EuropeanUrology European Urology 43 (2003) 609–614

The Role of the Complexed-to-Total Prostate-SpecificAntigen Ratio in Predicting the Final Pathological Stageof Clinically Localized Prostate CancerManuel Martıneza, Silvia Navarroc, Pilar Medinac, Humberto Villavicenciod,Luis Riojae, Eduardo Solsonaf, Amparo Estellesc, Justo Aznarb,J. Fernando Jimenez-Cruza, Francisco Espanac,*

aDepartment of Urology, La Fe University Hospital, Valencia, SpainbDepartment of Clinical Pathology, La Fe University Hospital, Valencia, SpaincResearch Center, La Fe University Hospital, Av. Campanar, 21, 46009 Valencia, SpaindFundacion Puigvert, Barcelona, SpaineHospital Miguel Servet, Zaragoza, SpainfInstituto Valenciano de Oncologıa, Valencia, Spain

Accepted 20 March 2003

AbstractObjectives: To examine the association between the complexed-to-total (C:T) prostate-specific antigen (PSA) ratioand prostate cancer pathological stage to assess whether the C:T PSA ratio may predict the final pathological stage inpatients with clinically localized prostate cancer.Patients and Methods: In a prospective study, 101 men with clinically localized prostate cancer underwent a stagingpelvic lymphadenectomy and radical prostatectomy. Total PSA (tPSA) and PSA complexed to a1-antichymotrypsin(cPSA) were measured from preoperative plasma and were correlated with the clinical and pathological stage, andwith surgical margin status. The pathological stage was determined as organ-confined (n ¼ 59) and extracapsularextension (n ¼ 42).Results: The distributions of tPSA and cPSAwere significantly different in men with locally confined and those withlocally extended disease. This finding was not observed for the C:T PSA ratio. The area under the receiver operatingcharacteristic (ROC) curve to predict the final pathological stage was significantly greater for tPSA (0.684) andcPSA (0.677) than for the C:T PSA ratio ( p < 0:032). TPSA (0.685) and cPSA (0.670) also showed areas under theROC curve greater than that of the C:T PSA ratio (0.542) ( p < 0:05) for prediction of positive surgical margins.Conclusions: Our results show that the C:T PSA ratio does not improve the performance of total PSA for predictingthe final pathological stage in patients with clinically localized prostate cancer.# 2003 Elsevier Science B.V. All rights reserved.

Keywords: Complexed prostate-specific antigen; Prostate cancer staging; Plasma

1. Introduction

Prostate-specific antigen (PSA) is the best tumormarker presently available and detects a significantproportion of organ-confined prostate cancers [1]. How-ever, PSA is of limited help in differentiating between

clinically important and unimportant carcinomas, con-fined versus extended disease, and in predicting thepathological stage [2]. Several attempts have been madeto enhance the differentiation between various gradesand stages of prostate cancer, and it has been shown thatthe combination of total PSA, clinical stage and Gleasonscore can improve preoperative staging [3].

PSA circulates in several molecular forms, includingfree PSA and PSA complexed to a1-antichymotrypsin(a1ACT) (complexed PSA) and a2-macroglobulin

* Corresponding author. Tel.: þ34-96-386-2797; Fax: þ34-96-197-3018.

E-mail address: espanya_fra@gva.es (F. Espana).

0302-2838/03/$ – see front matter # 2003 Elsevier Science B.V. All rights reserved.

doi:10.1016/S0302-2838(03)00149-0

[4,5]. Several studies have demonstrated that the pro-portion of free or complexed PSA enhances the clinicalutility of total PSA in the early detection of prostatecancer by improving the differentiation between pros-tate cancer and benign conditions [6–14]. However,conflicting results have been reported regarding theusefulness of the free-to-total (F:T) PSA ratio in theprediction of pathological stage [15–17]. Moreover,PSA:a1ACT complex is more stable than free PSA[18], and we have demonstrated that the C:T PSA ratioperformed better than the F:T PSA ratio in the differ-entiation between prostate cancer and benign prostatichyperplasia [9,13,19]. However, no data are availableon the usefulness of the C:T PSA ratio in the predictionof pathological stage.

The objective of the present study was to determinewhether the C:T PSA ratio could be utilized to betterdefine who would best benefit from a radical prosta-tectomy.

2. Patients andmethods

2.1. Sample collectionPlasma samples were obtained from 101 patients with clinically

localized prostate cancer before undergoing radical prostatectomy.All the patients were men with clinical complaints who had beenreferred to an urologist for prostatic evaluation. From January toApril 1999, consecutive samples were obtained from four differentinstitutions (42 from Hospital La Fe, Valencia, 14 from InstitutoValenciano de Oncologıa, Valencia; 25 from Fundacion Puigvert,Barcelona, and 20 from Hospital Miguel Servet, Zaragoza, Spain).The average age was 64 years (range 51–74). In all the men, prostatecancer was confirmed by sextant needle biopsies of the prostateexcept three men who were diagnosed after TURP. The clinicalstage was assessed by digital rectal examination and transrectalechography, according to the 1997 TNM classification (Table 1).Inclusion criteria were PSA less than 40 ng/ml, a negative abdom-inal computerized tomography and bone scan, and the absence ofany preoperative hormonal or radiation therapy. All the patientsunderwent pelvic lymph node dissection and radical retropubic

prostatectomy. Unfortunately, tumor volume was not determinedin the majority of patients and was not included in the study.

All the patient samples were obtained before any surgical orhormonal therapies or urethral or prostatic manipulation. Theprocedures followed were in accordance with the Helsinki Declara-tion of 1975. Blood samples were collected by venipuncture in0.13 M sodium citrate (1:9, citrate:blood, v:v) and centrifuged at1,500g for 30 min. Plasma was frozen in small portions, stored at�80 8C for not more than 6 months and thawed immediately beforeanalysis. Each aliquot was thawed only once.

2.2. PathologyUsing standard pathological techniques, radical prostatectomy

specimens and lymph nodes removed were examined. Pathologicstage was determined according to the 1997 TNM classification.Invasion into the prostate apex or into (but not beyond) the prostaticcapsule was not classified as T3, but as T2. Favorable pathology (FP)was defined as either organ-confined or extraprostatic extensionwith negative surgical margins and a Gleason score <7. Unfavorablepathology (UP) was defined as either extraprostatic extension withpositive surgical margins, and/or Gleason score �7, and/or seminalvesicle involvement, and/or lymph node involvement.

2.3. AssaysThe total plasma PSA concentration and the concentration of

PSA complexed to a1ACT were determined by specific sandwichELISAs as reported previously [9,19]. Although the assay forcomplexed PSA used in the present work is not commerciallyavailable, a commercially available assay using a different meth-odology to measure complexed PSA has been reported [20].

2.4. Statistical analysisIn the Box-and-whisker plot (Fig. 1), the central box represents

the values from the lower to upper quartile (25 to 75 percentile), themiddle line represents the median, and the vertical line extendsfrom the minimum to the maximum value, excluding outsides(smaller than the lower quartile minus 1.5 times the interquartilerange, or larger than the upper quartile plus 1.5 times the inter-quartile range) values which are displayed as separate points.

Student and paired t-tests, Mann–Whitney mean nonparametricU-test and the area under the receiver operating characteristic

Table1Clinical stage and Gleason score distribution in 101 men with clinically

localized prostate cancer

Patients (n)

Clinical stage

T1b 3

T1c 34

T2a 56

T2b 8

Gleason score

2–4 26

5–6 34

7 35

8–10 6Fig. 1.

610 M. Martınez et al. / European Urology 43 (2003) 609–614

(ROC) curves were done using a computer program for medicalstatistics [21]. The predictive value of the C:T PSA ratio wasanalyzed with a logistic regression model which included totalPSA, clinical stage and Gleason, using the statistical package SPSS6.0. A value of p < 0:05 was considered statistically significant.

3. Results

Table 1 shows the distribution of the preoperativeclinical stages and Gleason score. The histologicalstudy of the prostatectomy specimen showed that 59patients (58.4%) had an organ-confined prostate cancer(pT2) and 42 (41.6%) had extracapsular extension(>pT2). The histological analysis also showed positivesurgical margins in 49 patients (48.5%), 13 of themwith an organ-confined tumor.

As can be seen in Table 2, total PSA and PSA:a1ACTcomplex levels showed statistically significant differ-ences in relation to the pathological stage and surgicalmargin involvement. However, the C:T PSA ratio wassimilar in all cases.

In order to determine the utility of using complexedPSA and/or the C:T PSA ratio for the prediction ofpathological stage, we compared these parameters withthe established combination of total PSA with Gleasonscore and clinical stage in a logistic regression analysis.There was no significant additional benefit in usingcomplexed PSA or C:T PSA ratio to predict thepathological stage. For the prediction of favorable aswell as unfavorable pathology, complexed PSA or C:TPSA ratio provided similar prediction when comparedwith total PSA.

When the ROC curves were plotted to predict the finalpathological stage, total PSAwas the parameter with thehighest (0.684) area under the curve (AUC), followed byPSA:a1ACT (0.677, p ¼ 0:658), and both showed sta-tistically significant differences with respect to the C:TPSA ratio (0.522; p ¼ 0:031 and p ¼ 0:02, respec-tively). Of the three diagnostic methods for evaluatingsurgical margin involvement, total PSA had the highestAUC (0.685) but the differences between it and

PSA:a1ACT (0.670; p ¼ 0:37) were not significant.The AUC of the C:T PSA ratio (0.542), however, wassignificantly lower than those of total PSA ( p ¼ 0:046)and PSA:a1ACT complex ( p ¼ 0:041). The distribu-tion, therefore, according to tumor stage and surgicalmargin involvement was 46 patients (45.5%) with alocalized tumor and with negative surgical margin, 13(12.8%) with organ-confined prostate cancer but prob-able incisional margins, 36 (35.6%) with extracapsularextension and positive surgical margins, and six (5.9%)with a locally advanced tumor but with negative surgicalmargins, three of them with a Gleason score < 7.Therefore, according to the established prognosisclassification, 62 patients fulfilled favorable pathologycriteria and 39 had unfavorable pathology criteria.

When we compared the different groups, the totalPSA and PSA:a1ACT complex levels showed statisti-cally significant variations between the patients withfavorable pathology and those with unfavorable pathol-ogy (Table 3). It was again impossible to differentiatebetween these two groups on the basis of the C:T PSAratio. When the ROC curves for the diagnosis of thepatients with favorable pathology were plotted, totalPSA was the parameter with the highest AUC (0.696),followed by the PSA:a1ACT complex (0.686,p ¼ 0:49), and both differed from the C:T PSA ratiovalue (0.518) in a statistically significant manner( p ¼ 0:014 and p ¼ 0:010, respectively).

Table 2Total PSA (T), PSA:a1ACT complex (C) and C:T PSA ratio in patients with prostate cancer who were going to undergo radical prostatectomy, according to

their pathological stage and surgical margin involvement

Assay Pathological stage Surgical margins

Localized (�pT2)

(n ¼ 59)

Extracapsular (�pT3)

(n ¼ 42)

p Negative margins

(n ¼ 52)

Positive margins

(n ¼ 49)

p

Total PSA (T) (ng/ml) 5.90 (7.25 � 5.60) 9.70 (10.68 � 7.85) 0.002 5.20 (6.90 � 4.86) 9.40 (11.01 � 8.57) 0.001

PSA:a1ACT (C) (ng/ml) 4.84 (6.44 � 5.33) 7.90 (9.44 � 7.24) 0.003 4.43 (6.23 � 4.81) 7.80 (9.60 � 7.74) 0.003

C:T PSA ratio 0.86 (0.86 � 0.10) 0.87 (0.87 � 0.09) 0.725 0.88 (0.87 � 0.10) 0.85 (0.86 � 0.09) 0.581

The values are expressed as median with the mean � S:D: in parentheses.

Table 3Total PSA (T), PSA:a1ACT complex (C) and C:T PSA ratio in patients

with prostate cancer before undergoing radical prostatectomy

FP (n ¼ 62) UP (n ¼ 39) p

Total PSA

(ng/ml) (T)

5.92 (6.93 � 4.61) 10.0 (11.52 � 8.86) 0.001

PSA:a1ACT

(ng/ml) (C)

4.92 (6.17 � 4.52) 7.80 (10.10 � 8.01) 0.002

C:T PSA ratio 0.86 (0.86 � 0.10) 0.85 (0.86 � 0.09) 0.873

The groups of patients with favorable (FP) and unfavorable pathology (UP)

are compared. The values are expressed as median with the mean � S:D: in

parentheses.

M. Martınez et al. / European Urology 43 (2003) 609–614 611

To evaluate the significance of the differences in totalPSA and PSA:a1ACT complex values between thepatients with favorable and unfavorable pathology,we determined the specificity and sensitivity thatwould be obtained with different values of each ofthese parameters. As can be seen in Table 4, a totalPSA � 26 ng/ml would have given a correct diagnosisof all the patients with a good prognosis (100% sensi-tivity), and avoiding radical prostatectomy in fivepatients with an unfavorable histological study. If wedecrease the cutoff value to �16 ng/ml, sensitivity fallsto 96.8% with a specificity of 16.2%. Our results withPSA:a1ACT complex can almost be superimposed onthose obtained with total PSA, and in both cases areclearly better than those obtained with the C:T PSAratio.

4. Discussion

The finding that PSA is present in plasma in differentmolecular forms [4,5] and that by making use of themthe diagnostic performance of total PSA can beimproved [6–14] has led to increased interest in thismarker. The difficulties involved in PSA:a1ACT com-plex, together with the development of commercialassays for quantifying free PSA have encouraged manyresearchers to focus on the F:T PSA ratio [10,11,15–17,22–25]. In synthesis and almost unanimously thepublished reports indicate that the basic advantage ofF:T PSA ratio over total PSA lies in the improvedspecificity without any apparent loss of sensitivity.Together this means a decrease in the number ofprostate biopsies done unnecessarily. However, there

are theoretical reasons for using PSA:a1ACT complexand its ratio. On the one hand, it constitutes the mostabundant fraction in plasma, and therefore allows moreprecise quantification, especially in patients with lowtotal PSA levels in whom only a minimal fraction ofPSA is in the free form and most of it is complexed witha1ACT. On the other hand, PSA:a1ACT complex ismore stable than free PSA, although the stability of thelatter improves when plasma is used instead serum[25]. We recently demonstrated this in a study in whichwe compared PSA:a1ACT complex and free PSA, andexamined the effects of using plasma instead of serum[19]. The findings indicated that the C:T PSA ratioperformed better than the F:T PSA ratio both in plasmaand serum samples, and that the best diagnosis wasobtained using the C:T PSA ratio in plasma.

Once the efficacy of the C:T PSA ratio in diagnosingprostate cancer had been demonstrated, the next stepwas to evaluate its usefulness as a prognostic factor andespecially its ability for predicting the pathologicalstage of patients with clinically localized prostatecancer. In men who underwent radical prostatectomy,the probability of long-term, disease-free survivalrelates directly to pathological stage at the time ofsurgery [25]. Multivariate models integrating variablesthat can be derived from clinical and pathologicalassessment have been shown to be reliable and usefulin urology practice. Among these variables, the com-bination of clinical stage, serum PSA, and biopsyGleason score provides reliable assessment of the riskfor extraprostatic disease that can be used readily forcounseling individual patients [26]. Furthermore,pathological stage is the most reliable means of pre-dicting outcome of definite treatment in men withclinically localized prostate cancer. It is also the mostobjective and reliable mean of defining the anatomicprogression of the tumor. Pound et al. have shown thatgrouping patients according to a combination of patho-logical stage, Gleason score, and surgical margin statusprovides the best stratification of actuarial progression-free probability [27]. Logistic regression analysis ofour data clearly indicates the scant predictive value ofthe C:T PSA ratio; it does not reveal significant differ-ences between patients with organ-confined and extra-capsular extension, with positive or negative surgicalmargins or with a good or bad prognosis. Furthermore,in all cases the AUC is about 0.5, which demonstratethe poor prediction performance. However, thePSA:a1ACT complex does, by itself, show significantdifferences between the different groups, and has anAUC slightly lower than that of total PSA. Never-theless, the fact that the total PSA and PSA:a1ACTcomplex levels in the patients with the best prognosis

Table 4Sensitivity and specificity of different total PSA (T), PSA:a1ACT (C) and

C:T PSA ratio values for predicting which patients constitute the group

with favorable pathology

Cutoff value Sensitivity (%) Specificity (%)

Total PSA (T) (ng/ml) �2.45 11.3 100

�4.1 29 94.6

�16 96.8 16.2

�26 100 10.2

PSA:a1ACT(C) (ng/ml) �1.6 9.7 100

�3.54 25.8 96.3

�14.85 95.2 17.9

�27 100 8.1

C:T PSA ratio �0.65 4.8 100

�0.69 9.7 94.6

�0.99 82.3 21.6

�1.00 100 0

The values that give sensitivities and specificities around 95% and 100%

were chosen.

612 M. Martınez et al. / European Urology 43 (2003) 609–614

and in the other patients differ significantly does notnecessarily mean that they are clinically useful para-meters. In this case, the substantial overlapping of thevalues makes it difficult to define a cutoff value withwhich we could identify the patients who did not needto undergo radical surgery. For example, a total PSAlevel �26 ng/ml would have permitted correct diag-nosis of all the patients with favorable prognosticfactors (100% sensitivity) but would have obviatedradical prostatectomy in only four patients (10.2%specificity). With the PSA:a1ACT complex the speci-ficity is slightly lower for 100% sensitivity but almostidentical for 95.7% sensitivity. If a PSA:a1ACT com-plex level �14.85 ng/ml were used as the criterion forradical prostatectomy, it would make surgery unneces-sary in seven of the 39 patients with unfavorablecriteria, but, in contrast, would obviate a potentiallycurative radical prostatectomy in three out of the 62patients with a good prognosis. Therefore, thePSA:a1ACT complex and especially its ratio do notimprove upon the predictive capacity of total PSA inpatients undergoing radical prostatectomy.

This is the first report on the usefulness of the C:TPSA ratio for predicting the pathological stage inpatients who undergo radical prostatectomy. Withrespect to the F:T PSA ratio, most authors agree thatit does not provide more information than total PSA asto the final pathological [15,17,28]. However, Arcan-geli et al. [16] found a significant correlation betweenlow percentage of free PSA (below 14%) and poten-tially more aggressive tumor, or in other words withextracapsular extension, positive surgical margins,higher Gleason score and larger volume. Along thesame line, Morote et al. [29] concluded after logistic

regression analysis of a series of 92 patients subjectedto radical prostatectomy that the F:T PSA ratio is theonly factor capable of predicting potential surgicalfailure, concept introduced by Katz et al. [30] thatestablishes as the group of patients with adversepathology that including all cases with positive surgicalmargins or infiltration of seminal vesicles or lymphnode involvement. More specifically, these authorsfound that a combination of a serum concentrationof total PSA <10 ng/ml and a F:T PSA ratio >0.10makes it possible to detect a subgroup of patients with agood prognosis and a probability of less than 10% ofhaving potential surgical failure. However, as theseauthors acknowledge, the prognostic value of the F:TPSA ratio is not clearly defined.

In conclusion, PSA:a1ACT complex and its ratiodo not improve the ability of total PSA to predict thefinal pathological stage in patients with clinicallylocalized prostate cancer who are going to undergoradical prostatectomy. Further study is needed to ascer-tain whether the molecular forms of PSA provideadditional information to that gathered from the clin-ical stage, Gleason score and total PSA. Quantificationof these forms is not, therefore, at present to berecommended.

Acknowledgements

Supported in part by research grants 01/1148,PI020125, and PI020136 from the Plan Nacional deInvestigacion Cientıfica, Desarrollo e InnovacionTechnologica (Instituto de Salud Carlos III, Fondode Investigacion Sanitaria, Spain).

References

[1] Catalona WJ, Smith DS, Ratliff TL, Basler JW. Detection of organ-

confined cancer is increased through prostate-specific antigen based

screening. JAMA 1993;270:948–54.

[2] Epstein JI, Walsh PC, Carmichael M, Brendler CB. Pathological and

clinical findings to predict tumor extent of nonpalpable (T1c)

prostate cancer. JAMA 1994;271:368–74.

[3] Partin AW, Kattan MW, Subong EN, Walsh PC, Wojno KJ,

Oesterling JE, et al. Combination of prostate specific antigen,

clinical stage and Gleason score to predict pathological stage of

localized prostate cancer. JAMA 1997;277:1445–51.

[4] Lilja H, Christensson A, Dahlen U, Matikainen M-T, Nilsson O,

Petterson K, et al. Prostate specific antigen in serum occurs

predominantly in complex with a1-antichymotrypsin. Clin Chem

1991;37:1618–25.

[5] Heeb MJ, Espana F, Gittes R, Griffin JH. Prostate specific antigen-

a2-macroglobulin complexes in prostate cancer patient sera.

Biochem Mol Biol Internat 1995;37:917–23.

[6] Stenman UH, Leinonen J, Alfthan H, Rannikko S, Tuhkanen K,

Alfthan O. A complex between prostate-specific antigen and

a1-antichymotrypsin is the mayor form of prostate-specific antigen

in serum of patients with prostate cancer: assay of the complex

improves clinical sensitivity for cancer. Cancer Res 1991;51:

222–6.

[7] Wood WG, van der Sloot E, Bohle A. The establishment and

evaluation of luminescent-labelled immunometric assays for pros-

tate-specific antigen-a1-antichymotrypsin complexes in serum. Eur J

Clin Chem Clin Biochem 1991;29:787–94.

[8] Christensson A, Bjork T, Nilsson O, Dahlen U, Matikainen M,

Cockett ATK, et al. Serum prostate antigen complexed to a1-

antichymotrypsin as an indicator of prostate cancer. J Urol 1993;150:

100–5.

[9] Espana F, Martınez M, Sanchez-Cuenca J, Vera CD, Estelles A,

Jimenez-Cruz JF. Prostate-specific antigen and its complexes with

a1-antichymotrypsin in the plasma of patients with prostatic disease.

Eur Urol 1996;30:512–8.

[10] Luderer AA, Chen Y-T, Soriano TF, Kramp WJ, Carlson G, Cuny C,

et al. Measurement of the proportion of free to total prostate specific

antigen improves diagnostic performance of prostate specific antigen

M. Martınez et al. / European Urology 43 (2003) 609–614 613

in the diagnostic gray zone of total prostate specific antigen. Urology

1995;46:187–94.

[11] Catalona WJ, Smith DS, Wolfert RL, Wang TJ, Rittenhouse HG,

Ratliff TL, et al. Evaluation of percentage of free serum prostate

specific antigen to improve specificity of prostate cancer screening.

JAMA 1995;274:1214–20.

[12] Espana F, Martınez M, Royo M, Vera CD, Estelles A, Aznar J, et al.

Reference ranges for the concentrations of total and complexed

plasma prostate-specific antigen and their ratio in patients with

benign prostate hyperplasia. Eur Urol 1997;32:268–72.

[13] Martınez M, Espana F, Royo M, Vera CD, Estelles A, Jimenez-Cruz

JF, et al. Prostate-specific antigen complexed to a1-antitrypsin in the

early detection of prostate cancer. Eur Urol 2000;38:85–90.

[14] Martınez M, Espana F, Royo M, Alapont JM, Navarro S, Estelles A,

et al. The proportion of prostate-specific antigen (PSA) complexed to

a1-antichymotrypsin improves the discrimination between prostate

cancer and benign prostatic hyperplasia in men with a total PSA of

10 to 30 mg/l. Clin Chem 2002;48:1251–6.

[15] Pannek J, Subong EN, Jones KA, Marschake PL, Epstein JI, Chan

DW, et al. The role of free/total prostate-specific antigen ratio in the

prediction of final pathologic stage for men with clinically localized

prostate cancer. Urology 1996;48:51–4.

[16] Arcangeli CG, Shepherd DL, Smith DS, Humphrey PA, Keetch DW,

Catalona WJ. Correlation of percent free PSA with pathologic

features of prostatic carcinomas. J Urol 1996;155:415A.

[17] Bangma CH, Kranse R, Blijenberg BG, Schroeder FH. The free-to-

total serum prostate specific antigen ratio for staging prostate

carcinoma. J Urol 1997;157:544_547.

[18] Piironen T, Petterson K, Suonpaa M, Stenman HU, Oesterling JE,

Lovgren T, et al. In vitro stability of free prostate-specific antigen

(PSA) and prostate specific antigen (PSA) complexed to a1-

antichymotrypsin in blood samples. Urology 1996;48:81–7.

[19] Espana F, Royo M, Martınez M, Enguıdanos MJ, Vera CD, Estelles

A, et al. Free and complexed prostate specific antigen in the

differentiation of benign prostatic hyperplasia and prostate cancer:

studies in serum and plasma samples. J Urol 1998;160:2081–8.

[20] Allard WJ, Zhou Z, Yeung KK. Novel immunoassay for the

measurement of complexed prostate-specific antigen in serum. Clin

Chem 1998;44:1216–23.

[21] Schoojans F, Zalata A, Depuydt CE, Comhaire FH. MedCalc: a new

computer program for medical statistics. Comput Meth Progr

Biomed 1995;48:257–62.

[22] Prestigiacomo AF, Lilja H, Petterson K, Wolfert RL, Stamey TA. A

comparison of the free fraction of serum prostate specific antigen in

men with benign and cancerous prostates: the best case scenario. J

Urol 1996;156:350–4.

[23] Filella X, Alcover J, Molina R, Rodrıguez A, Carretero P, Ballesta

AM. Clinical evaluation of free PSA/total PSA (prostate-specific

antigen) ratio in the diagnosis of prostate cancer. Eur J Cancer

1997;33:1226–9.

[24] Ravery V, Boccongibod L. Free/total prostate-specific antigen ratio.

Hope and controversies. Eur Urol 1997;31:385–8.

[25] Pound CR, Partin AW, Eisenberger MA, Chan DW, Walsh PC.

Natural history of progression after PSA elevation following radical

prostatectomy. JAMA 1999;281:1591–7.

[26] Feneley MR, Partin AW. Indicators of pathologic stage of prostate

cancer and their use in clinical practice. Urol Clin North Am 2001;

28:443–57.

[27] Pound CR, Partin AW, Epstein JI, Walsh PC. Prostate-specific antigen

after anatomic radical retropubic prostatectomy: Patterns of recur-

rence and cancer control. Urol Clin North Am 1997;24:395–406.

[28] Graefen M, Hammerer P, Henke P, Hilz H, Huland E, Huland H.

Percentage of free PSA does not correlate with pathological

outcome. J Urol 1996;155:370A.

[29] Morote J, Raventos CX, Lorente JA, Encabo G, de Torres I, Lopez

MA, et al. Analisis del porcentaje de PSA lbre en la prediccion del

fallo quirurgico potencial en pacientes sometidos a prostatectomıa

radical. Actas Urol Esp 1999;23:127–34.

[30] Katz AE, de Vries GM, Benson MC, Buttyan RE, O’Toole K, Rubin

MA, et al. The role of the reverse transcriptase polymerase chain

reaction assay for prostate specific antigen in the selection of patients

for radical prostatectomy. Urol Clin North Am 1996;23:541–9.

614 M. Martınez et al. / European Urology 43 (2003) 609–614