Subclassification of the extent of extraprostatic extension (EPE) in radical prostatectomy (RP) specimens may enhance prognostication for prostate cancer (PCa) patients. Yet optimal criteria for staging, validation within a large cohort, and long-term follow-up are lacking.
To compare biochemical recurrence–free survival (BRFS), PCa-specific survival (PCSS), and overall survival (OS) for focal EPE (pT3aF) and nonfocal EPE (pT3aNF) in a large tertiary-referral center with the Epstein criteria for EPE extent.
Design, setting, and participants
Between 1982 and 2012, 20 434 men underwent RP, and 15 565 men (76.2%) had available pathologic and survival data. A total of 4216 men with isolated EPE were subclassified into pT3aF (1869 men, 44%) and pT3aNF (2347 men, 56%).
Outcome measurements and statistical analysis
Predictors of BRFS, PCSS, and OS were identified with multivariate Cox proportional hazard models. Covariates included age, preoperative prostate-specific antigen, body mass index, surgery year, Gleason score, and surgical margin status.
Results and limitations
With a median follow-up of 9.0 yr (range: 1–27), 314 men died of PCa, with 1300 deaths from any cause. In a multivariate model, pT3aNF compared with pT3aF was an independent predictor of BRFS (hazard ratio [HR]: 1.39; 95% confidence interval [CI], 1.18–1.62;p < 0.001), but not of PCSS (HR: 1.38; 95% CI, 0.89–2.11;p = 0.146) or OS (HR: 1.13; 95% CI, 0.94–1.36;p = 0.197). Ten-year BRFS, PCSS, and OS for pT3aF and pT3aNF were 76% versus 59%, 98% versus 96%, and 95% versus 90%, respectively.
In a large RP cohort, subclassification of EPE extent with the Epstein criteria improves correlation with BRFS. PCSS and OS in men with isolated EPE (pT3a) are excellent and are not significantly different between pT3aF and pT3aNF. The EPE extent should be subclassified to identify a subgroup of men with a higher risk of recurrence (pT3aNF) and to consider them for additional therapy.
Subclassification of extraprostatic extension (EPE) with the Epstein criteria improves correlation with biochemical recurrence–free survival. Prostate cancer–specific survival and overall survival in men with isolated EPE (pT3a) are excellent and are not significantly different between pT3aF and pT3aNF.
Keywords: Prostate cancer, Extraprostatic extension, Recurrence, Survival.
The clinical course is variable for men whose radical prostatectomy (RP) specimens reveal extraprostatic extension (EPE) in isolation, without seminal vesicle invasion or lymph node micrometastases. Without any additional treatment, more than half of these men do not progress over 10 yr  . Yet others experience biochemical recurrence of their prostate cancer (PCa), necessitating enhanced segregation of these men into meaningful subcategories, , and .
Whereas initial data imply that EPE extent may predict biochemical recurrence–free survival (BRFS), validation within a large cohort and long-term follow-up are lacking, , , , and . Various methods of determining EPE extent exist, including the Epstein criteria, the Wheeler method, and the radial distance method, , and . A survey by the International Society of Urological Pathology (ISUP) identified that most pathologists supported the suggestion that EPE should be quantified. However, ISUP does not endorse any method, citing the absence of validation in a large cohort  . Similarly, EPE extent is not included in the American Joint Committee on Cancer (AJCC) staging system  . In the present study, we investigate long-term PCa-specific survival (PCSS), overall survival (OS), and BRFS in a large cohort of RP patients at a single center at which urologic pathologists have evaluated EPE extent using the Epstein criteria dating back to 1982.
2. Materials and methods
The Johns Hopkins institutional review board–approved Radical Prostatectomy Database was queried. Between 1982 and 2012, 20 434 patients underwent RP for clinically localized PCa. Of these men, 15 565 (76.2%) had available pathologic and survival data. The mean age and prostate-specific antigen (PSA), with standard deviation, were 58.3 ± 7.8 yr and 6.8 ± 5.8 ng/ml, respectively.
RP specimens were inked and sliced serially from base to apex for microscopic examination by experienced urologic pathologists. The entire prostate was embedded for all cases. We have used an updated grading system since 2004 using a modified Gleason scoring system and did not reclassify cases before 2004.
EPEwas defined as tumor extending out of the prostate into periprostatic soft tissue. As described by Epstein et al.  , EPE was subdivided into focal EPE (a few neoplastic glands outside the prostate in two or fewer slides) and nonfocal EPE. Epstein et al. previously discriminated EPE extent as focal or established (here referred to asnonfocal). Some cases with a bulging capsule were diagnosed as pT3a, even without fatty tissue. The Epstein criteria were applied to these cases, depending on the number of glands that protrude from an imaginary normal rounded contour. The cases were categorized asfocal EPEif a few glands were located immediately out of the prostate or asnonfocal EPEif more glands were present, either in extent or in distance away from the prostate.
Patients were divided into five groups according to pathologic T stage with the Epstein subdivision: T2, T3aF (focal EPE), T3aNF (nonfocal EPE), T3b, and lymph node positive. Perioperative parameters such as age, preoperative PSA, obesity, surgery year, biopsy and pathologic Gleason score, surgical margin status, and follow-up data for progression and mortality were collected prospectively and reviewed retrospectively.
Biochemical recurrencewas defined as occurring if any postoperative PSA was ≥0.2 ng/ml. Survival status and cause-of-death information were obtained from patient follow-up, the Social Security Administration Death Index, and the Centers for Disease Control National Death Index. PCa-specific death was designated when the underlying cause of death was PCa or the patient had castration-resistant PCa at the time of death.
2.2. Statistical analysis
Patients with pT3aF and pT3aNF were compared using the studentttest or the Wilcoxon rank-sum test for continuous variables and a χ2test for categorical variables. The log-rank test and univariate and multivariate Cox proportional hazard regression models were used to compare BRFS, PCSS, and OS in each group. Covariates included age, preoperative PSA, BMI, surgery year, Gleason score, and surgical margin. To evaluate the clinical significance of EPE extent on prognosis after RP, the concordance indexes (c-indexes) for two multivariate models were compared: one distinguishing pT3aF and pT3aNF, and the other with a combined pT3a group. All analyses were performed using Stata v.12 software (StataCorp LP, College Station, TX, USA).
With a median follow-up of 9.0 yr (range: 1–27), 1707 men progressed and 314 men died of PCa, with 1300 deaths from any cause. A total of 4216 men were diagnosed with pT3a disease. Of these men, 1869 (44%) were designated pT3aF, and 2347 (56%) had pT3aNF disease ( Table 1 ). For pT3aF and pT3aNF, respectively, 10-yr BRFS was 75.6% and 58.8%, 10-yr PCSS was 98.4% and 96.0%, and 10-yr OS was 94.7% and 90.0% (Fig 1, Fig 2, and Fig 3).
|Characteristics||pT3aF||pT3aNF||p value||pT3a||Total patients|
|Age, mean, yr||58.5||59.0||< 0.01||58.8||58.0|
|Preoperative PSA, mean, ng/ml||7.1||8.9||<0.01||8.1||6.9|
|Biopsy Gleason score, %|
|Pathology Gleason score, %||<0.01|
|Positive surgical margin, %||22.6||36.6||<0.001||30.4||13.9|
BMI = body mass index; PSA = prostate-specific antigen.
Although more than half of men with pT3a disease had nonfocal EPE (pT3aNF), only 26 patients in the pT3aF group (1.4%) and 127 patients in the pT3aNF group (5.4%) underwent secondary radiation therapy (RT) within a year, because adjuvant RT was not routinely performed at our institution. In addition, only 91 men in the pT3aF group (4.9%) and 218 men in the pT3aNF group (9.3%) received hormonal therapy during the follow-up.
Kaplan-Meier survival analysis of a staging system with subclassification of pT3a by EPE extent showed that BRFS is significantly worse for advancing pathologic stage ( Fig. 1 ). In a multivariate Cox proportional hazards model, pT3aF (hazard ratio [HR]: 2.30;p < 0.001) and pT3aNF (HR: 3.20;p < 0.001) were independent predictors of BRFS compared with pT2 disease ( Table 2 ). With the same covariates, pT3aNF was directly compared with pT3aF. Compared with pT3aF, pT3aNF was found to be an independent predictor of BRFS (HR: 1.39; 95% confidence interval [CI], 1.18–1.62;p < 0.001) but not of PCSS (HR: 1.39; 95% CI, 0.89–2.11;p = 0.146) or OS (HR: 1.13; 95% CI, 0.94–1.36;p = 0.197) ( Table 2 ). The difference between the c-indexes for models separating or combining pT3aF and pT3aNF was <0.002 for all outcome measures (BRFS, PCSS, and OS) (data not shown). Subanalysis restricted to the PSA era (1989 to present) resulted in the same conclusions (data not shown).
|HR (95% CI)||p value||HR (95% CI)||p value||HR (95% CI)||p value|
|New T substage|
|T3aF||2.30 (1.94–2.74)||<0.001||2.48 (1.43-4.32)||0.001||1.05 (0.88-1.25)||0.593|
|T3aNF||3.20 (2.75–3.73)||<0.001||3.42 (2.12–5.51)||<0.001||1.18 (1.01–1.39)||0.035|
|T3b||5.00 (4.14–6.03)||<0.001||7.97 (4.81–13.20)||<0.001||1.62 (1.29–2.04)||<0.001|
|LN+||7.44 (6.16–8.99)||<0.001||10.13 (6.13–16.74)||<0.001||2.03 (1.62–2.56)||<0.001|
|T3aNF (versus T3aF)||1.39 (1.18–1.62)||<0.001||1.38 (0.89–2.11)||0.146||1.13 (0.94–1.36)||0.197|
|BMI||1.03 (1.02–1.05)||<0.001||1.02 (0.98–1.05)||0.377||1.03 (1.02–1.05)||<0.001|
|Age||1.00 (0.99–1.00)||0.529||0.99 (0.97–1.01)||0.378||1.07 (1.06–1.08)||<0.001|
|PSA||1.01 (1.01–1.02)||<0.001||1.00 (0.99–1.01)||0.594||1.00 (0.99–1.01)||0.510|
|Positive surgical margin (versus no surgical margin)|
|Yes||1.69 (1.52–1.89)||<0.001||1.41 (1.11–1.79)||0.004||1.05 (0.91–1.21)||0.547|
|Gleason score (versus 2–6)|
|7 (3 + 4)||2.73 (2.35–3.16)||<0.001||6.92 (3.98–12.05)||<0.001||1.38 (1.19–1.59)||<0.001|
|7 (4 + 3)||5.12 (4.33–6.06)||<0.001||12.93 (7.26–23.02)||<0.001||1.79 (1.49–2.16)||<0.001|
|8–10||7.53 (6.37–8.90)||<0.001||29.69 (16.41–50.17)||<0.001||2.72 (2.27–3.27)||<0.001|
|Surgery year||1.00 (0.99–1.01)||0.758||0.94 (0.92–0.97)||<0.001||0.96 (0.95–0.97)||<0.001|
BMI = body mass index; BRFS = biochemical recurrence–free survival; CI = confidence interval; HR = hazard ratio; OS = overall survival; PCSS = prostate cancer–specific survival, PSA = prostate-specific antigen.
At present, EPE extent is not included in the AJCC staging system for PCa, but this omission is being called into question as accumulating data suggest that the extent of EPE on the RP specimen may be a means of identifying higher-risk men, , and . Still, the 2010 ISUP consensus conference could not identify optimal criteria for measuring EPE extent, highlighting the need for validation within a large cohort and .
Various methods of determining EPE extent in pathologic specimens exist, notably including the Epstein criteria, which is based on number of glands and slides; the Wheeler method, which is similarly based on number of involved sections and high-power fields; and the radial distance method, which uses an ocular micrometer, , and . Sung et al.  evaluated eight methods of subclassification for their ability to predict BRFS, but only 83 specimens were available for analysis. The largest study to date used the Wheeler method for 688 patients, yet only 271 had EPE, and 9% who were lymph node positive or had seminal vesical invasion (SVI) were not excluded from the analysis  . Since being lymph node positive and having SVI are strong predictors of BRFS, , and , the true potential value of the EPE extent relies on its utility for men with isolated EPE. Because of the protracted clinical course of PCa, long-term follow-up is required to evaluate accurate oncologic outcomes.
The present study of a large tertiary-referral center with the Epstein criteria for EPE extent since 1982 is the first to report the relationship of EPE extent with PCSS and OS. A multivariate model with covariates of known predictors for biochemical recurrence showed that pT3aF versus pT3aNF (Epstein criteria) was not an independent predictor of PCSS (HR: 1.38;p = 0.146) or OS (HR: 1.13;p = 0.197). While EPE extent was a significant predictor of BRFS (HR: 1.39;p < 0.001), the degree of improvement for BRFS prediction was not clinically significant, based on a comparison of c-indexes for models combining versus distinguishing pT3aNF and pT3aF (within 0.002 for all outcome measures).
One possible explanation for the inconsistent findings for BRFS versus prostate cancer-specific survival (PCSS) or OS is the relatively few deaths compared with recurrence events, despite the large size and long follow-up of the present cohort. With an even longer follow-up and more death events, the EPE extent may be identified as a significant, independent predictor of PCSS or OS in the future analysis. Additionally, EPE extent was available to clinicians at our institution on pathology reports. If this information influenced the decision to give adjuvant therapy, this might explain why BRFS but not PCSS differed by EPE extent. However, this idea is unlikely, since only a small proportion of patients with pT3a disease underwent secondary RT within a year at our institution.
Our study showed that there was no statistically significant difference in PCSS between pT3aF and pT3aNF. However, men with any EPE experienced worse BRFS and PCSS compared with men with pT2 disease. More important, men with pT3aNF have significantly worse BRFS, as well as PCSS, compared with men with pT2 disease. Subclassification of EPE will help in identifying a subgroup of men with a higher risk of recurrence (pT3aNF).
Several pathologic considerations are worth noting. Challenges identifying the true edge of the prostate capsule may have affected pathologic specimen review, as previously reported  . Difficulty in diagnosing EPE arises when the tumor has induced a dense desmoplastic response in the periprostatic adipose tissue  . Because of the desmoplastic response, it can be difficult to judge whether the tumor has extended out of the gland or is within the fibrous tissue of the prostate  . Although the presence of fat invasion is diagnostic of EPE, one can have EPE in the absence of fat, except anteriorly. In this situation, we observed whether the normal rounded contour of the gland had been altered by a protuberance—corresponding to extension of tumor into the periprostatic tissue. One can also look for the outer limits of the condensed smooth muscle of the prostate as the boundary of the prostate in the posterior and posterolateral regions.
Further clarification is also warranted for the case of EPE at the apex and bladder neck. If tumor is present at the apical margin without benign glands, we designate it as T2x with positive surgical margin because we cannot determine whether it is an area of intraprostatic incision or EPE because of ambiguities of the histologic boundary of the prostate in this region. Regarding the bladder neck, microscopic involvement has historically been called pT3a, not pT4.
Other potential limitations include the retrospective design of the present study. Although multivariate analysis was performed, this statistical method may not capture all confounders that affect survival after RP. Specifically, information about adjuvant or salvage therapy was not included in the analysis, and these therapies are known to affect survival and . Finally, the present report concerns men treated at a single center, limiting generalizability.
In a large RP cohort, subclassification of EPE extent with the Epstein criteria improves correlation with BRFS. PCSS and OS in men with isolated EPE (pT3a) are excellent and not significantly different between pT3aF and pT3aNF. The EPE extent should be subclassified to identify a subgroup of men with a higher risk of recurrence (pT3aNF) and to consider them for additional therapy.
Author contributions:Misop Han had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design:Jeong, Han.
Acquisition of data:Humphreys, Han.
Analysis and interpretation of data:Jeong, Han.
Drafting of the manuscript:Jeong, Chalfin, Han.
Critical revision of the manuscript for important intellectual content:Epstein, Partin, Walsh.
Statistical analysis:Feng, Trock.
Administrative, technical, or material support:Chalfin.
Financial disclosures:Misop Han certifies that all conflicts of interest, including specific financial interests and relationships and affiliations relevant to the subject matter or materials discussed in the manuscript (eg, employment/affiliation, grants or funding, consultancies, honoraria, stock ownership or options, expert testimony, royalties, or patents filed, received, or pending), are the following: None.
Funding/Support and role of the sponsor:None.
-  J.I. Epstein, A.W. Partin, J. Sauvageot, P.C. Walsh. Prediction of progression following radical prostatectomy: a multivariate analysis of 721 men with long-term follow-up. Am J Surg Pathol. 1996;20:286-292 Crossref
-  J.E. McNeal, A.A. Villers, E.A. Redwine, F.S. Freiha, T.A. Stamey. Capsular penetration in prostate cancer: significance for natural history and treatment. Am J Surg Pathol. 1990;14:240-247 Crossref
-  M. Ohori, T.M. Wheeler, M.W. Kattan, Y. Goto, P.T. Scardino. Prognostic significance of positive surgical margins in radical prostatectomy specimens. J Urol. 1995;154:1818-1824
-  J.I. Epstein, G. Pizov, P.C. Walsh. Correlation of pathologic findings with progression after radical retropubic prostatectomy. Cancer. 1993;71:3582-3593 Crossref
-  T.M. Wheeler, O. Dillioglugil, M.W. Kattan, et al. Clinical and pathological significance of the level and extent of capsular invasion in clinical stage T1–2 prostate cancer. Hum Pathol. 1998;29:856-862 Crossref
-  M.T. Sung, H. Lin, M.O. Koch, D.D. Davidson, L. Cheng. Radial distance of extraprostatic extension measured by ocular micrometer is an independent predictor of prostate-specific antigen recurrence: a new proposal for the substaging of pT3a prostate cancer. Am J Surg Pathol. 2007;31:311-318 Crossref
-  A. Billis, L.L. Meirelles, L.L. Freitas, L.A. Magna, L.O. Reis, U. Ferreira. Influence of focal and diffuse extraprostatic extension and positive surgical margins on biochemical progression following radical prostatectomy. Int Braz J Urol. 2012;38:175-184 Crossref
-  B.A. van Veggel, I.M. van Oort, J.A. Witjes, L.A. Kiemeney, C.A. Hulsbergen-van de Kaa. Quantification of extraprostatic extension in prostate cancer: different parameters correlated to biochemical recurrence after radical prostatectomy. Histopathology. 2011;59:692-702 Crossref
-  J.I. Epstein, M.J. Carmichael, G. Pizov, P.C. Walsh. Influence of capsular penetration on progression following radical prostatectomy: a study of 196 cases with long-term followup. J Urol. 1993;150:135-141
-  C. Magi-Galluzzi, A.J. Evans, B. Delahunt, et al. International Society of Urological Pathology (ISUP) Consensus Conference on Handling and Staging of Radical Prostatectomy Specimens: working group 3: extraprostatic extension, lymphovascular invasion and locally advanced disease. Mod Pathol. 2011;24:26-38 Crossref
-  S.B. Edge, C.C. Compton, F.L. Greene, A. TRotti. AJCC cancer staging manual. ed. 7. (Springer, New York, NY, 2009)
-  I.D. Fleming, D.E. Henson, R.V. Hutter, et al. AJCC cancer staging manual. (Lippincott-Raven, Philadelphia, PA, 1997)
-  M.V. Tefilli, E.L. Gheiler, R. Tiguert, et al. Prognostic indicators in patients with seminal vesicle involvement following radical prostatectomy for clinically localized prostate cancer. J Urol. 1998;160:802-806
-  G.S. Palapattu, M.E. Allaf, B.J. Trock, J.I. Epstein, P.C. Walsh. Prostate specific antigen progression in men with lymph node metastases following radical prostatectomy: results of long-term followup. J Urol. 2004;172:1860-1864 Crossref
-  S.J. Freedland, W.J. Aronson, J.C. Presti Jr., et al. Predictors of prostate-specific antigen progression among men with seminal vesicle invasion at the time of radical prostatectomy. Cancer. 2004;100:1633-1638 Crossref
-  A.G. Ayala, J.Y. Ro, R. Babaian, et al. The prostatic capsule: does it exist? Its importance in the staging and treatment of prostatic carcinoma. Am J Surg Pathol. 1989;13:21-27 Crossref
-  A.J. Evans, P.C. Henry, T.H. Van der Kwast, et al. Interobserver variability between expert urologic pathologists for extraprostatic extension and surgical margin status in radical prostatectomy specimens. Am J Surg Pathol. 2008;32:1503-1512 Crossref
-  M. Bolla, H. van Poppel, B. Tombal, et al. Postoperative radiotherapy after radical prostatectomy for high-risk prostate cancer: long-term results of a randomised controlled trial (EORTC trial 22911). Lancet. 2012;380:2018-2027 Crossref
-  I.M. Thompson Jr., C.M. Tangen, J. Paradelo, et al. Adjuvant radiotherapy for pathologically advanced prostate cancer: a randomized clinical trial. JAMA. 2006;296:2329-2335 Crossref
a Department of Urology, James Buchanan Brady Urological Institute, Johns Hopkins Medical Institutions, Baltimore, MD, USA
b Department of Urology, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, Korea
c Department of Urology, Seoul National University Boramae Hospital, Seoul, Korea
d Department of Pathology, Johns Hopkins Medical Institutions, Baltimore, MD, USA
Corresponding author. James Buchanan Brady Urological Institute, Johns Hopkins Medical Institutions, 600 N. Wolfe St./Marburg 146, Baltimore, MD 21287-2101, USA. Tel. +1 410 502 7454; Fax: +1 410 955 0833.
© 2014 European Association of Urology, Published by Elsevier B.V.