PCA3 Score Before Radical Prostatectomy Predicts Extracapsular Extension and Tumor Volume
Article Outline
Purpose
PCA3 is a prostate specific, nonprotein coding RNA that is over expressed in prostate cancer. Recent studies showed the diagnostic potential of a urine based PCA3 for predicting biopsy outcome. We assessed the relationship between urine PCA3 and pathological features in whole mount radical prostatectomy specimens.
Materials and Methods
Post-digital rectal examination urine specimens were obtained from 72 men with prostate cancer before radical prostatectomy. PCA3 and PSA mRNA were measured. The ratio of PCA3 to PSA mRNA was recorded as a PCA3 score and correlated with data on each prostate specimen.
Results
Patients with extracapsular extension had a significantly higher median PCA3 score than patients without extracapsular extension (48.8 vs 18.7, p = 0.02). PCA3 score significantly correlated with total tumor volume (r = 0.38, p <0.01). On multivariate analysis PCA3 score was an independent predictor of extracapsular extension (p = 0.01) and total tumor volume less than 0.5 cc (p = 0.04). At a cutoff PCA3 score of 47 extracapsular extension was predicted with 94% specificity and an 80% positive predictive value. When combined with serum PSA and biopsy Gleason score, the ROC AUC for predicting extracapsular extension was 0.90.
Conclusions
PCA3 detected in the post-digital rectal examination urine of patients with prostate cancer correlated with pathological findings. Therefore, it could provide prognostic information. To our knowledge this is the first report of a molecular urine assay that predicts extracapsular extension.
Key Words: prostate, prostatic neoplasms, urine, tumor marker, biological, gene expression
Abbreviations and Acronyms: CaP, prostate cancer, ECE, extracapsular extension, DRE, digital rectal examination, GS, Gleason score, LR, logistic regression, PCA3, CaP gene 3, PSA, prostate specific antigen
Widespread use of PSA has led to earlier CaP detection but it has fallen short for identifying the extent of disease in an individual. Several nomograms have been developed to stratify the patient risk of nonorgan confined disease at prostatectomy using PSA, biopsy results and DRE findings1, 2, 3 but they lack accuracy. Up to 50% of cases are upgraded at prostatectomy.4, 5 Therefore, recommending active surveillance in patients remains problematic, although many may not require treatment in their lifetime.2, 6 For these reasons biomarkers that can predict pathological stage, grade or tumor volume are desperately needed.
Investigations of novel serum biomarkers for CaP have yielded only slight improvements over PSA.7, 8, 9 A recent focus is moving toward the molecular analysis of exfoliated prostatic epithelial cells from the gland by manipulation. Prostatic secretions can be difficult to obtain but adequate cellular material has been obtained in urine collected after an attentive DRE.7, 10 Of the many urine based molecular assays that are currently being evaluated the measurement of PCA3 has shown the most promise in several studies.11, 12
PCA3 is a prostate specific mRNA that is over expressed in most CaP and has high tumor cell specificity.13, 14, 15 PCA3 does not code for a protein but it can be amplified and quantified from whole urine after DRE.16 While urine based PCA3 assays show promise for predicting biopsy results,11, 12 limited data are available to address the possible prognostic value after a diagnosis is made.17 We correlated the preoperative urine PCA3 score with pathological prostatectomy features of prognostic significance, specifically stage, grade and tumor volume.
Patients and Methods
Institutional review board approval was obtained before the commencement of this study and all patients provided written informed consent. Between September 2006 and November 2007, 72 men volunteered for the study. Patients on medications that affect PSA, eg 5α-reductase inhibitors or herbal medications, were not eligible for study. Urine specimens were obtained before biopsy in 33 men or at least 6 weeks after biopsy in 39 but before prostatectomy. Table 1 lists patient characteristics.
Table 1. Preoperative clinical data
| Median age (range) | 58 (42–73) |
| Median ng/ml serum PSA (range) | 4.7(1.0–31.6) |
| Median cc total tumor vol(range) | 0.87(0.004–14.39) |
| Median PCA3 score (range) | 25.7(4.0–269.0) |
| No. race (%): | |
| White | 54(75.0) |
| Black | 18(25.0) |
| No. clinical stage (%): | |
| cT1 | 52(71.2) |
| cT2 | 21(28.8) |
| No. GS biopsy grade (%): | |
| 3+3 | 50(69.4) |
| 3+4 | 9(12.5) |
| 4+3 | 6(8.3) |
| 8−9 | 7(9.7) |
Urine specimens were obtained immediately after DRE, which involved 3 sweeps of the prostate on each lateral lobe. PCA3 and PSA mRNA were then measured using a DTS® 400 System, as previously described.11, 12, 16 PCA3 mRNA was normalized to PSA mRNA in each sample and the ratio was multiplied by 103 to obtain the PCA3 score. This score was correlated with comprehensive pathological data on radical prostatectomy whole mount specimens.
All prostatectomy specimens were processed and analyzed by the same genitourinary pathologists (BF and IS). Specimens were handed to the pathologist in the operating room at Walter Reed Army Medical Center, where they were inked and palpable tumors were incised to assess gross ECE. Specimens were then carried to the Armed Forces Institute of Pathology and processed using the institutional standard whole mount technique. Each prostate was formalin fixed, paraffin embedded and sectioned at 2.2 mm intervals before whole sections were mounted on slides.18 Tumor histology was graded using the Gleason grading system. All tumors were measured in 3 dimensions and the product was multiplied by 0.4 to estimate volume. This estimation technique was shown to be accurate in previous studies.19 Total tumor volume was determined by adding the volume of each tumor. Microscopic tumor foci were noted but considered to be of negligible volume in this study.
Spearman rank correlation analysis was used to determine associations among commonly used clinical variables, PCA3 score and total tumor volume. The median test was performed to determine the association between PCA3 score and patient clinicopathological characteristics. Stepwise LR was used to identify independent preoperative predictors (urine PCA3 score, serum PSA, biopsy grade, clinical stage, race, etc) of total tumor volume, characterized as less vs greater than 0.5 cc or ECE. ROC analysis was used to assess prediction results of the univariate and multivariate LR models with p <0.05 considered statistically significant. SAS®, version 9.1.3 was used for all data analysis.
Results
In our patient cohort we observed upgrading in 21 of 72 cases (29.2%) and down grading in 11 (15.3%) of GS from biopsy to prostatectomy specimens. Table 2 lists overall prostatectomy characteristics. Current American Joint Committee on Cancer staging criteria were used to assign a pathological T stage to each specimen. By this commonly used staging system ECE defines pT3 disease.
Table 2. Pathological data
| No. Pts(%) | |
|---|---|
| Pathological T stage: | |
| pT2 | 51(70.9) |
| pT3a | 15(20.8) |
| pT3b | 6(8.3) |
| GS prostatectomy grade: | |
| 3+3 | 42(58.3) |
| 3+4 | 20(27.8) |
| 4+3 | 3(4.1) |
| 8−9 | 7(9.7) |
| Total tumor vol(cc): | |
| Less than 0.5 | 28(38.9) |
| 0.5–2.0 | 25(34.7) |
| Greater than 2.0 | 19(26.4) |
Table 3 shows stepwise multivariate logistic regression analysis to evaluate the ability of preoperative clinical variables and PCA3 score to predict ECE. Biopsy grade greater than GS 6, preoperative PSA and PCA3 score independently predicted ECE. When ROC analysis was performed with these 3 factors combined, the AUC was 0.90. When PCA3 score was evaluated independently, the AUC was 0.732 and the ROC was noticeably skewed toward greater specificity (see figure). Using a cutoff PCA3 score of 47 in our series the resulting sensitivity, specificity and accuracy were 57%, 94% and 83%, respectively. Positive and negative predictive values using this cutoff were 80% and 84%, respectively. Of 47 patients with biopsy GS 6 and PSA less than 10 ng/ml ECE was seen in the prostatectomy specimens of 6 (13%). This finding was correctly predicted by the PCA3 score in 4 of the 6 cases (67%) with a cutoff score of 47. Results in only 3 of 47 cases were falsely positive in this subset.
Table 3. Stepwise multivariate LR analysis to predict ECE using preoperative variables in 72 patients
| Variable | Univariate | Stepwise Multivariate | ||
|---|---|---|---|---|
| OR(95% CI) | p Value | OR(95% CI) | p Value | |
| Age | 0.985 (0.919–1.087) | 0.68 | — | — |
| Race(black vs white) | 3.500 (1.136–10.779) | 0.02 | 4.517 (0.720–28.352) | 0.10 |
| Clinical stage(cT2 vs cT1) | 2.438 (0.828–7.176) | 0.10 | 4.621 (0.855–24.980) | 0.07 |
| Biopsy grade (6 vs greater than 6) | 15.710 (4.556–54.172) | <0.01 | 10.227 (2.145–48.762) | <0.01 |
| Log PSA | 5.096 (1.577–16.471) | <0.01 | 7.863 (1.406–43.987) | 0.01 |
| Log PCA3 score | 3.440 (1.519–7.795) | <0.01 | 4.155 (1.320–13.077) | 0.01 |
ROC for predicting ECE by preoperative serum PSA, PCA3 score, and combined PSA, PCA3 score and Gleason score 6 or greater than 6 (red). LR, logistic regression. bGS, biopsy Gleason score.
Spearman analysis showed that PCA3 score significantly correlated with total tumor volume in 72 patients (r = 0.38, p <0.01). Patients with a larger tumor volume of greater than 2.0 cc had a significantly higher PCA3 score (median 47.6, range 7.5 to 269.0) than patients with a smaller tumor volume of 0.5 to 2.0 cc (median 17.5) and less than 0.5 cc (median 18.7) (p = 0.01). Clinical stage, preoperative PSA and PCA3 score were independent predictors of a total tumor volume of less than 0.5 cc (p = 0.04, table 4).
Table 4. Stepwise multivariate LR analysis to predict total tumor volume less than 0.5 cc stage using preoperative variables in 72 patients
| Variable | Univariate | Stepwise Multivariate | ||
|---|---|---|---|---|
| OR(95% CI) | p Value | OR(95% CI) | p Value | |
| Age | 1.012 (0.948–1.080) | 0.73 | — | — |
| Race(black vs white) | 1.000 (0.335–2.987) | 1.00 | — | — |
| Clinical stage(cT2 vs cT1) | 3.778 (1.115–12.797) | 0.03 | 12.253 (2.181–68.835) | <0.01 |
| Biopsy grade(6 vs greater than 6) | 6.333 (1.662–24.137) | <0.01 | — | — |
| Log PSA | 3.339 (1.323–8.425) | 0.01 | 5.547 (1.748–17.596) | <0.01 |
| Log PCA3 score | 2.064 (1.063–4.008) | 0.03 | 2.222 (1.022–4.830) | 0.04 |
Discussion
The 44.4% rate of grade migration (32 of 72 cases) in this series highlights one of the major challenges of treating CaP today. There is an urgent need for markers to better predict pathological stage and the significance of disease in CaP. It would be valuable to have a marker that accurately predicts ECE and/or low volume disease in patients with otherwise low risk disease by biopsy GS and preoperative serum PSA. In these patients a predictive marker would affect decisions about lymph node dissection with prostatectomy and be invaluable for contemplating active surveillance. The data presented show that PCA3 may be useful in this regard. It is important to confirm these results in larger scale multicenter studies.
The figure shows the most remarkable result of this study. The AUC of the combination of serum PSA, biopsy GS and PCA3 score shows that this new assay works synergistically with established prognostic tools. Individually the PCA3 score does not have an overall advantage over serum PSA for predicting ECE but it adds specificity, which greatly improves the predictive ability when combined. Patients with a biopsy GS of less than 6, low PSA and a PCA3 score of less than 47 would be excellent candidates for active surveillance in this cohort. The additional correlation with tumor volume makes PCA3 score an excellent potential marker for following patients on active surveillance. If the PCA3 score were to increase while on surveillance, patients would then require treatment.
Nakanishi et al recently reported that PCA3 score correlates with tumor volume as well as prostatectomy grade but they did not find a correlation with pathological stage.17 This inconsistency may be explained by differences among subject groups or in the pathological evaluation of specimens. We found an increased frequency of pT3 disease in our cohort than they found in their series (21 of 72 cases or 29.2% vs 17 of 96 or 17.7%), which could have been a result of increased detection from analyzing the prostate in whole mounted sections with smaller intervals (2.2 vs 4 mm). The other finding that varied was a correlation with prostatectomy grade. In this study the median PCA3 score for prostatectomy GS 6 was less than the median score for GS greater than 6 but the difference was not significant (21.3 vs 30.8, p = 0.22). A larger study could better differentiate the data. It is interesting to note that recent evidence correlated preoperative tumor volume, as estimated from biopsy results, with prostatectomy grade.4 Another aspect of this population that differs from populations in other series was the relatively high percent of black men (25%). There are interesting new findings supporting different genetic changes in this group and studies are under way to perform subset analyses to evaluate possible differences.20
Previous studies have indicated that in a pre-biopsy population a PCA3 score cutoff of 35 provides an optimal balance of sensitivity and specificity for predicting CaP vs no CaP.11, 12 In this pre-prostatectomy cohort the mean tumor volume in patients with a PCA3 score of greater than 35 was 1.1 cc (median 0.56). Of these 48 patients 23 (47.9%) had a tumor volume of less than 0.5 cc and 42 (87.5%) had a tumor volume of less than 2.0 cc. In contrast, 19 of 24 patients (79.2%) with a PCA3 score of greater than 35 had a tumor volume of greater than 0.5 cc and 13 (66.7%) had a tumor volume of greater than 2.0 cc (mean 3.0, median 2.09). For a PCA3 score of less vs greater than 35 the difference in the percent of patients with a tumor volume of less than 0.5 cc was significant (p = 0.02). Determining the optimal PCA3 score cutoff point to predict prognosis, ie stratifying patients for active surveillance, is under study. It is also worth noting that for diagnostic or prognostic applications maximal predictive accuracy would likely be achieved by using PCA3 score as a continuous (vs dichotomous) variable in combination with other clinical and/or pathological data.
The theory behind the correlation of PCA3 detected in urine and tumor volume is based on the premise that cells from larger tumors are more likely to exfoliate into the prostatic ducts. These tumor cells are dislodged during DRE and expelled in first catch urine. One could also postulate that a tumor with a worse histological grade might shed cells more readily. It is difficult to understand how invasion through the prostatic capsule can be related to increased gene expression or exfoliation of cells. There could be a common cellular mechanism or this finding could be related to the anatomical location of the tumor near the capsule, such that direct manipulation by DRE releases cells more efficiently. The correlation could also be explained by tumor volume affecting PCA3 score and ECE.
Conclusions
PCA3 score in post-DRE urine of patients with CaP was a strong independent predictor of ECE that functioned synergistically with other clinical information. We also confirmed previous results showing a correlation between PCA3 score and tumor volume. Therefore, the PCA3 urine test has the potential to provide valuable prognostic information. Further studies of the value of this assay in various populations are currently under way.
References
- Pretreatment nomogram for prostate-specific antigen recurrence after radical prostatectomy or external-beam radiation therapy for clinically localized prostate cancer. J Clin Oncol. 1999;17:168
- Nomogram incorporating PSA level to predict cancer-specific survival for men with clinically localized prostate cancer managed without curative intent. Cancer. 2008;112:69
- Updated nomogram to predict pathologic stage of prostate cancer given prostate-specific antigen level, clinical stage, and biopsy Gleason score (Partin tables) based on cases from 2000 to 2005. Urology. 2007;69:1095
- . Prostate cancer volume at biopsy predicts clinically significant upgrading. J Urol. 2008;179:896
- . Correlation between prostate needle biopsy and radical prostatectomy Gleason gradings of 111 cases with prostatic adenocarcinoma. Urol J. 2004;1:246
- Management and survival of screen-detected prostate cancer patients who might have been suitable for active surveillance. Eur Urol. 2006;50:475
- Molecular detection of prostate cancer in urine by GSTP1 hypermethylation. Clin Cancer Res. 2001;7:2727
- . Is there a role for serum human tissue kallikrein in detection of prostate cancer?. Urology. 2007;70:519
- . Biomarkers for prostate cancer detection. J Urol. 2007;178:2252
- DD3(PCA3)-based molecular urine analysis for the diagnosis of prostate cancer. Eur Urol. 2003;44:8
- PCA3: a molecular urine assay for predicting prostate biopsy outcome. J Urol. 2008;179:1587
- PCA3 molecular urine assay for prostate cancer in men undergoing repeat biopsy. Urology. 2007;69:532
- Frequent overexpression of ETS-related gene-1 (ERG1) in prostate cancer transcriptome. Oncogene. 2005;24:3847
- DD3: a new prostate-specific gene, highly overexpressed in prostate cancer. Cancer Res. 1999;59:5975
- DD3(PCA3), a very sensitive and specific marker to detect prostate tumors. Cancer Res. 2002;62:2695
- APTIMA PCA3 molecular urine test: development of a method to aid in the diagnosis of prostate cancer. Clin Chem. 2006;52:1089
- PCA3 molecular urine assay correlates with prostate cancer tumor volume: implication in selecting candidates for active surveillance. J Urol. 2008;179:1804
- Mapping of TMPRSS2-ERG fusions in the context of multi-focal prostate cancer. Mod Pathol. 2008;21:67
- . A streamlined three-dimensional volume estimation method accurately classifies prostate tumors by volume. Am J Surg Pathol. 2003;27:1291
- Admixture mapping identifies 8q24 as a prostate cancer risk locus in African-American men. Proc Natl Acad Sci USA. 2006;103:14068
Study received institutional review board approval.
Supported by a grant from Gen-Probe, Inc., San Diego, California (DGM, SS).
For another article on a related topic see page 2206.
PII: S0022-5347(08)01863-6
doi:10.1016/j.juro.2008.07.060
© 2008 American Urological Association. Published by Elsevier Inc. All rights reserved.
Refers to article:
- The Significance of Monoamine Oxidase-A Expression in High Grade Prostate Cancer , 22 September 2008