pregnant women, others have found no significant differences between the two

59

groups (155, 156). My results showed that CA 15-3 concentration was low at birth but spiked

later on during the neonatal period, which may further contribute to the controversy surrounding

the status of CA 15-3 as an oncofetal antigen. To my knowledge, this is the first study to

examine CA 15-3 levels in a neonatal population, thus, additional investigations are required to

confirm or refute the status of CA 15-3 as an oncofetal antigen, and to elucidate the reasons for

the unusual pattern of expression observed during the neonatal period.

4.4 Sex Differences

Three analytes demonstrated gender differences according to our analysis: Free PSA, Total PSA,

and CA 125 (Figure 10, 12, 13). Specifically, both Free and Total PSA demonstrated higher

levels in males than in females after puberty. This is not surprising, as PSA levels are known to

be regulated by IGF-1 and Testosterone, both of which also increase during puberty.

Additionally, previous studies have demonstrated that PSA levels increase with Tanner stage

(157), lending further support to the hypothesis that changes linked to puberty cause an increase

in PSA levels in young males. Despite the fact that PSA levels are higher in males than in

females after puberty, it is also interesting to note that a number of females in the study

population had PSA levels that were comparable to the levels found in males of a similar age.

This highlights the important fact that PSA is not specific to the prostate as it’s name might

suggest. Rather, PSA has been discovered in the endometrium, brain and breast tissue (103,

104). It is also possible that these participants have a common metabolic syndrome known as

Polycystic Ovarian Syndrome which is associated with elevated levels of androgens(119).

CA 125 also showed gender differences in the teenage years, as females 11 - < 19 years had

higher concentrations of CA 125 than males. This increase in CA 125 in the female population

may be due to female-specific physiological processes, including development of the female

reproductive system at puberty, phase of menstrual cycle and pregnancy (158). The importance

of these influencing factors has been previously noted in the adult population. In a previously

discussed study, Gutierrez et al. observed that female hormones influence CA 125 levels,

therefore, they concluded that there should be differences in reference values for CA 125

between males and females. For this reason, their group established sex-specific reference

intervals for CA 125. In addition, significant differences were discovered between post-

menopausal and fertile women and, therefore, female reference intervals were also stratified by

60

menopausal status (138). However, to my knowledge, this research is the first study to uncover a

sex difference in CA 125 levels in the pediatric population and, thus, further work is required to

elucidate the reasons for these observed sex-differences.

In the future, for markers like CA 125 and PSA that required sex-specific partitions, it will be

important to examine the relationship between these markers and Tanner stage as well as

menstrual status. Tanner staging assesses the level of pubertal development of an individual

(143). The CALIPER project has collected Tanner staging information for a subset of the

participants in our biobank, using a self-report questionnaire. The self-report Tanner instrument

includes pictures of Tanner’s stages of development. For females, images of breast development

and pubic hair growth are included, whereas for males images of pubic hair growth and testicular

size are included. The questionnaire asks the participant to indicate which of the images in each

specific category is the closest to the way their own body looks. This self-report method was

validated by comparing self-report staging to physician staging (143). A study examining the

relationship between Tanner stage and marker concentration could help determine if there is a

link between changes in biomarker levels and pubertal development. This would be helpful in

determining potential reasons for the differences observed in CA 125 and PSA levels between

males and females. For example, if increasing Tanner stage correlates with increasing marker

concentration this could indicate a potential link between the marker and the changes in hormone

levels that occur during puberty. In this particular study, the sample size of individuals with

Tanner information was not sufficient to complete analysis on the effect of Tanner stage on

cancer marker concentration; therefore, it is a potential question for future research.

In addition to examining potential correlations between marker concentration and Tanner stage,

it is also important to assess the correlation between marker concentration and menstrual status.

As noted earlier, previous studies that examined reference values in the adult population have

provided evidence that both PSA and CA 125 levels fluctuate with menstrual cycle phase (61,

62, 118, 152, 159). Exploring this relationship in the pediatric population would also help

elucidate the reasons why sex differences were observed for these two markers. For example, it

may be possible that spikes in CA 125 concentrations occur when females are in a specific

menstrual cycle stage and this may contribute to the higher cut-off observed in females aged 11 -

< 19 years compared to males 11 - < 19 years. Similar to the problems mentioned previously

with obtaining sufficient Tanner staging information, in this particular study, the sample size of

61

individuals with menstrual cycle data was not sufficient to complete analysis on the effect of

menstrual cycle phase on cancer marker concentration; therefore, it is a potential question for

future research.

4.5 Comparison to previously established pediatric reference intervals

4.5.1 Key Observations

Overall, comparisons of this research with previously established pediatric reference intervals for

these analytes revealed similar trends in expression across the pediatric age range. However,

dramatic differences in reference interval values were observed for many of the analytes

examined here. There are a variety of potential explanations for these discrepancies. Firstly,

many of the studies outlined below did not use a healthy community sample but, instead, a

specific subset of hospitalized patients. A recent study published by CALIPER questioned the

use of hospitalized patients to establish reference ranges (160), and suggested that reference

intervals established from patient subsets likely do not replace the need for rigorously established

reference intervals based on healthy community samples.

In addition, discrepancies may be due to the fact that for several of the studies cited here,

different analytical methods were used for sample testing. Many groups have noted the

importance of establishing platform-specific reference intervals, and this issue is discussed

further in Section 4.8 – Future Directions.

4.5.2 Typical Oncofetal Antigens

In relation to the patterns observed for the typical oncofetal antigens (CA 19-9, AFP, ProGRP,

SCC, CEA and HE4), similar trends to the ones observed here have been previously

demonstrated for levels of AFP, ProGRP and CA 19-9 across the pediatric age range (3, 53, 92).

Reference intervals were also previously established for both AFP as well as CA 19-9. To my

knowledge, no studies have examined levels of HE4, SCC or CEA in a pediatric population.

AFP reference intervals were established on an earlier generation of AFP Abbott Architect assay

by CALIPER in 2013. The reference intervals established in the current study had, overall,

higher upper limits than those established in the previous CALIPER study. It should be noted

62

that this may be due to differences in the two assays, specifically, this may relate to a higher limit

of detection in the AFP assay used in the present study.

Regarding CA 19-9, the upper limit of the 0 - < 1 year reference interval established in the

current study (64.02 U/mL) was vastly different compared to the upper limit of the 95th

percentile confidence interval (22 U/L) established by Lahdenne et al. for the 0.1 – 0.5 year age

range (53). This may be due, in part, to the fact that Lahdenne et al. used a radioimmunoassay

(RIA) method, while the present study used a chemiluminescence immunoassay (53). In

addition, participants used in the previous study were infants waiting for minor surgical

operations and, therefore, may not reflect the values observed in a healthy population as closely

as the results produced here, which were established using a healthy community sample. Finally,

it should be noted that the previous study established reference intervals for two unique partitions

in the 0.1 – 0.5 year age range using only 39 samples. This sample size is significantly smaller

than what is recommended by the CLSI A 28-3 guidelines.

4.5.3 Atypical Oncofetal Antigens

Similar patterns to the ones observed here have been previously demonstrated for CA 125. In

addition, reference intervals were previously established for the birth – 1.5 year age range (53).

Again, although the general patterns of expression were similar, the established reference values

differed greatly from those of Lahdenne et al. - while the upper limit of the 0 - < 4 month

reference interval established in the current study was 21.5 U/mL, the upper limit of the 95th

percentile confidence interval established by Lahdenne et al. for the 0.1 – 0.5 year age range was

45 U/L (53).

As previously mentioned, in addition to the differences in analytical methods used, the results of

the Lahdenne et al. study may have differed from those observed here due to the fact that the

former study used infants awaiting minor surgical operations rather than healthy infants, and

established age-stratified reference intervals based on a very small sample population (only 39

samples) (53).

4.5.4 Analytes Requiring Sex-Specific Partitions

To my knowledge, this is the first study to illustrate sex-differences in the teenage age range for

CA 125. However, several studies have previously demonstrated differences in PSA levels

63

between males and females in the pediatric population. Notably, Randell et al. (122) had

previously established reference intervals for Total PSA in males and females aged birth – 18

years. Since this study established 7 different age partitions for both males and females it is

difficult to compare values, however, similar patterns of expression can be observed for males

and females in both Randell et al.’s study and the current study.

4.5.5 Anti-Thyroglobulin

Anti-Tg was the only marker in the present study that did not require any age or sex partitioning.

This contradicts a recent report by Taubner et al., who observed gender differences with higher

levels of Anti-Tg in females aged 6 - < 20 years (35). This is in line with previous findings that

thyroid dysfunction and autoimmunity is more common in adult females and this study by

Taubner et al. may illustrate the beginning of this trend in the female population (161).

There are a number of potential reasons for why our current findings differ from those of

Taubner et al. This discrepancy may be due to the fact that the previous study was carried out

using a different immunoassay on the Roche Modular System. The poor concordance between

Anti-Tg assays has been previously noted (31) and highlights the importance of establishing

reference intervals specific to each platform. In addition, it should be noted that approximately

83% of the reference population used in the Taubner et al. study consisted of hospital and clinic

patients(35).

4.6 Comparison to previously established adult reference intervals

It is important to note that key differences were observed between the pediatric reference

intervals established here and recommended cut-offs in the adult population. For a summary of

adult reference limits reported by Abbott Diagnostics™ and commonly used decision limits see

Table 7.

Both Anti-Tg and Total PSA remained well below the recommended adult cut offs of 22 IU/mL

and 2 – 4 ng/mL throughout the entire pediatric age range (35, 40).

Not surprisingly, four of the oncofetal antigens examined here (SCC, HE4, CEA and AFP) fell

above the recommended adult cut-offs during the neonatal period, then subsequently dropped to

64

below or approximately at the adult cut-off (141, 158, 162, 163). On the other hand, two of the

oncofetal antigens examined (ProGRP and CA 19-9) were persistently higher than the

recommended adult cut-offs (158, 164).

At birth, the upper limit of the reference intervals for both CA 15-3 and CA 125 (23.6 U/mL and

21.5 U/mL) fell below the recommended adult cut-offs of 25 U/mL and 35 U/mL, respectively

(141, 158). However, while the CA 15-3 reference interval returned to levels below the adult

cut-off at around 1 year of age after reaching a peak during the 1 week - < 1 year age range,

levels of CA 125 in teenage females were higher than the adult cut-off of 35 U/mL with an upper

limit of 39.1 U/mL. These observed differences highlight the importance of establishing

reference intervals specific to the pediatric population.

65

Table 7. Adult Reference Limits and Decision Limits

Analyte Abbott Reported Percentile^ Decision Limit

AFP (ng/mL) 8.8* 200 (141)

Anti-Tg (IU/mL) 4.11 Δ 22 (35)

CA 15-3 (U/mL) 31 " 25 (158)

CA 19-9 (U/mL) 37* 37 (158)

CA 125 (U/mL) 35 35 (141)

CEA (ng/mL) 5* 2.5 (158)

Free PSA (ng/mL) 0.5 + -

HE4 (pmol/L) 70* 70 (162)

ProGRP (pg/mL) 63* 50 (102)

SCC (ng/mL) 1.5* 1.5 (163)

Total PSA (ng/mL) 4* 2 – 4 (40)

^ From Abbott Architect ci4100 package insert

* 95th percentile " 99th percentile Δ 97.5th percentile + 87th percentile

66

4.7 Study Limitations

4.7.1 Single Platform Study

There are certain limitations to the present study that are important to discuss. Firstly, a

biomarker by definition must be able to be measured in a minimally invasive, cost-effective and

efficient manner(12). Therefore, this study was limited to assays that are available on automated

large-scale platforms and, furthermore, since this was an Abbott Diagnostics™ study we were

further limited to assays available on the Abbott Architect ci4100. It is important to note that the

reference intervals established in this study are specific to this platform and that in order for

health centers not using this platform to benefit from this information, a transference study must

be carried out. This concept is discussed further in Section 4.8.3.

4.7.2 Storage Effects

Many of the samples used in this analysis have been stored at - 80°C for several years. The

CALIPER project has previously examined the stability of 57 biochemical markers also stored at

- 80°C over a 10 – 13 month period and discovered that all analytes with the exception of

parathyroid hormone (PTH) were stable (165). However, with the exception of AFP, none of the

cancer markers examined in the present study were included in this stability study.

A limited number of studies have examined the issue of analyte stability in relation to cancer

biomarkers. Specifically, ProGRP was found to be stable at – 30 °C for up to 4 weeks, the

longest time period examined (166). An additional study examined the long-term stability of

Total PSA in - 70 °C storage over a two year period. It was discovered that 100% of samples

stored at this temperature showed no statistically significant difference in measurement after 2

years (167). An additional study examined both Free and Total PSA levels in serum stored at -

70 °C. Both a regression and a Bland-Altman analysis lead the authors to conclude that Free and

Total PSA can be stored and used up to 5 years after collection (168). Other studies on the

stability of cancer biomarkers typically did not assess long-term stability at temperatures

comparable to the - 80°C at which CALIPER samples are stored but, rather, examined short term

stability at room temperature, or the impact of multiple freeze-thaw cycles; neither of which are

applicable to the current study (169, 170).

67

4.7.3 Confidence Interval Size

Upon calculation of the reference intervals, it was observed that, primarily for the upper limits of

certain partitions, the confidence intervals were unusually wide. Often, wide confidence

intervals can be attributed to inadequate sample size, however, in the case of this study, wide

confidence intervals were observed even in partitions with large sample sizes (2). I attribute this

observation to the highly skewed nature of the data. Specifically, in most cases where

confidence interval size was an issue, it was observed that the data was highly skewed towards

the lower concentration end of the distribution. In contrast, there were fewer data points in the

higher concentration end of the distribution and, furthermore, there tended to be a high degree of

variability between the data points at higher concentrations. Box-Cox transformations were

carried out as discussed in Section 2.3 to try to normalize the data and reduce the size of the

confidence intervals, however, the confidence intervals for many of the partitions remained

wider than recommended by the CLSI C28-A3 guidelines, which state that confidence intervals

should be less than 0.2 times the wide of a 95% reference interval (2). It is possible that this issue

relates to a high degree of between-subject biological variation, and the importance of exploring

biological variation parameters for these markers in future studies is discussed further in Section

4.8.

4.7.4 Neonatal Sample Size Restraints

Finally, due to the difficulty in collecting a large sample volume of blood from the neonatal

population, I was limited in the amount of neonatal results that could be obtained for Free PSA,

the assay that required the largest sample volume. As such, my ability to partition data in the

first year of life was limited despite the fact that the scatterplot seemed to suggest the need for a

0 - < 1 week age partition.

4.8 Future Directions

4.8.1 Establishing Reference Intervals for Additional Cancer Biomarkers

In the future, it will be important to continue working to establish reference intervals for

additional cancer biomarkers that may be relevant to the pediatric population. In particular, it

will be important to begin recruitment efforts to collect both plasma and urine samples from

68

healthy children in order to examine cancer biomarkers that are detected in these fluids rather

than in serum.

The establishment of reference intervals for hematological parameters such as reticulocyte,

leukocyte and lymphocyte counts will be important specifically for leukemia patients. Recently,

Aldrimer et al., established pediatric reference intervals for a variety of such hematological

parameters, however, this study was conducted in a largely Caucasian Swedish population (171).

In addition, the authors used a non-parametric rank method on partitions with a sample size of

less than 120, which is not recommended by the CLSI C28-A3 guidelines (2).

The establishment of reference intervals for urinary tumor markers such as HVA (homovanillic

acid) and VMA (Vanillylmandelic acid) would be of use for pediatric neuroblastoma patients.

Pediatric reference intervals for these markers were recently established, however, the study was

carried out using a hospital population (172).

4.8.2 Examining Ethnic Differences

Due to limitations owing to sample size, I did not have the statistical power necessary to examine

the impact of ethnicity on the concentrations of the 11 cancer biomarkers examined in this study.

In the future, it will be important to examine the impact of this key covariate, as previous studies,

discussed below, have provided some preliminary evidence that ethnicity may have an impact on

reference values for these markers.

Other groups have begun to investigate differences in marker levels between ethnic groups and

highlight the importance of examining this issue further by demonstrating the need for ethnicity-

specific reference intervals. AFP is one of the markers that has been studied in this context and

important differences in AFP levels between ethnic groups have been observed. One study

examined the levels of AFP in maternal serum of approximately 9000 pregnant women. It was

discovered that black women showed consistently higher values for AFP, and the study authors

recommended that a correction for race should be applied when interpreting AFP values (173). A

second study, which also examined AFP levels in the maternal serum of pregnant women, found

that AFP levels were generally higher in Asian and black women compared to Hispanic and

white women. This group also recommended use of ethnicity-specific interpretation of AFP

measurements (174). Studies have also been conducted in male populations and, again, similar

69

results were discovered. African men had significantly higher levels of AFP compared to

populations sampled from Singapore and Lyon (175). Although the latter study seems to muddle

the concepts of ethnic background and nationality, by categorizing based on the country in which

the participant resides rather than ethnic origin, they both provide evidence that there may be

important differences in AFP between various ethnic populations and this may have important

consequences for the interpretation of AFP values in patients.

Other markers have not been examined as closely as AFP, however, some interesting findings

have emerged that also point to the importance of ethnicity in the interpretation of biomarker

levels. For example, one study examined levels of HE4 in a healthy, Asian, female population.

The study revealed that levels of HE4 in the Indian population were higher than in the Malaysian

population. On the other hand, no differences were noted in HE4 levels between the Malaysian

and Chinese population (176). Significant differences between ethnic groups were also observed

in a study that examined CA 125 levels in a healthy post-menopausal population. Specifically,

CA 125 levels in the African and Asian populations were lower than those in the Caucasian

population (177).

To my knowledge, only one study examined ethnic differences in CEA measurements, with a

fairly small sample size. It was discovered that healthy, black, non-smoking men had higher

levels of CEA compared to healthy, white non-smoking men. These differences were not found

in the female population (178).

Important findings have also been published regarding the impact of ethnicity on PSA levels.

Specifically, men of African descent had higher serum PSA levels than Caucasian men (179). In

addition, healthy men of African descent showed more rapid increases in PSA over time

compared with white men (180). This is an important consideration for prostate cancer screening

programs.

To my knowledge, no studies have been conducted to examine whether or not ethnic differences

exist in levels of SCC, ProGRP or Anti-Tg. However, the aforementioned studies provide solid

evidence that highlights the importance of examining these types of questions in all cancer

markers and, potentially, developing ethnicity-specific reference intervals for the markers

examined here.

70

4.8.3 Transferring Reference Intervals to Additional Platforms

In addition to the establishment of reference intervals for other cancer biomarkers, it will be

important to carry out a transference study in order to establish reference intervals for the 11

markers examined here on additional analytical platforms (7). The United Kingdom National

External Quality Assessment (EQA) Service has reported between-method coefficients of

variation (CVs) in excess of 20% for some tumor markers (181). These differences could be due

to variations in methods or in antibody specificity, among other factors (182). The findings in

this EQA report highlight the risk involved with using reference values that are not established

specifically for the platform used in testing.

CALIPER previously carried out a transference study to expand the scope of pediatric reference

intervals for 40 biochemical markers initially established on the Abbott ARCHITECT to four

other major platforms used in children’s hospitals across Canada (Beckman Coulter DxC800,

Ortho Vitros 5600, Roche Cobas 6000 and Siemens Vista 1500) (7). This type of study involved

two main steps: transference and validation. Transference involved testing 200 hospital patient

samples on the new analytical systems. If the assays were highly correlated, the mathematical

relationship between the two analytical systems was used to calculate new reference intervals for

the platform(s) in question. The validation steps involved testing 100 healthy CALIPER

participant samples on all platforms in question and calculating the percentage of individuals that

fell within the newly calculated reference intervals for each specific platform. The validation was

considered successful if ≥ 90% of the values fell within the reference interval (7). A similar

approach should be used in the future to transfer the reference intervals for the 11 biomarkers

examined in this study to other analytical platforms, in order to ensure that accurate reference

intervals are available for each of the various platforms used in children’s hospitals across

Canada.

4.8.4 Establishing Biological Variation Parameters

Finally, as previously noted, many of the markers examined in this study may be important for

monitoring cancer patients, to aid in either the assessment of treatment progress/effectiveness, or

to monitor patients in remission for potential recurrences (Table 1). If these markers are to be

used in this context, it will be important to establish key biological variation parameters in order

to make informed choices about what constitutes a significant change between serial

71

measurements (183). This is necessary because changes in serial measurements can be caused by

a variety of factors that may not relate to clinically significant changes in a patient’s condition.

Specifically, changes in serial measurements could be caused by factors like preanalytical

variation, analytical variation or within-subject biological variation (183).

Establishing these biological variation parameters will provide the information necessary to

calculate reference change values (RCV), which indicate the necessary change in concentration

between two measurements that is likely to signal a clinically significant change (183). In

addition, biological variation can be used to establish quality specifications such as bias,

precision, and total allowable error (184). Previous studies have examined biological variation

parameters in adults for CEA, AFP and CA 19-9, and suggest that biological variation is an

important consideration when monitoring cancer patients (183, 185). Biological variation

parameters for these markers, as well as for the other 9 cancer biomarkers examined here, will

need to be established in a pediatric population as previous studies have noted that certain

differences in biological variation do exist between the pediatric and adult populations (8).

4.9 Conclusions

The establishment of pediatric reference intervals for tumor biomarkers will aid in harnessing the

true predictive, diagnostic, and monitorial potential of tumor markers in a pediatric population.

These reference intervals will not only be useful in practice by contributing to a more accurate

assessment of tumor markers in cancer diagnosis and treatment, but also in research aimed at

determining the prognostic and/or diagnostic value of tumor marker in various types of cancers

and populations.

72

References

1. Horowitz GL. Estimating reference intervals. American journal of clinical pathology 2010;133:175-7.

2. CLSI. Defining, establishing, and verifying reference intervals in the clinical laboratory; approved guideline—third edition. CLSI document C28-A3, Vol. Wayne (PA).

3. Bailey D, Colantonio D, Kyriakopoulou L, Cohen AH, Chan MK, Armbruster D, Adeli K. Marked biological variance in endocrine and biochemical markers in childhood: Establishment of pediatric reference intervals using healthy community children from the caliper cohort. Clinical chemistry 2013;59:1393-405.

4. Colantonio DA, Kyriakopoulou L, Chan MK, Daly CH, Brinc D, Venner AA, et al. Closing the gaps in pediatric laboratory reference intervals: A caliper database of 40 biochemical markers in a healthy and multiethnic population of children. Clinical chemistry 2012;58:854-68.

5. Konforte D, Shea JL, Kyriakopoulou L, Colantonio D, Cohen AH, Shaw J, et al. Complex biological pattern of fertility hormones in children and adolescents: A study of healthy children from the caliper cohort and establishment of pediatric reference intervals. Clinical chemistry 2013;59:1215-27.

6. Daly CH, Liu X, Grey VL, Hamid JS. A systematic review of statistical methods used in constructing pediatric reference intervals. Clinical biochemistry 2013;46:1220-7.

7. Estey MP, Cohen AH, Colantonio DA, Chan MK, Marvasti TB, Randell E, et al. Clsi-based transference of the caliper database of pediatric reference intervals from abbott to beckman, ortho, roche and siemens clinical chemistry assays: Direct validation using reference samples from the caliper cohort. Clinical biochemistry 2013;46:1197-219.

8. Bailey D, Bevilacqua V, Colantonio DA, Pasic MD, Perumal N, Chan MK, Adeli K. Pediatric within-day biological variation and quality specifications for 38 biochemical markers in the caliper cohort. Clinical chemistry 2014;60:518-29.

9. Raizman JE, Cohen AH, Teodoro-Morrison T, Wan B, Khun-Chen M, Wilkenson C, et al. Pediatric reference value distributions for vitamins a and e in the caliper cohort and establishment of age-stratified reference intervals. Clinical biochemistry 2014;47:812-5.

10. Rosychuk RJ, Witol A, Stobart K. Childhood cancer trends in a western canadian province: A population-based 22-year retrospective study. Pediatric blood & cancer 2010;55:1348-55.

11. Cancer in children. Sep 20, 2012. Retrieved from: http://www.cancer.org/cancer/cancerinchildren/index [Accessed: Feb 11, 2013]

73

12. Biomarkers Definitions Working G. Biomarkers and surrogate endpoints: Preferred definitions and conceptual framework. Clinical pharmacology and therapeutics 2001;69:89-95.

13. Diamandis EP. Cancer biomarkers: Can we turn recent failures into success? Journal of the National Cancer Institute 2010;102:1462-7.

14. Sahab ZJ, Semaan SM, Sang QX. Methodology and applications of disease biomarker identification in human serum. Biomarker insights 2007;2:21-43.

15. Konforte D, Diamandis EP. Is early detection of cancer with circulating biomarkers feasible? Clinical chemistry 2013;59:35-7.

16. Bader D, Riskin A, Vafsi O, Tamir A, Peskin B, Israel N, et al. Alpha-fetoprotein in the early neonatal period--a large study and review of the literature. Clinica chimica acta; international journal of clinical chemistry 2004;349:15-23.

17. Puck TT, Waldren CA, Jones C. Mammalian cell growth proteins. I. Growth stimulation of fetuin. Proceedings of the National Academy of Sciences of the United States of America 1968;59:192-9.

18. Aoyagi Y, Ikenaka T, Ichida F. Alpha-fetoprotein as a carrier protein in plasma and its bilirubin-binding ability. Cancer research 1979;39:3571-4.

19. Wajner M, Papiha SS, Wagstaff TI. Response of human peripheral blood lymphocytes in the presence of cord sera: Relationship of lymphocyte transformation with number of pregnancies and levels of alpha-fetoprotein. Clinical and experimental immunology 1983;52:381-6.

20. Isaacs H, Jr. Fetal and neonatal hepatic tumors. Journal of pediatric surgery 2007;42:1797-803.

21. von Schweinitz D, Hecker H, Schmidt-von-Arndt G, Harms D. Prognostic factors and staging systems in childhood hepatoblastoma. International journal of cancer Journal international du cancer 1997;74:593-9.

22. Labdenne P, Heikinheimo M. Clinical use of tumor markers in childhood malignancies. Annals of medicine 2002;34:316-23.

23. Baranzelli MC, Kramar A, Bouffet E, Quintana E, Rubie H, Edan C, Patte C. Prognostic factors in children with localized malignant nonseminomatous germ cell tumors. Journal of clinical oncology : official journal of the American Society of Clinical Oncology 1999;17:1212.

24. Pauniaho SL, Tatti O, Lahdenne P, Lindahl H, Pakarinen M, Rintala R, Heikinheimo M. Tumor markers afp, ca 125, and ca 19-9 in the long-term follow-up of sacrococcygeal teratomas in infancy and childhood. Tumour biology : the journal of the International Society for Oncodevelopmental Biology and Medicine 2010;31:261-5.

74

25. Ishak KG. Primary hepatic tumors in childhood. Progress in liver diseases 1976;5:636-67.

26. Bergstrand CG. Alphafetoprotein in paediatrics. Acta paediatrica Scandinavica 1986;75:1-9.

27. Blair JI, Carachi R, Gupta R, Sim FG, McAllister EJ, Weston R. Plasma alpha fetoprotein reference ranges in infancy: Effect of prematurity. Archives of disease in childhood 1987;62:362-9.

28. Lin JD. Thyroglobulin and human thyroid cancer. Clinica chimica acta; international journal of clinical chemistry 2008;388:15-21.

29. Li L, Paul S, Tyutyulkova S, Kazatchkine MD, Kaveri S. Catalytic activity of anti-thyroglobulin antibodies. Journal of immunology 1995;154:3328-32.

30. Ringel MD, Nabhan F. Approach to follow-up of the patient with differentiated thyroid cancer and positive anti-thyroglobulin antibodies. The Journal of clinical endocrinology and metabolism 2013;98:3104-10.

31. Pickett AJ, Jones M, Evans C. Causes of discordance between thyroglobulin antibody assays. Annals of clinical biochemistry 2012;49:463-7.

32. Spencer CA. Clinical review: Clinical utility of thyroglobulin antibody (tgab) measurements for patients with differentiated thyroid cancers (dtc). The Journal of clinical endocrinology and metabolism 2011;96:3615-27.

33. Cho EM, Kim UH, Choi BH, Ko CW. Changes of antithroglobulin antibody in children with congenital hypothyroidism. Annals of pediatric endocrinology & metabolism 2013;18:179-82.

34. Bitterman O, Bongiovanni M, Giuliani C, Roma G, Toscano V, Napoli A. Anti thyroperoxidase and anti thyroglobulin antibodies in diabetic pregnancies. Journal of endocrinological investigation 2014.

35. Taubner K, Schubert G, Pulzer F, Pfaeffle R, Korner A, Dietz A, et al. Serum concentrations of anti-thyroid peroxidase and anti-thyroglobulin antibodies in children and adolescents without apparent thyroid disorders. Clinical biochemistry 2014;47:3-7.

36. Spicer AP, Rowse GJ, Lidner TK, Gendler SJ. Delayed mammary tumor progression in muc-1 null mice. The Journal of biological chemistry 1995;270:30093-101.

37. Burchell J, Wang D, Taylor-Papadimitriou J. Detection of the tumour-associated antigens recognized by the monoclonal antibodies hmfg-1 and 2 in serum from patients with breast cancer. International journal of cancer Journal international du cancer 1984;34:763-8.

38. Simmons AR, Baggerly K, Bast RC, Jr. The emerging role of he4 in the evaluation of epithelial ovarian and endometrial carcinomas. Oncology 2013;27:548-56.

75

39. Iwahori K, Suzuki H, Kishi Y, Fujii Y, Uehara R, Okamoto N, et al. Serum he4 as a diagnostic and prognostic marker for lung cancer. Tumour biology : the journal of the International Society for Oncodevelopmental Biology and Medicine 2012;33:1141-9.

40. Sturgeon CM, Duffy MJ, Stenman UH, Lilja H, Brunner N, Chan DW, et al. National academy of clinical biochemistry laboratory medicine practice guidelines for use of tumor markers in testicular, prostate, colorectal, breast, and ovarian cancers. Clinical chemistry 2008;54:e11-79.

41. Molina R, Barak V, van Dalen A, Duffy MJ, Einarsson R, Gion M, et al. Tumor markers in breast cancer- european group on tumor markers recommendations. Tumour biology : the journal of the International Society for Oncodevelopmental Biology and Medicine 2005;26:281-93.

42. Duffy MJ, Evoy D, McDermott EW. Ca 15-3: Uses and limitation as a biomarker for breast cancer. Clinica chimica acta; international journal of clinical chemistry 2010;411:1869-74.

43. Hayes DF, Zurawski VR, Jr., Kufe DW. Comparison of circulating ca15-3 and carcinoembryonic antigen levels in patients with breast cancer. Journal of clinical oncology : official journal of the American Society of Clinical Oncology 1986;4:1542-50.

44. Colomer R, Ruibal A, Genolla J, Rubio D, Del Campo JM, Bodi R, Salvador L. Circulating ca 15-3 levels in the postsurgical follow-up of breast cancer patients and in non-malignant diseases. Breast cancer research and treatment 1989;13:123-33.

45. Hashimoto T, Matsubara F. Changes in the tumor marker concentration in female patients with hyper-, eu-, and hypothyroidism. Endocrinologia japonica 1989;36:873-9.

46. Symeonidis A, Kouraklis-Symeonidis A, Apostolopoulos D, Arvanitopoulou E, Giannakoulas N, Vassilakos P, Zoumbos N. Increased serum ca-15.3 levels in patients with megaloblastic anemia due to vitamin b12 deficiency. Oncology 2004;67:359-67.

47. Steinberg W. The clinical utility of the ca 19-9 tumor-associated antigen. The American journal of gastroenterology 1990;85:350-5.

48. Atkinson BF, Ernst CS, Herlyn M, Steplewski Z, Sears HF, Koprowski H. Gastrointestinal cancer-associated antigen in immunoperoxidase assay. Cancer research 1982;42:4820-3.

49. Arends JW, Verstynen C, Bosman FT, Hilgers J, Steplewski Z. Distribution of monoclonal antibody-defined monosialoganglioside in normal and cancerous human tissues: An immunoperoxidase study. Hybridoma 1983;2:219-29.

50. Kannagi R. Carbohydrate antigen sialyl lewis a--its pathophysiological significance and induction mechanism in cancer progression. Chang Gung medical journal 2007;30:189-209.

76

51. Goonetilleke KS, Siriwardena AK. Systematic review of carbohydrate antigen (ca 19-9) as a biochemical marker in the diagnosis of pancreatic cancer. European journal of surgical oncology : the journal of the European Society of Surgical Oncology and the British Association of Surgical Oncology 2007;33:266-70.

52. Herlyn M, Steplewski Z, Herlyn D, Koprowski H. Colorectal carcinoma-specific antigen: Detection by means of monoclonal antibodies. Proceedings of the National Academy of Sciences of the United States of America 1979;76:1438-42.

53. Lahdenne P, Pitkanen S, Rajantie J, Kuusela P, Siimes MA, Lanning M, Heikinheimo M. Tumor markers ca 125 and ca 19-9 in cord blood and during infancy: Developmental changes and use in pediatric germ cell tumors. Pediatric research 1995;38:797-801.

54. Bast RC, Jr., Spriggs DR. More than a biomarker: Ca125 may contribute to ovarian cancer pathogenesis. Gynecologic oncology 2011;121:429-30.

55. Akita K, Tanaka M, Tanida S, Mori Y, Toda M, Nakada H. Ca125/muc16 interacts with src family kinases, and over-expression of its c-terminal fragment in human epithelial cancer cells reduces cell-cell adhesion. European journal of cell biology 2013;92:257-63.

56. Rump A, Morikawa Y, Tanaka M, Minami S, Umesaki N, Takeuchi M, Miyajima A. Binding of ovarian cancer antigen ca125/muc16 to mesothelin mediates cell adhesion. The Journal of biological chemistry 2004;279:9190-8.

57. Theriault C, Pinard M, Comamala M, Migneault M, Beaudin J, Matte I, et al. Muc16 (ca125) regulates epithelial ovarian cancer cell growth, tumorigenesis and metastasis. Gynecologic oncology 2011;121:434-43.

58. Patankar MS, Jing Y, Morrison JC, Belisle JA, Lattanzio FA, Deng Y, et al. Potent suppression of natural killer cell response mediated by the ovarian tumor marker ca125. Gynecologic oncology 2005;99:704-13.

59. Birgen D, Ertem U, Duru F, Sahin G, Yuksek N, Bozkurt C, et al. Serum ca 125 levels in children with acute leukemia and lymphoma. Leukemia & lymphoma 2005;46:1177-81.

60. Kutluk T, Varan A, Erbas B, Buyukpamukcu M. Serum ca 125 levels in children with non-hodgkin's lymphoma. Pediatric hematology and oncology 1999;16:311-9.

61. Bon GG, Kenemans P, Dekker JJ, Hompes PG, Verstraeten RA, van Kamp GJ, Schoemaker J. Fluctuations in ca 125 and ca 15-3 serum concentrations during spontaneous ovulatory cycles. Human reproduction 1999;14:566-70.

62. Erbagci AB, Yilmaz N, Kutlar I. Menstrual cycle dependent variability for serum tumor markers cea, afp, ca 19-9, ca 125 and ca 15-3 in healthy women. Disease markers 1999;15:259-67.

77

63. Sarandakou A, Protonotariou E, Rizos D. Tumor markers in biological fluids associated with pregnancy. Critical reviews in clinical laboratory sciences 2007;44:151-78.

64. Tuxen MK, Soletormos G, Dombernowsky P. Tumor markers in the management of patients with ovarian cancer. Cancer treatment reviews 1995;21:215-45.

65. Gallup DG, Talledo E. Management of the adnexal mass in the 1990s. Southern medical journal 1997;90:972-81.

66. Chen DX, Schwartz PE, Li XG, Yang Z. Evaluation of ca 125 levels in differentiating malignant from benign tumors in patients with pelvic masses. Obstetrics and gynecology 1988;72:23-7.

67. Hammarstrom S. The carcinoembryonic antigen (cea) family: Structures, suggested functions and expression in normal and malignant tissues. Seminars in cancer biology 1999;9:67-81.

68. Konstantopoulos K, Thomas SN. Cancer cells in transit: The vascular interactions of tumor cells. Annual review of biomedical engineering 2009;11:177-202.

69. Tibbetts LM, Doremus CM, Tzanakakis GN, Vezeridis MP. Liver metastases with 10 human colon carcinoma cell lines in nude mice and association with carcinoembryonic antigen production. Cancer 1993;71:315-21.

70. Blumenthal RD, Osorio L, Hayes MK, Horak ID, Hansen HJ, Goldenberg DM. Carcinoembryonic antigen antibody inhibits lung metastasis and augments chemotherapy in a human colonic carcinoma xenograft. Cancer immunology, immunotherapy : CII 2005;54:315-27.

71. Kammerer R, von Kleist S. Cea expression of colorectal adenocarcinomas is correlated with their resistance against lak-cell lysis. International journal of cancer Journal international du cancer 1994;57:341-7.

72. Perkins GL, Slater ED, Sanders GK, Prichard JG. Serum tumor markers. American family physician 2003;68:1075-82.

73. Fletcher RH. Carcinoembryonic antigen. Annals of internal medicine 1986;104:66-73.

74. Clinical practice guidelines for the use of tumor markers in breast and colorectal cancer. Adopted on may 17, 1996 by the american society of clinical oncology. Journal of clinical oncology : official journal of the American Society of Clinical Oncology 1996;14:2843-77.

75. Angel CA, Pratt CB, Rao BN, Schell MJ, Parham DM, Lobe TE, Fleming ID. Carcinoembryonic antigen and carbohydrate 19-9 antigen as markers for colorectal carcinoma in children and adolescents. Cancer 1992;69:1487-91.

78

76. Li J, Chen H, Mariani A, Chen D, Klatt E, Podratz K, et al. He4 (wfdc2) promotes tumor growth in endometrial cancer cell lines. International journal of molecular sciences 2013;14:6026-43.

77. Bingle L, Singleton V, Bingle CD. The putative ovarian tumour marker gene he4 (wfdc2), is expressed in normal tissues and undergoes complex alternative splicing to yield multiple protein isoforms. Oncogene 2002;21:2768-73.

78. Schalkwijk J, Wiedow O, Hirose S. The trappin gene family: Proteins defined by an n-terminal transglutaminase substrate domain and a c-terminal four-disulphide core. The Biochemical journal 1999;340 ( Pt 3):569-77.

79. Kirchhoff C. Molecular characterization of epididymal proteins. Reviews of reproduction 1998;3:86-95.

80. Hellstrom I, Raycraft J, Hayden-Ledbetter M, Ledbetter JA, Schummer M, McIntosh M, et al. The he4 (wfdc2) protein is a biomarker for ovarian carcinoma. Cancer research 2003;63:3695-700.

81. Jacob F, Meier M, Caduff R, Goldstein D, Pochechueva T, Hacker N, et al. No benefit from combining he4 and ca125 as ovarian tumor markers in a clinical setting. Gynecologic oncology 2011;121:487-91.

82. Moore RG, Brown AK, Miller MC, Skates S, Allard WJ, Verch T, et al. The use of multiple novel tumor biomarkers for the detection of ovarian carcinoma in patients with a pelvic mass. Gynecologic oncology 2008;108:402-8.

83. Nolen B, Velikokhatnaya L, Marrangoni A, De Geest K, Lomakin A, Bast RC, Jr., Lokshin A. Serum biomarker panels for the discrimination of benign from malignant cases in patients with an adnexal mass. Gynecologic oncology 2010;117:440-5.

84. Moore RG, McMeekin DS, Brown AK, DiSilvestro P, Miller MC, Allard WJ, et al. A novel multiple marker bioassay utilizing he4 and ca125 for the prediction of ovarian cancer in patients with a pelvic mass. Gynecologic oncology 2009;112:40-6.

85. Nagy B, Jr., Krasznai ZT, Balla H, Csoban M, Antal-Szalmas P, Hernadi Z, Kappelmayer J. Elevated human epididymis protein 4 concentrations in chronic kidney disease. Annals of clinical biochemistry 2012;49:377-80.

86. Hertlein L, Stieber P, Kirschenhofer A, Krocker K, Nagel D, Lenhard M, Burges A. Human epididymis protein 4 (he4) in benign and malignant diseases. Clinical chemistry and laboratory medicine : CCLM / FESCC 2012;50:2181-8.

87. LeBleu VS, Teng Y, O'Connell JT, Charytan D, Muller GA, Muller CA, et al. Identification of human epididymis protein-4 as a fibroblast-derived mediator of fibrosis. Nature medicine 2013;19:227-31.

79

88. Moore RG, Miller MC, Eklund EE, Lu KH, Bast RC, Jr., Lambert-Messerlian G. Serum levels of the ovarian cancer biomarker he4 are decreased in pregnancy and increase with age. American journal of obstetrics and gynecology 2012;206:349 e1-7.

89. Lowe KA, Shah C, Wallace E, Anderson G, Paley P, McIntosh M, et al. Effects of personal characteristics on serum ca125, mesothelin, and he4 levels in healthy postmenopausal women at high-risk for ovarian cancer. Cancer epidemiology, biomarkers & prevention : a publication of the American Association for Cancer Research, cosponsored by the American Society of Preventive Oncology 2008;17:2480-7.

90. Urban N, Thorpe J, Karlan BY, McIntosh MW, Palomares MR, Daly MB, et al. Interpretation of single and serial measures of he4 and ca125 in asymptomatic women at high risk for ovarian cancer. Cancer epidemiology, biomarkers & prevention : a publication of the American Association for Cancer Research, cosponsored by the American Society of Preventive Oncology 2012;21:2087-94.

91. Ischia J, Patel O, Shulkes A, Baldwin GS. Gastrin-releasing peptide: Different forms, different functions. BioFactors 2009;35:69-75.

92. Adachi N, Aoyagi K, Saito M, Matsuda I, Yamaguchi K. Age-related changes of serum progastrin-releasing peptide levels during childhood. Acta paediatrica Japonica; Overseas edition 1997;39:336-8.

93. Gonzalez N, Moody TW, Igarashi H, Ito T, Jensen RT. Bombesin-related peptides and their receptors: Recent advances in their role in physiology and disease states. Current opinion in endocrinology, diabetes, and obesity 2008;15:58-64.

94. Ohki-Hamazaki H, Iwabuchi M, Maekawa F. Development and function of bombesin-like peptides and their receptors. The International journal of developmental biology 2005;49:293-300.

95. Jensen RT, Battey JF, Spindel ER, Benya RV. International union of pharmacology. Lxviii. Mammalian bombesin receptors: Nomenclature, distribution, pharmacology, signaling, and functions in normal and disease states. Pharmacological reviews 2008;60:1-42.

96. Patel O, Shulkes A, Baldwin GS. Gastrin-releasing peptide and cancer. Biochimica et biophysica acta 2006;1766:23-41.

97. Patel O, Clyde D, Chang M, Nordlund MS, Steel R, Kemp BE, et al. Pro-grp-derived peptides are expressed in colorectal cancer cells and tumors and are biologically active in vivo. Endocrinology 2012;153:1082-92.

98. Inaji H, Komoike Y, Motomura K, Higashiyama M, Ohtsuru M, Funai H, et al. Demonstration and diagnostic significance of pro-gastrin-releasing peptide in medullary thyroid carcinoma. Oncology 2000;59:122-5.

80

99. Stieber P, Yamaguchi, K. Progrp enables diagnosis of small-cell lung cancer. Washington: AACC Press, 2002.

100. Harmsma M, Schutte B, Ramaekers FC. Serum markers in small cell lung cancer: Opportunities for improvement. Biochimica et biophysica acta 2013;1836:255-72.

101. Priest JR, Williams GM, Hill DA, Dehner LP, Jaffe A. Pulmonary cysts in early childhood and the risk of malignancy. Pediatric pulmonology 2009;44:14-30.

102. Molina R, Auge JM, Alicarte J, Filella X, Vinolas N, Ballesta AM. Pro-gastrin-releasing peptide in patients with benign and malignant diseases. Tumour biology : the journal of the International Society for Oncodevelopmental Biology and Medicine 2004;25:56-61.

103. Balk SP, Ko YJ, Bubley GJ. Biology of prostate-specific antigen. Journal of clinical oncology : official journal of the American Society of Clinical Oncology 2003;21:383-91.

104. Stone JG, Rolston RK, Ueda M, Lee HG, Richardson SL, Castellani RJ, et al. Evidence for the novel expression of human kallikrein-related peptidase 3, prostate-specific antigen, in the brain. International journal of clinical and experimental pathology 2009;2:267-74.

105. Rehault S, Monget P, Mazerbourg S, Tremblay R, Gutman N, Gauthier F, Moreau T. Insulin-like growth factor binding proteins (igfbps) as potential physiological substrates for human kallikreins hk2 and hk3. European journal of biochemistry / FEBS 2001;268:2960-8.

106. Cohen P, Graves HC, Peehl DM, Kamarei M, Giudice LC, Rosenfeld RG. Prostate-specific antigen (psa) is an insulin-like growth factor binding protein-3 protease found in seminal plasma. The Journal of clinical endocrinology and metabolism 1992;75:1046-53.

107. Fielder PJ, Rosenfeld RG, Graves HC, Grandbois K, Maack CA, Sawamura S, et al. Biochemical analysis of prostate specific antigen-proteolyzed insulin-like growth factor binding protein-3. Growth regulation 1994;4:164-72.

108. Clements JA, Willemsen NM, Myers SA, Dong Y. The tissue kallikrein family of serine proteases: Functional roles in human disease and potential as clinical biomarkers. Critical reviews in clinical laboratory sciences 2004;41:265-312.

109. Obiezu CV, Diamandis EP. Human tissue kallikrein gene family: Applications in cancer. Cancer letters 2005;224:1-22.

110. Fortier AH, Nelson BJ, Grella DK, Holaday JW. Antiangiogenic activity of prostate-specific antigen. Journal of the National Cancer Institute 1999;91:1635-40.

81

111. Fortier AH, Holaday JW, Liang H, Dey C, Grella DK, Holland-Linn J, et al. Recombinant prostate specific antigen inhibits angiogenesis in vitro and in vivo. The Prostate 2003;56:212-9.

112. Heidtmann HH, Nettelbeck DM, Mingels A, Jager R, Welker HG, Kontermann RE. Generation of angiostatin-like fragments from plasminogen by prostate-specific antigen. British journal of cancer 1999;81:1269-73.

113. Goh CL, Saunders EJ, Leongamornlert DA, Tymrakiewicz M, Thomas K, Selvadurai ED, et al. Clinical implications of family history of prostate cancer and genetic risk single nucleotide polymorphism (snp) profiles in an active surveillance cohort. BJU international 2013;112:666-73.

114. Schroder FH, Hugosson J, Roobol MJ, Tammela TL, Ciatto S, Nelen V, et al. Screening and prostate-cancer mortality in a randomized european study. The New England journal of medicine 2009;360:1320-8.

115. Andriole GL, Crawford ED, Grubb RL, 3rd, Buys SS, Chia D, Church TR, et al. Mortality results from a randomized prostate-cancer screening trial. The New England journal of medicine 2009;360:1310-9.

116. Ito K, Yamamoto T, Ohi M, Kurokawa K, Suzuki K, Yamanaka H. Free/total psa ratio is a powerful predictor of future prostate cancer morbidity in men with initial psa levels of 4.1 to 10.0 ng/ml. Urology 2003;61:760-4.

117. Luo LY, Diamandis EP, Look MP, Soosaipillai AP, Foekens JA. Higher expression of human kallikrein 10 in breast cancer tissue predicts tamoxifen resistance. British journal of cancer 2002;86:1790-6.

118. Nagar R, Msalati AA. Changes in serum psa during normal menstrual cycle. Indian journal of clinical biochemistry : IJCB 2013;28:84-9.

119. Mardanian F, Heidari N. Diagnostic value of prostate-specific antigen in women with polycystic ovary syndrome. Journal of research in medical sciences : the official journal of Isfahan University of Medical Sciences 2011;16:999-1005.

120. Tchetgen MB, Oesterling JE. The effect of prostatitis, urinary retention, ejaculation, and ambulation on the serum prostate-specific antigen concentration. The Urologic clinics of North America 1997;24:283-91.

121. Catalona WJ, Richie JP, Ahmann FR, Hudson MA, Scardino PT, Flanigan RC, et al. Comparison of digital rectal examination and serum prostate specific antigen in the early detection of prostate cancer: Results of a multicenter clinical trial of 6,630 men. The Journal of urology 1994;151:1283-90.

122. Randell EW, Diamandis EP, Ellis G. Serum prostate-specific antigen measured in children from birth to age 18 years. Clinical chemistry 1996;42:420-3.

82

123. Suminami Y, Nawata S, Kato H. Biological role of scc antigen. Tumour biology : the journal of the International Society for Oncodevelopmental Biology and Medicine 1998;19:488-93.

124. Kato H. Expression and function of squamous cell carcinoma antigen. Anticancer research 1996;16:2149-53.

125. Suminami Y, Nagashima S, Vujanovic NL, Hirabayashi K, Kato H, Whiteside TL. Inhibition of apoptosis in human tumour cells by the tumour-associated serpin, scc antigen-1. British journal of cancer 2000;82:981-9.

126. Murakami A, Fukushima C, Yositomi K, Sueoka K, Nawata S, Fujimoto M, et al. Tumor-related protein, the squamous cell carcinoma antigen binds to the intracellular protein carbonyl reductase. International journal of oncology 2010;36:1395-400.

127. Numa F, Takeda O, Nakata M, Nawata S, Tsunaga N, Hirabayashi K, et al. Tumor necrosis factor-alpha stimulates the production of squamous cell carcinoma antigen in normal squamous cells. Tumour biology : the journal of the International Society for Oncodevelopmental Biology and Medicine 1996;17:97-101.

128. Torre GC. Scc antigen in malignant and nonmalignant squamous lesions. Tumour biology : the journal of the International Society for Oncodevelopmental Biology and Medicine 1998;19:517-26.

129. Bolger BS, Dabbas M, Lopes A, Monaghan JM. Prognostic value of preoperative squamous cell carcinoma antigen level in patients surgically treated for cervical carcinoma. Gynecologic oncology 1997;65:309-13.

130. Duk JM, Groenier KH, de Bruijn HW, Hollema H, ten Hoor KA, van der Zee AG, Aalders JG. Pretreatment serum squamous cell carcinoma antigen: A newly identified prognostic factor in early-stage cervical carcinoma. Journal of clinical oncology : official journal of the American Society of Clinical Oncology 1996;14:111-8.

131. Gadducci A, Tana R, Cosio S, Genazzani AR. The serum assay of tumour markers in the prognostic evaluation, treatment monitoring and follow-up of patients with cervical cancer: A review of the literature. Critical reviews in oncology/hematology 2008;66:10-20.

132. Esajas MD, Duk JM, de Bruijn HW, Aalders JG, Willemse PH, Sluiter W, et al. Clinical value of routine serum squamous cell carcinoma antigen in follow-up of patients with early-stage cervical cancer. Journal of clinical oncology : official journal of the American Society of Clinical Oncology 2001;19:3960-6.

133. Niibe Y, Kazumoto T, Toita T, Yamazaki H, Higuchi K, Ii N, et al. Frequency and characteristics of isolated para-aortic lymph node recurrence in patients with uterine cervical carcinoma in japan: A multi-institutional study. Gynecologic oncology 2006;103:435-8.

83

134. Kim BG. Squamous cell carcinoma antigen in cervical cancer and beyond. Journal of gynecologic oncology 2013;24:291-2.

135. Balk SJ, Council on Environmental H, Section on D. Ultraviolet radiation: A hazard to children and adolescents. Pediatrics 2011;127:e791-817.

136. Molina R, Filella X, Torres MD, Ballesta AM, Mengual P, Cases A, Balaque A. Scc antigen measured in malignant and nonmalignant diseases. Clinical chemistry 1990;36:251-4.

137. Saah AJ, Hoover DR. "Sensitivity" and "specificity" reconsidered: The meaning of these terms in analytical and diagnostic settings. Annals of internal medicine 1997;126:91-4.

138. Gutierrez A, Martinez-Serra J, Barcelo B, Sampol A, Vinas L, Gonzalez G, et al. Prognostic value of serum ca125 levels in diffuse large b-cell lymphoma: Potential role of a new sex- and age-adjusted reference value. International journal of laboratory hematology 2010;32:582-9.

139. Lahdenne P, Kuusela P, Siimes MA, Ronnholm KA, Salmenpera L, Heikinheimo M. Biphasic reduction and concanavalin a binding properties of serum alpha-fetoprotein in preterm and term infants. The Journal of pediatrics 1991;118:272-6.

140. Wong JR, Harris JK, Rodriguez-Galindo C, Johnson KJ. Incidence of childhood and adolescent melanoma in the united states: 1973-2009. Pediatrics 2013;131:846-54.

141. Duffy MJ. Tumor markers in clinical practice: A review focusing on common solid cancers. Medical principles and practice : international journal of the Kuwait University, Health Science Centre 2013;22:4-11.

142. Yamashita S, Tokuishi K, Moroga T, Yamamoto S, Ohbo K, Miyahara S, et al. Serum level of he4 is closely associated with pulmonary adenocarcinoma progression. Tumour biology : the journal of the International Society for Oncodevelopmental Biology and Medicine 2012;33:2365-70.

143. Morris NM, Udry JR. Validation of a self-administered instrument to assess stage of adolescent development. Journal of youth and adolescence 1980;9:271-80.

144. Statistics Canada. Ethnic origins, 2006 counts, for Canada, provinces and territories - 20% sample data [Table]. http://www12statcan.ca/censusrecensement/2006/dp-pd/hlt/07-562/pages/page.cfm?Lang=E&Geo=PR&Code=35&Data=Count&Table= 2&StartRec=1&Sort=3&Display=All (Accessed April 2012).

145. Harris EK, Boyd JC. On dividing reference data into subgroups to produce separate reference ranges. Clinical chemistry 1990;36:265-70.

146. Box GEPC, D.R. . An analysis of transformations. Journal of the Royal Statistical Society, Series B 1964;26:211-52.

84

147. Filiben JJ. The probability plot correlation coefficient test for normality Technometrics 1975;17:111-7.

148. Tukey JW. Exploratory data analysis. Boston: Addison-Wesley; 1977. 688 p.

149. Hubert M, Van der Veeken S. Outlier detection for skewed data. J. Chemometrics 2008;22:235– 46.

150. Horn PS PA. Reference intervals: A user's guide. Washington (DC): AACC Press, 2005.

151. Efron B, Tibishirani, R. An introduction to the bootstrap. New York.: Chapman & Hall Inc., 1993.

152. Zarghami N, Grass L, Sauter ER, Diamandis EP. Prostate-specific antigen in serum during the menstrual cycle. Clinical chemistry 1997;43:1862-7.

153. Fleuren GJ, Nap M, Aalders JG, Trimbos JB, de Bruijn HW. Explanation of the limited correlation between tumor ca 125 content and serum ca 125 antigen levels in patients with ovarian tumors. Cancer 1987;60:2437-42.

154. Quirk JG, Jr., Brunson GL, Long CA, Bannon GA, Sanders MM, O'Brien TJ. Ca 125 in tissues and amniotic fluid during pregnancy. American journal of obstetrics and gynecology 1988;159:644-9.

155. Botsis D, Sarandakou A, Kassanos D, Kontoravdis A, Rizos D, Protonotariou E, et al. Breast cancer markers during normal pregnancy. Anticancer research 1999;19:3539-41.

156. Panidis D, Vlassis G, Matalliotakis J, Skiadopoulos S, Kalogeropoulos A. Serum levels of the oncofetal antigens ca-125, ca 19-9 and ca 15-3 in patients with endometriosis. Journal of endocrinological investigation 1988;11:801-4.

157. Kim MR, Gupta MK, Travers SH, Rogers DG, Van Lente F, Faiman C. Serum prostate specific antigen, sex hormone binding globulin and free androgen index as markers of pubertal development in boys. Clinical endocrinology 1999;50:203-10.

158. Ghosh I, Bhattacharjee D, Das AK, Chakrabarti G, Dasgupta A, Dey SK. Diagnostic role of tumour markers cea, ca15-3, ca19-9 and ca125 in lung cancer. Indian journal of clinical biochemistry : IJCB 2013;28:24-9.

159. Mitchell G, Sibley PE, Wilson AP, Sauter E, A'Hern R, Eeles RA. Prostate-specific antigen in nipple aspiration fluid: Menstrual cycle variability and correlation with serum prostate-specific antigen. Tumour biology : the journal of the International Society for Oncodevelopmental Biology and Medicine 2002;23:287-97.

160. Shaw JL, Cohen A, Konforte D, Binesh-Marvasti T, Colantonio DA, Adeli K. Validity of establishing pediatric reference intervals based on hospital patient data: A comparison of the modified hoffmann approach to caliper reference intervals obtained in healthy children. Clinical biochemistry 2014;47:166-72.

85

161. Quinn FA, Tam MC, Wong PT, Poon PK, Leung MS. Thyroid autoimmunity and thyroid hormone reference intervals in apparently healthy chinese adults. Clinica chimica acta; international journal of clinical chemistry 2009;405:156-9.

162. Van Gorp T, Cadron I, Despierre E, Daemen A, Leunen K, Amant F, et al. He4 and ca125 as a diagnostic test in ovarian cancer: Prospective validation of the risk of ovarian malignancy algorithm. British journal of cancer 2011;104:863-70.

163. Cheah PL, Liam CK, Yap SF, Looi LM. Squamous cell carcinoma antigen as an adjunct tumour marker in primary carcinoma of the lung. Journal of clinical pathology 1994;47:535-7.

164. Molina R, Filella X, Auge JM. Progrp: A new biomarker for small cell lung cancer. Clinical biochemistry 2004;37:505-11.

165. Brinc D, Chan MK, Venner AA, Pasic MD, Colantonio D, Kyriakopolou L, Adeli K. Long-term stability of biochemical markers in pediatric serum specimens stored at -80 degrees c: A caliper substudy. Clinical biochemistry 2012;45:816-26.

166. Nordlund MS, Bjerner J, Warren DJ, Nustad K, Paus E. Progastrin-releasing peptide: Stability in plasma/serum and upper reference limit. Tumour biology : the journal of the International Society for Oncodevelopmental Biology and Medicine 2008;29:204-10.

167. Woodrum D, York L. Two-year stability of free and total psa in frozen serum samples. Urology 1998;52:247-51.

168. Scaramuzzino DA, Schulte K, Mack BN, Soriano TF, Fritsche HA. Five-year stability study of free and total prostate-specific antigen concentrations in serum specimens collected and stored at -70 degrees c or less. The International journal of biological markers 2007;22:206-13.

169. Gao YC, Yuan ZB, Yang YD, Lu HK. Effect of freeze-thaw cycles on serum measurements of afp, cea, ca125 and ca19-9. Scandinavian journal of clinical and laboratory investigation 2007;67:741-7.

170. Banfi G, Parma P, Pontillo M. Stability of tumor markers ca 19.9, ca 125, and ca 15.3 in serum obtained from plain tubes and tubes containing thixotropic gel separator. Clinical chemistry 1997;43:2430-1.

171. Aldrimer M, Ridefelt P, Rodoo P, Niklasson F, Gustafsson J, Hellberg D. Population-based pediatric reference intervals for hematology, iron and transferrin. Scandinavian journal of clinical and laboratory investigation 2013;73:253-61.

172. Davidson DF, Hammond PJ, Murphy D, Carachi R. Age-related medical decision limits for urinary free (unconjugated) metadrenalines, catecholamines and metabolites in random urine specimens from children. Annals of clinical biochemistry 2011;48:358-66.

86

173. Crandall BF, Lebherz TB, Schroth PC, Matsumoto M. Alpha-fetoprotein concentrations in maternal serum: Relation to race and body weight. Clinical chemistry 1983;29:531-3.

174. O'Brien JE, Dvorin E, Drugan A, Johnson MP, Yaron Y, Evans MI. Race-ethnicity-specific variation in multiple-marker biochemical screening: Alpha-fetoprotein, hcg, and estriol. Obstetrics and gynecology 1997;89:355-8.

175. Sizaret P, Tuyns A, Martel N, Jouvenceaux A, Levin A, Ong YW, Rive J. Alpha-fetoprotein levels in normal males from seven ethnic groups with different hepatocellular carcinoma risks. Annals of the New York Academy of Sciences 1975;259:136-55.

176. Mokhtar N, Thevarajah M, Ma N, M I. Human epididymis protein 4 reference intervals in a multiethnic asian women population. Asian Pacific journal of cancer prevention : APJCP 2012;13:6391-5.

177. Pauler DK, Menon U, McIntosh M, Symecko HL, Skates SJ, Jacobs IJ. Factors influencing serum ca125ii levels in healthy postmenopausal women. Cancer epidemiology, biomarkers & prevention : a publication of the American Association for Cancer Research, cosponsored by the American Society of Preventive Oncology 2001;10:489-93.

178. Haines AP, Levin AG, Fritsche HA. Ethnic-group differences in serum levels of carcinoembryonic antigen. Lancet 1979;2:969.

179. Henderson RJ, Eastham JA, Culkin DJ, Kattan MW, Whatley T, Mata J, et al. Prostate-specific antigen (psa) and psa density: Racial differences in men without prostate cancer. Journal of the National Cancer Institute 1997;89:134-8.

180. Sarma AV, St Sauver JL, Jacobson DJ, McGree ME, Klee GG, Lieber MM, et al. Racial differences in longitudinal changes in serum prostate-specific antigen levels: The olmsted county study and the flint men's health study. Urology 2014;83:88-93.

181. Sturgeon CM, Ellis A., Al-Sadie R. Annual review for 2007. Vol. http://www.ukneqas.org.uk., 2007.

182. Sturgeon CM, Seth J. Why do immunoassays for tumour markers give differing results?--a view from the uk national external quality assessment schemes. European journal of clinical chemistry and clinical biochemistry : journal of the Forum of European Clinical Chemistry Societies 1996;34:755-9.

183. Erden G, Barazi AO, Tezcan G, Yildirimkaya MM. Biological variation and reference change values of ca 19-9, cea, afp in serum of healthy individuals. Scandinavian journal of clinical and laboratory investigation 2008;68:212-8.

184. Fraser CG. Data on biological variation: Essential prerequisites for introducing new procedures? Clin Chem 1994;40:1671-3.

87

185. Plebani M, Giacomini A, Beghi L, de Paoli M, Roveroni G, Galeotti F, et al. Serum tumor markers in monitoring patients: Interpretation of results using analytical and biological variation. Anticancer research 1996;16:2249-52.

pediatric population reference value distributions for ... · pediatric population reference value...

Documents