Download - Diagnostic value of p16INK4A, Ki-67, and human papillomavirus L1 capsid protein immunochemical staining on cell blocks from residual liquid-based gynecologic cytology specimens

Diagnostic Value of p16INK4A, Ki-67, andHuman Papillomavirus L1 Capsid ProteinImmunochemical Staining on Cell BlocksFrom Residual Liquid-Based GynecologicCytology SpecimensLi Yu, MD1,2; Liantang Wang, MD, PhD1; Juemin Zhong, MD1; and Shangwu Chen, MD, PhD3

BACKGROUND: This study was conducted to evaluate the reliability and role of cell block preparations in the diagno-

sis of neoplastic and preneoplastic lesions of the cervix and to improve the value of cell block preparations in diag-

nosing and predicting the prognosis of cervical lesions through immunostaining of p16INK4A (p16), Ki-67, and human

papillomavirus (HPV) L1 capsid protein (HPV L1). METHODS: In total, 138 specimens were diagnosed on liquid-based

cytology (LBC) and cell block preparations, and 63 specimens were subjected subsequently to tissue follow-up and

immunostaining for p16, Ki-67, and HPV L1 on cell block sections. RESULTS: In 42 specimens that were diagnosed as

low-grade squamous intraepithelial lesion, high-grade squamous intraepithelial lesion (HSIL), and squamous cell car-

cinoma (SCC) on cell blocks, 38 specimens (90.5%) were confirmed by histopathologic reports, and there was

slightly better than 81.6% agreement between LBC and tissue follow-up. Immunointensity and cells that were positive

for p16 were enhanced according to increased pathologic grade and differed statistically between cervical intraepi-

thelial neoplasia 1 (CIN-1) and CIN-2/CIN-3 as well as SCC. The positive rates of HPV L1 decreased gradually according

to the severity of cervical neoplasia, and HPV L1/p16 expression patterns were related to the severity of cervical

lesions. CONCLUSIONS: The cell block preparation technique was complementary to LBC, and the authors concluded

that the application of LBC combined with cell block preparations may improve the diagnostic accuracy of cytology.

Immunostaining for p16 and Ki-67 on cell block preparations can help to improve the diagnostic accuracy of HSIL

and SCC. A combined expression pattern of p16 and HPV L1 may serve as a valuable index for predicting prognosis

and follow-up of cervical dysplastic lesions. Cancer (Cancer Cytopathol) 2010;118:47–55. VC 2010 American Cancer

Society.

KEYWORDS: cell block preparations, immunocytochemistry, liquid-based cytology, human papillomavirus L1 capsid

protein, p16INK4A, Ki-67, cervical intraepithelial neoplasia.

Cervical cancer is the second most frequent tumor in women worldwide,1 and cervical intraepithelial neoplasia(CIN) is the key stage of tumorigenesis for cervical epithelium. Therefore, the early diagnosis of CIN is crucially importantfor the prevention and treatment of cervical neoplasia. Morbidity and mortality from cervical carcinoma have declinedremarkably through mass screening with the Papanicolaou (Pap) smear.2 The great drawback of the Pap test is the highrate of false-negative and false-positive results.3 Although liquid-based cytology (LBC) has improved sensitivity, it has notimproved specificity for diagnosing cervical lesions.4,5 In recent years, cell block preparation has been used as a diagnostictechnique to complement liquid-based, monolayer cervicovaginal specimens.6-8 In addition, many dysplasia-associatedbiomarkers have been identified and used to improve the diagnostic accuracy of neoplastic and preneoplastic lesions of thecervix in histology and cytology.9,10

DOI: 10.1002/cncy.20061, Received: September 25, 2009; Revised: October 28, 2009; Accepted: November 4, 2009, Published online January 12, 2010 in Wiley

InterScience (www.interscience.wiley.com)

Corresponding authors: Shangwu Chen, MD, PhD, Department of Biochemistry, School of Life Sciences, Sun Yat-sen (Zhongshan) University, Guangzhou,

510275, P.R. China; Fax: (011) 86-20-39332950; [email protected] and Li Yu, MD, Department of Histopathology, the First Affiliated Hospital, Sun Yat-sen

(Zhongshan) University, Guangzhou 510080, P.R. China; Fax: (011) 86-20-87331780; [email protected]

1Department of Histopathology, First Affiliated Hospital, Sun Yat-sen (Zhongshan) University, Guangzhou, People’s Republic of China; 2Department of Histopathol-

ogy, Yuexiu Hospital of Guangzhou, Guangzhou, People’s Republic of China; 3State Key Laboratory for Biocontrol, Guangdong Key Laboratory of Pharmaceutical

Functional Genes, Department of Biochemistry, School of Life Sciences, Sun Yat-sen (Zhongshan) University, Guangzhou, People’s Republic of China

Cancer Cytopathology February 25, 2010 47

Original Article

Human papilloma virus (HPV) infection is anessential factor for the development of cervical cancers.p16INK4A (p16) and Ki-67 are 2 surrogate markers ofhigh-risk HPV infection and high-grade CIN.11 Nor-mally, p16 is a cell-cycle inhibitor that binds to cyclin-de-pendent kinase 4 (CDK4)/CDK6 and prevents thephosphorylation and subsequent inactivation of the reti-noblastoma protein (pRB).12 The down-regulation ofp16 has been associated with carcinogenesis in a variety oforgan systems.13,14 Integration of HPV DNA into thehost genome results in the overexpression of viral proteinsE6 and E715; E7, in turn, binds to and inactivates pRBand induces the production of p16 through a negativefeedback loop.16 Ki-67 is a well known cell proliferationmarker and is highly correlated with the grade of dysplas-tic changes in the cervical epithelium.17,18 Positive immu-nostaining for both p16 and Ki-67 in the upper two-thirds of the squamous epithelium is a good indicator ofneoplastic atypical squamous lesions.19

HPV L1 (L1) capsid protein is a nuclear proteinexpressed by all HPV subtypes during a productive infec-tion. The production of this protein is linked to the matu-ration process of basal to superficial epithelial cells;therefore, and L1 is expressed strongly at the superficiallayer of the epithelium.9 L1 is expressed in the early pro-ductive phase of HPV infection and progressively is lost inthe later transformation phase, when p16 becomes overex-pressed.10 It has been demonstrated that L1 has prognos-tic relevance in mild-to-moderate dysplastic lesions, inwhich its expression indicates a greater tendency to regresscompared with L1-negative lesions.20,21

In the current study, we modified the preparation ofcell blocks from residual liquid-based gynecologic cytol-ogy specimens and evaluated the reliability and role of cellblock preparations in the diagnosis of neoplastic and pre-neoplastic lesions of the cervix. In addition, immunocyto-chemical staining for p16, Ki-67, and L1 capsid proteinon cell blocks was performed to improve the value of cellblock preparations in diagnosing cervical lesions and esti-mating their prognostic relevance to CIN. To our knowl-edge, this is the first time immunoreactivity to L1 capsidprotein has been detected on cell blocks for the predictionof CIN progression.

MATERIALS AND METHODS

Specimens and Study Design

Cervicovaginal cytology specimens were collected from147 women (mean age, 39 years; range, 18-83 years) who

were selected on the basis of the SurePath liquid-basedpreparation slides (TriPath Imaging, Burlington, NC).From the 147 specimens that were selected, we were ableto prepare sections from 138 cell blocks (9 specimenscould not be prepared, because they were limited by lowcellularity). Cytology diagnoses of liquid-based Papsmears were obtained using terminology according to TheBethesda System (2001), and the specimens were diag-nosed as follows: 22 specimens were negative for squa-mous intraepithelial lesions or malignancy (NILM), 43specimens were diagnosed as atypical squamous cells ofundetermined significance (ASCUS), 12 specimens werediagnosed as low-grade squamous intraepithelial lesion(LSIL), 58 specimens were diagnosed as high-grade squa-mous intraepithelial lesion (HSIL), and 3 specimens werediagnosed as squamous cell carcinoma (SCC). Sixty-threeof 138 specimens, including 12 diagnosed as NILM, sub-sequently were subjected to tissue follow-up and immu-nostaining for p16, Ki-67, and L1 capsid protein on cellblock sections. Tissues that were biopsy proven as eitherpositive or negative for squamous cell dysplasia were usedas positive and negative controls, respectively, forimmunocytochemistry.

Cell Block Preparation

Residual SurePath samples were centrifuged for 5 minutesat 1500 revolutions per minute. The supernatant fluidwas poured off, the pellet was resuspended with 5 to 8drops of plasma, 1 or 2 drops of 2% CaCl2 were addedand mixed gently, and the mixture was set aside forapproximately 20 minutes until a clot formed. The clotwas wrapped with lens tissue, placed onto a histology cas-sette, and fixed in 10% neutral buffered formalin for 10minutes. The cassette was then processed as a routine sur-gical specimen. Sections (4 lm) were cut from the cellblocks and stained with hematoxylin and eosin for mor-phologic evaluation.

Immunocytochemistry

Cell block sections were deparaffinized in xylene, washedin ethanol, and finally washed in phosphate-buffered sa-line, pH 7.4. To increase specificity and sensitivity, sec-tions were pretreated in a microwave for 20 minutes onhigh in Tris/ethylene diamine tetraacetic acid buffer, pH9.0. After cooling to room temperature and rinsing in dis-tilled water, endogenous peroxidase activity was blockedwith 3% H2O2 for 15 minutes. The sections were sub-jected to immunocytochemistry with either mouse anti-human p16 (clone JC8) monoclonal antibody (dilution,

Original Article

48 Cancer Cytopathology February 25, 2010

1:50; GeneTex, San Antonio, Tex) or rabbit anti-Ki-67(clone SP6) monoclonal antibody (dilution, 1:100; LabVision, Fremont, Calif). Briefly, after incubation with pri-mary antibody at 4�C overnight, the sections were washedwith phosphate-buffered saline 3 times and incubated foranother 20 minutes at room temperature with the horse-radish peroxidase (HRP)-labeled secondary antibody thatwas included in the MaxVision antimouse/rabbit andHRP-conjugated Polymer immunohistochemistry kit(MaxVision, Fuzhou, China). Color was developedthrough incubation with diaminobenzidine (Dako Corp,Carpinteria, Calif) at room temperature for 5 to 10minutes.

Positive p16 immunostaining cells appeared as abrown color within the nucleus with or without cytoplas-mic staining. Positive Ki-67 immunostaining appeared asa brown color only within the nucleus. At least 10 cellsstained for p16 or Ki-67 were considered positive to pre-vent nonspecific background staining,22 because p16staining also can be observed in tubal metaplasia, giantcells, malignant and benign endometrial cells, trichomo-nas, atypical glandular cells, atrophy, and inflamma-tion23,24; and Ki-67 is expressed normally in the parabasalcells of mature squamous epithelium.25 Dark brown indi-cated strong staining, and light brown indicated weakstaining. Positively stained cells distributed from the baseto the top of an epithelial fragment indicated full-thick-ness epithelial staining.

Detecting Human Papillomavirus L1Capsid Protein

HPV L1 capsid protein was detected through a techniqueof combined in situ hybridization with immunohisto-chemistry using the CytoReact-Cell/tissue broad-spec-trum HPV L1 Kit (Advanced Medical Science &Technology, LLC, Baltimore, Md) according to the man-ufacture’s instructions. The antibody recognizes the majorL1 capsid proteins of high-risk and low-risk HPV types,including HPV type 6 (HPV-6), HPV-11, HPV-16,HPV-18, HPV-31, HPV-33, HPV-42, HPV-51, HPV-52, HPV-56, and HPV-58. Then, the slides were counter-stained with hematoxylin and examined by light micros-copy. Nuclear staining was considered HPV L1-positive.

Data Processing and Statistical Analysis

Specimens from 138 selected patients were subjected toLBC diagnosis and cell block diagnosis simultaneously,and the specimens diagnosed as SCC, HSIL, LSIL,ASCUS, NILM were counted. The results were listed as a

table to evaluate the correlation of both diagnosticapproaches. For 63 specimens that had tissue follow-up,the degree of agreement between each histologic diagnosisand interpretation with LBC or cell block preparation wascompared and calculated. Subsequently, the 63 specimenswere subjected to immunostaining for p16, Ki-67, and L1capsid protein on cell block sections, and the expressionstatus of all 3 biomarkers as well as the combined expres-sion of p16 and L1 were analyzed. On the basis of histopa-thologic diagnoses, the 63 specimens that had tissuefollow-up available were classified into 4 groups: chroniccervicitis (CC), CIN-1, CIN-2/CIN-3, and SCC. Thesoftware program SPSS 13.0 (SPSS, Inc, Chicago, Ill) andthe chi-square test were used for statistical analyses, and Pvalues<.05 were considered indicative of statistical signifi-cance. For multiple comparisons, P values<.0083 (.05/.6)were considered statistically significant.

RESULTS

Correlation of Cell Block Diagnosis WithLiquid-Based Cytology Interpretation

The cytologic morphology of various cervical lesions on cellblocks and Pap smears is illustrated in Figure 1A-F. Dys-plastic cells were observed on both cell block preparationsand Pap smears, but squamous epithelial fragments wereobserved on the cell block sections (Fig. 1D-F). Similar tocorresponding tissue sections (Fig. 1G-I), tissue architecturewas observed in the fragments (Fig. 1D-F). Of 138 cellblocks, 4 were interpreted as SCC, 70 were interpreted asHSIL, 20 were interpreted as LSIL, 23 were interpreted asASCUS, and 21 were interpreted as NILM (Table 1). Of 4specimens that were diagnosed as SCC on cell blocks, 3were consistent with the LBC diagnosis, and the 1 exceptionwas had an LBC diagnosis of HSIL. Among the 70 speci-mens that were diagnosed as HSIL on cell blocks, 4 had anoriginal LBC diagnosis of LSIL, and 16 had an originalLBC diagnosis of ASCUS. Of the 20 specimens that werediagnosed as LSIL on cell blocks, only 8 specimens had thesame diagnosis with both methods; with LBC, 7 specimenswere diagnosed as HSIL, and 5 specimens were diagnosedas ASCUS. Of the 23 specimens that were diagnosed asASCUS on cell blocks, only 1 specimen was evaluated origi-nally as NILM in LBC. All 21 NILM specimens were diag-nosed consistently in both LBC and cell blocks. Overall, thediagnoses for 75.4% of specimens (104 of 138) agreed wellwith both methods. For specimens that had differing diag-noses, the diagnosis reported in most (27 specimens) was

Immunochemical Staining on Cell Blocks/Yu et al


Figure 1. These photomicrographs provide a morphologic comparison of various cervical lesions on Papanicolaou (Pap) smears,cell block preparations, and tissue sections and immunostaining of p16INK4A (p16,) Ki-67, and human papillomavirus (HPV) cap-sid protein L1 (HPV L1) on cell blocks. The Pap smear demonstrates nuclear enlargement and hyperchromasia of atypical squa-mous cells in (A) low-grade squamous intraepithelial lesion (LSIL); (B) a hyperchromatic, crowded group of cells in high-gradesquamous intraepithelial lesion (HSIL); and (C) a cluster of cancer cells with blood background in squamous cell carcinoma(SCC). (D-F) Squamous epithelial fragments are observed on these cell block sections in which the degree of dysplasia wasenhanced according to the severity of cervical lesions. Cervical biopsy revealed that the dysplasia was distributed (G) under thelower 1-third of the squamous epithelium in LSIL or (H) across the full thickness of the squamous epithelium in HSIL, and (I) can-cer nests invaded into muscle tissue in SCC. Immunodetection demonstrated sporadic staining for (J) p16, (M) Ki-67, and (P)HPV L1 in LSIL. The majority of squamous epithelial cells are immunostained for p16 and Ki-67 in (K,N) HSIL and (L,O) SCC. NoHPV L1 staining is observed in (Q) HSIL or (R) SCC. LBC indicates liquid-based cytology; CB, cell block.

Original Article


more severe in cell blocks than in LBC preparations (withthe exception of 7 specimens).

Tissue Follow-Up

To evaluate the accuracy of cell block and LBC diagnoses,63 of 138 specimens were subjected to tissue follow-up andwere diagnosed histopathologically as follows: 5 SCC, 31CIN-2/CIN-3, 15 CIN-1, and 12 CC (Table 2). In 42specimens that were diagnosed as LSIL, HSIL, and SCC incell blocks, 38 specimens (90.5%) were confirmed by histo-pathologic reports, for slightly greater than 81.6% agree-ment (31 of 38 specimens) between LBC and tissue follow-up with no statistically significant difference (P¼ .249).

Immunocytochemistry for p16 and Ki-67 onCell Block Preparations

The rates of positive staining for p16 in cell block prepara-tions that were diagnosed as CC, CIN-1, CIN-2/CIN-3,

and SCC were 0%, 40%, 100%, and 100%, respectively(Table 3). It was clear that the intensity of p16 stainingand the numbers of positive cells were enhanced accordingto increased pathologic grade (Fig. 1J-L). Cell block prep-arations for all SCC specimens and for approximately25% of CIN-2/CIN-3 specimens (26%) exhibited full-thickness epithelial staining for p16, whereas none of theCIN-1 or CC specimens exhibited full-thickness epithe-lial staining. Similar to p16 staining, the detecting ratesfor Ki-67 in CC, CIN-1, CIN-2/CIN-3, and SCC were0%, 40%, 87%, and 100%, respectively. The differencesin positive staining rates and staining intensity for p16between CIN-1 and CIN-2/CIN-3 and for SCC werestatistically significant. Positive rates for Ki-67 betweenspecimens that were diagnosed as CIN-1 and CIN-2/CIN-3 on cell block preparations also differedsignificantly.

Table 1. Correlation of Cell Block Diagnosis With Liquid-Based Cytology Diagnosis

LBC Diagnoses: No. of Specimens

Cell Block Diagnosis SCC HSIL LSIL ASCUS NILM Total

SCC 3 1 0 0 0 4

HSIL 0 50 4 16 0 70

LSIL 0 7 8 5 0 20

ASCUS 0 0 0 22 1 23

NILM 0 0 0 0 21 21

Total 3 58 12 43 22 138

LBC indicates liquid-based cytology; SCC, squamous cell carcinoma; HSIL, high-grade squamous intraepithelial lesion; LSIL, low-grade squamous intraepithe-

lial lesion; ASCUS, atypical squamous cells of undetermined significance; NILM, negative for squamous intraepithelial lesions or malignancy.

Table 2. Correlation of Histologic Diagnosis With Cell Block and Liquid-Based Cytology Diagnosis

No. of Histologic Diagnoses (%), n563

Diagnosis CC, n512 CIN-1, n515 CIN-2/CIN-3, n531 SCC, n55

Cell block diagnosisSCC 0 (0) 0 (0) 0 (0) 4 (80)

HSIL 0 (0) 2 (13) 26 (84) 1 (20)

LSIL 0 (0) 8 (53) 1 (3) 0 (0)

ASCUS 0 (0) 5 (33) 4 (13) 0 (0)

NILM 12 (100) 0 (0) 0 (0) 0 (0)

LBC diagnosisSCC 0 (0) 0 (0) 0 (0) 3 (60)

HSIL 0 (0) 2 (13) 20 (64) 1 (20)

LSIL 0 (0) 8 (53) 4 (13) 0 (0)

ASCUS 0 (0) 5 (33) 7 (23) 1 (20)

NILM 12 (100) 0 (0) 0 (0) 0 (0)

CC indicates chronic cervicitis; CIN, cervical intraepithelial neoplasia; SCC, squamous cell carcinoma; HSIL, high-grade squamous intraepithelial lesion; LSIL,

low-grade squamous intraepithelial lesion; ASCUS, atypical squamous cells of undetermined significance; NILM, negativity for squamous intraepithelial lesions

or malignancy; LBC, liquid-based cytology.



Human Papillomavirus L1 Capsid ProteinDetection on Cell Blocks

HPV L1 capsid protein was detected on cell blocks from63 specimens that had tissue follow-up available. Thedetection rates of HPV L1 in the specimens that werediagnosed on cell block preparations as CIN-1, CIN-2/CIN-3, and SCC were 40%, 19.4%, and 0%, respec-

tively, which decreased gradually according to the severityof cervical neoplasia (Table 3, Fig. 1P-R). Because a previ-ous study had demonstrated that the detection of p16 andL1 expression was helpful for estimating the biologic riskof cervical lesions, the combined expression of p16 andHPV L1 capsid protein in CIN and SCC was analyzed(Table 4).

Table 3. Immunostaining of p16, Ki-67, and Human Papillomavirus L1 Capsid Protein in Cell Block Preparations

No. of Specimens (%)

Parameter CC CIN-1 CIN-2/CIN-3 SCC Pa

Total no. of specimens 12 15 31 5

p16 StainingPositive cells

�10 12 (100) 9 (60)b 0 (0)c,e 0 (0)d,f <.001

>10 0 (0) 6 (40) 31 (100) 5 (100)

Intensity

Weak 12 (100) 14 (93) 16 (52)c,e 0 (0)d,f <.001

Strong 0 (0) 1 (7) 15 (48) 5 (100)

Full-thickness epithelial staining

Absent 12 (100) 15 (100) 23 (74) 0 (0)d,f,g <.001

Present 0 (0) 0 (0) 8 (26) 5 (100)

Ki-67 stainingPositive cells

�10 12 (100) 9 (60) 4 (13)c,e 0 (0)d <.001

>10 0 (0) 6 (40) 27 (87) 5 (100)

Intensity

Weak 12 (100) 12 (80) 19 (61) 3 (60) .034

Strong 0 (0) 3 (20) 12 (39) 2 (40)

Full-thickness epithelial staining

Absent 12 (100) 15 (100) 25 (81) 4 (80) .101

Present 0 (0) 0 (0) 6 (19) 1 (20)

HPV L1 capsid protein stainingPositive 0 (0) 6 (40) 6 (19) 0 (0) .047

Negative 12 (100) 9 (60) 25 (81) 5 (100)

CC indicates chronic cervicitis; CIN, cervical intraepithelial neoplasia; SCC, squamous cell carcinoma; HPV, human papillomavirus.aP values indicate the statistical significance of relations between groups and parameters.bThe difference between CC and CIN-1 was statistically significant.cThe difference between CC and CIN-2/CIN-3 was statistically significant.dThe difference between CC and SCC was statistically significant.eThe difference between CIN-1 and CIN-2/CIN-3 was statistically significant.fThe difference between CIN-1 and SCC was statistically significant.gThe difference between CIN-2/CIN-3 and SCC was statistically significant.

Table 4. Combined Expression of p16 and Human Papillomavirus L1 Capsid Protein in CervicalIntraepithelial Neoplasia and Squamous Cell Carcinoma

No. of Specimens (%)

Diagnosis L12/p161 L11/p161 L11/p162 L12/p162

CIN-1, n¼15 2 (13) 4 (27) 3 (20) 6 (40)

CIN-2/CIN-3, n¼31 25 (81) 6 (19) 0 (0) 0 (0)

SCC, n¼5 5 (100) 0 (0) 0 (0) 0 (0)

LI indicates human papillomavirus L1 capsid protein; �, negative; p16, p16INK4A; þ, positive; CIN, cervical intraepithelial

neoplasia; SCC, squamous cell carcinoma.

Original Article


DISCUSSIONIncreasing the accuracy of cytology diagnosis of cervicallesions, especially for HSIL and SCC, is pivotal for theprevention and therapy of cervical tumors. In the currentstudy, of 5 SCC specimens that had tissue follow-up avail-able, 3 specimens were interpreted in LBC, and 1 morespecimen was diagnosed on cell blocks. In addition,64.5% of specimens that were categorized as CIN-2/CIN-3 originally were interpreted as HSIL on LBC, andthe accuracy rate increased further to 83.9% with cellblock diagnoses. However, diagnosis agreement rates withtissue follow-up for CIN-1 did not increase on cell blockpreparations compared with LBC, because only 8 in 15CIN-1 specimens were diagnosed as LSIL with both LBCand cell blocks. It is likely that cell block assessments forSCC and HSIL dovetailed better with tissue diagnosesthan LBC assessments, although more specimens need tobe evaluated.

The detection of both p16 and Ki-67 is a good indi-cator of high-grade CIN.11,12 Immunocytochemistry forp16 and Ki-67 also has been performed in ThinPrepsmears26 and cell block preparations,23,27 and data fromthose studies support the finding that p16 and Ki-67 areuseful as markers for the diagnosis of CIN on cellblocks.23,27 In the current study, greater immunoreactiv-ity for p16 and Ki-67 in HSIL and SCC specimens wasdetected in cell block sections, consistent with the resultsfrom tissue sections in our previous study28 and suggest-ing that p16 and Ki-67 immunostaining on a cell blockcould improve the accuracy of detecting HSIL and SCC.The intensity of p16 staining on a cell block also is an im-portant index for the determination of high-grade CIN.Consistent with this, strong staining for p16 was observedin the cell block preparations from all SCC sections andfrom nearly half of the CIN-2/CIN-3 sections (48%) inthe current study. Strong, full-thickness epithelial stainingwas observed only in the HSIL and SCC specimens,which can serve as a valuable indicator of high-grade cervi-cal lesions.29,30 Either weak staining was observed or lesspositive cells were exhibited in the cell block preparationsfrom low-grade CIN.

Because the majority of CIN-1 lesions and a consid-erable number of CIN-2 lesions regress spontaneously,the ability to predict the development of CIN also is animportant issue for cervical cancer prevention and treat-ment.31 The L1 capsid protein is the major target for thecell-mediated immune response to HPV infections andnormally is detectable during the productive stage ofHPV disease. It is produced abundantly in mild-to-mod-

erate dysplasia but rarely is observed in nonsuspicious Papsmears or severe dysplasia, and it is not produced in carci-nomas.21,32,33 The L1 capsid protein reportedly was posi-tive in 30% of LSILs, 12% of HSILs, and 0% of SCCsin LBC31 and in 43.7% of LSILs and 33.3% HSILs incervical smears.32 Our results are similar to those fromprevious studies. It was reported that p16 and L1 immu-nohistochemistry could be helpful for estimating the bio-logic potential of low-grade squamous cervical lesions.Particularly for sections in which the grade of the lesion isdifficult to assess morphologically, the L1/p16 expressionpattern could be useful for planning the clinical manage-ment of women with these lesions.34 The staining of L1/p16 can be divided into the following 4 patterns: L1-nega-tive/p16-negative, L1-positive/p16-negative, L1-positive/p16-positive, and L1-negative/p16-positive, with eachpattern representing different biologic significance.35 L1-negative indicates latent viral infection or integration ofHPV DNA into the host genome; L1-positive indicatesthat viral DNA is present as a productive form; and p16-positive and p16-negative indicate whether or not the cellcycle is disrupted, respectively. If an HPV DNA test ispositive, then the L1-negative/p16-negative and L1-posi-tive/p16-negative patterns indicate that the lesion is in anondysplastic state but without or with producing the vi-rus, respectively. The L1-positive/p16-positive and L1-negative/p16-positive patterns indicate that the lesion isin an early or advanced dysplastic phase, respectively.35

Thus, the sequence of the L1/p16 expression status couldchange in the order of L1-negative/p16-negative, L1-posi-tive/p16-negative, L1-positive/p16-positive, and L1-neg-ative/p16-positive with increasing severity of the cervicallesion.35 In our study, the L1-negative/p16-positive pat-tern was observed mainly in SCC and CIN-2/CIN-3 sam-ples (Table 4), whereas the L1-positive/p16-positivepattern was observed in the minority of CIN-1 and CIN-2/CIN-3 samples. The L1-positive/p16-negative and L1-negative/p16-negative pattern existed only in portions ofsome CIN-1 samples. Thus, the L1/p16 expression pat-terns are related to the severity of cervical lesions and mayserve as a valuable index for predicting prognosis anddetermining a follow-up strategy for dysplastic lesions ofthe cervix, as proposed by previous studies.34

Although tissue diagnosis is considered the ‘‘goldstandard’’ for CIN, it may not reflect the true disease statein all specimens because of the fallibility of colposcopy.Thus, the cytologic approach plays an important role inidentifying and diagnosing cervical neoplasia. In the cur-rent study, we compared the diagnoses in cell block



preparations with diagnosis from LBC and tissue sectionreports; performed immunocytochemistry for p16, Ki-67,and the HPV L1 capsid protein on cell block sections; andanalyzed the correlation of the expression of these bio-markers with the severity of cervical lesions. Our datademonstrated that the cell block technique may be com-plementary to LBC and that combining LBC with cellblocks may improve the diagnostic accuracy of cytology.In addition, cell blocks can be used for sequential section-ing and immunostaining. Immunostaining for p16 andKi-67 (and especially p16) on a cell block preparation ishelpful for improving the diagnostic accuracy of HSILand SCC. Detection of the combined expressions of p16and HPV L1 can provide valuable information for pre-dicting prognosis and determining follow-up schedulesfor women who have cervical dysplastic lesions. More-over, cell blocks in our study were prepared from cell pel-lets and plasma with CaCl2 (instead of the commonlyused thrombin), which is economic and suitable for gen-eral medical units.

CONFLICT OF INTEREST DISCLOSURESSupported by grants from Yuexiu Science and Technology Bu-reau of Guangzhou, China (2005-WS-001 and 2008-WS-003).

REFERENCES1. Parkin DM, Pisani P, Ferlay J. Estimates of the worldwide

incidence of 25 major cancers in 1990. Int J Cancer. 1999;80:827-841.

2. Koss LG. The Papanicolaou test for cervical cancer detec-tion. A triumph and a tragedy. JAMA. 1989;261:737-743.

3. DeMay R. The Pap smear. In: DeMay R, ed. The Art andScience of Cytopathology. Chicago, Ill: American Society ofClinical Pathologists Press;1996:61-205.

4. Nanda K, McCrory DC, Myers ER, et al. Accuracy of thePapanicolaou test in screening for and follow-up of cervicalcytologic abnormalities: a systematic review. Ann InternMed. 2000;132:810-819.

5. Klinkhamer PJ, Meerding WJ, Rosier PF, Hanselaar AG.Liquid-based cervical cytology. Cancer. 2003;99:263-271.

6. Richard K, Dziura B, Hornish A. Cell block preparation asa diagnostic technique complementary to fluid-based mono-layer cervicovaginal specimens. Acta Cytol. 1999;43:69-73.

7. Yeoh GP, Chan KW. Cell block preparation on residualThinPrep sample. Diagn Cytopathol. 1999;21:427-431.

8. Keyhani-Rofagha S, Vesey-Shecket M. Diagnostic value, fea-sibility, and validity of preparing cell blocks from fluid-based gynecologic cytology specimens. Cancer. 2002;96:204-209.

9. Ozbun MA, Meyers C. Temporal usage of multiple pro-moters during the life cycle of human papillomavirus type31b. J Virol. 1998;72:2715-2722.

10. Doorbar J. The papillomavirus life cycle. J Clin Virol.2005;32(suppl 1):S7-S15.

11. Halloush RA, Akpolat I, Jim Zhai Q, Schwartz MR, ModyDR. Comparison of ProEx C with p16INK4a and Ki-67immunohistochemical staining of cell blocks prepared fromresidual liquid-based cervicovaginal material: a pilot study.Cancer. 2008;114:474-480.

12. Kalof AN, Cooper K. p16INK4a immunoexpression: surro-gate marker of high-risk HPV and high-grade cervical intra-epithelial neoplasia. Adv Anat Pathol. 2006;13:190-194.

13. Leversha MA, Fielding P, Watson S, Gosney JR, Field JK.Expression of p53, pRB, and p16 in lung tumours: a valida-tion study on tissue microarrays. J Pathol. 2003;200:610-619.

14. Makitie AA, MacMillan C, Ho J, et al. Loss of p16 expres-sion has prognostic significance in human nasopharyngealcarcinoma. Clin Cancer Res. 2003;9:2177-2184.

15. Klaes R, Woerner SM, Ridder R, et al. Detection of high-risk cervical intraepithelial neoplasia and cervical cancer byamplification of transcripts derived from integrated papillo-mavirus oncogenes. Cancer Res. 1999;59:6132-6136.

16. Klaes R, Benner A, Friedrich T, et al. p16INK4a immuno-histochemistry improves interobserver agreement in the diag-nosis of cervical intraepithelial neoplasia. Am J Surg Pathol.2002;26:1389-1399.

17. Kruse AJ, Baak JP, de Bruin PC, et al. Ki-67 immunoquan-titation in cervical intraepithelial neoplasia (CIN): a sensitivemarker for grading. J Pathol. 2001;193:48-54.

18. Nam EJ, Kim JW, Hong JW, et al. Expression of p16 andKi-67 in relation to the grade of cervical intraepithelial neo-plasia and high-risk human papillomavirus infection. J Gyne-col Oncol. 2008;19:162-168.

19. Aoyama C, Liu P, Ostrzega N, Holschneider CH. Histo-logic and immunohistochemical characteristics of neoplas-tic and nonneoplastic subgroups of atypical squamouslesions of the uterine cervix. Am J Clin Pathol. 2005;123:699-706.

20. Griesser H, Sander H, Hilfrich R, Moser B, Schenck U.Correlation of immunochemical detection of HPV L1 cap-sid protein in Pap smears with regression of high-risk HPVpositive mild/moderate dysplasia. Anal Quant Cytol Histol.2004;26:241-245.

21. Rauber D, Mehlhorn G, Fasching PA, Beckmann MW,Ackermann S. Prognostic significance of the detection ofhuman papilloma virus L1 protein in smears of mild tomoderate cervical intraepithelial lesions. Eur J Obstet GynecolReprod Biol. 2008;140:258-262.

22. Saqi A, Pasha TL, McGrath CM, Yu GH, Zhang P, GuptaP. Overexpression of p16INK4A in liquid-based specimens(SurePath) as marker of cervical dysplasia and neoplasia.Diagn Cytopathol. 2002;27:365-370.

23. Liu H, Shi J, Wilkerson M, et al. Immunohistochemicaldetection of p16INK4a in liquid-based cytology specimenson cell block sections. Cancer. 2007;111:74-82.

24. Duncan L, Jacob S, Hubbard E. Evaluation of p16INK4aas a diagnostic tool in the triage of Pap smears demonstrat-ing atypical squamous cells of undetermined significance.Cancer. 2008;114:34-48.

25. Keating JT, Cviko A, Riethdorf S, et al. Ki-67, cyclin E,and p16INK4 are complimentary surrogate biomarkers forhuman papilloma virus-related cervical neoplasia. Am J SurgPathol. 2001;25:884-891.

26. Murphy N, Ring M, Killalea AG, et al. p16INK4A as amarker for cervical dyskaryosis: CIN and cGIN in cervicalbiopsies and ThinPrep smears. J Clin Pathol. 2003;56:56-63.

Original Article


27. Akpolat I, Smith DA, Ramzy I, Chirala M, Mody DR. Theutility of p16INK4a and Ki-67 staining on cell blocks pre-pared from residual thin-layer cervicovaginal material. Can-cer. 2004;102:142-149.

28. Yu L, Peng J, Chen S. Expression of p16INK4A and Ki-67in cervical intraepithelial neoplasia. J Clin Exp Pathol. 2007;23:665-668.

29. Klaes R, Friedrich T, Spitkovsky D, et al. Overexpression ofp16(INK4A) as a specific marker for dysplastic and neoplasticepithelial cells of the cervix uteri. Int J Cancer. 2001;92:276-284.

30. Branca M, Ciotti M, Santini D, et al. p16(INK4A) expres-sion is related to grade of cin and high-risk human papillo-mavirus but does not predict virus clearance after conizationor disease outcome. Int J Gynecol Pathol. 2004;23:354-365.

31. Ostor AG. Natural history of cervical intraepithelial neopla-sia: a critical review. Int J Gynecol Pathol. 1993;12:186-192.

32. Melsheimer P, Kaul S, Dobeck S, Bastert G. Immunocyto-chemical detection of HPV high-risk type L1 capsid pro-teins in LSIL and HSIL as compared with detection ofHPV L1 DNA. Acta Cytol. 2003;47:124-128.

33. Birner P, Bachtiary B, Dreier B, et al. Signal-amplified col-orimetric in situ hybridization for assessment of human pap-illomavirus infection in cervical lesions. Mod Pathol.2001;14:702-709.

34. Negri G, Bellisano G, Zannoni GF, et al. p16INK4a andHPV L1 immunohistochemistry is helpful for estimatingthe behavior of low-grade dysplastic lesions of the cervixuteri. Am J Surg Pathol. 2008;32:1715-1720.

35. Yoshida T, Sano T, Kanuma T, et al. Immunochemicalanalysis of HPV L1 capsid protein and p16 protein in liq-uid-based cytology samples from uterine cervical lesions.Cancer. 2008;114:83-88.



Download - Diagnostic value of p16INK4A, Ki-67, and human papillomavirus L1 capsid protein immunochemical staining on cell blocks from residual liquid-based gynecologic cytology specimens

Top Related