Natural history of surgically treated high-risk prostate cancer

Urologic Oncology: Seminars and Original Investigations ] (2015) ∎∎∎–∎∎∎

Original article

Natural history of surgically treated high-risk prostate cancer Alberto Briganti, M.D.a,*,1, Robert Jeffrey Karnes, M.D.b,1, Giorgio Gandaglia, M.D.a, Martin Spahn, M.D.c, Paolo Gontero, M.D.d, Lorenzo Tosco, M.D.e, Burkhard Kneitz, M.D.f, Felix K.H. Chun, M.D.g, Emanuele Zaffuto, M.D.a, Maxine Sun, M.D.h, Markus Graefen, M.D.i, Giansilvio Marchioro, M.D.j, Detlef Frohneberg, M.D.k, Simone Giona, M.D.d, Pierre I. Karakiewicz, M.D.h, Hein Van Poppel, M.D.e, Francesco Montorsi, M.D.a, Steven Joniau, M.D.e, on behalf of the European Multicenter Prostate Cancer Clinical and Translational Research Group (EMPaCT) a

Unit of Urology/Division of Oncology, URI, IRCCS Ospedale San Raffaele, Milan, Italy b Department of Urology, Mayo Medical School and Mayo Clinic, Rochester, MN c Department of Urology, University of Bern, Bern, Switzerland d Department of Urology, University of Turin, Torino, Italy e Department of Urology, University Hospitals Leuven, Leuven, Belgium f Department of Urology and Pediatric Urology, University Hospital Wurzburg, Wurzburg, Germany g Department of Urology, University Hospital Hamburg-Eppendorf, Hamburg, Germany h Cancer Prognostics and Health Outcomes Unit, University of Montreal Health Center, Montreal, Quebec, Canada i Martiniclinic, Prostate Cancer Center Hamburg-Eppendorf, Hamburg, Germany j Department of Urology, University of Piemonte Orientale, Novara, Italy k Department of Urology, Community Hospital Karlsruhe, Karlsruhe, Germany Received 17 September 2014; received in revised form 4 November 2014; accepted 26 November 2014

Abstract Background: No data exist on the patterns of biochemical recurrence (BCR) and their effect on survival in patients with high-risk prostate cancer (PCa) treated with surgery. The aim of our investigation was to evaluate the natural history of PCa in patients treated with radical prostatectomy (RP) alone. Materials and methods: Overall, 2,065 patients with high-risk PCa treated with RP at 7 tertiary referral centers between 1991 and 2011 were identified. First, we calculated the probability of experiencing BCR after surgery. Particularly, we relied on conditional survival estimates for BCR after RP. Competing-risks regression analyses were then used to evaluate the effect of time to BCR on the risk of cancer-specific mortality (CSM). Results: Median follow-up was 70 months. Overall, the 5-year BCR-free survival rate was 55.2%. Given the BCR-free survivorship at 1, 2, 3, 4, and 5 years, the BCR-free survival rates improved by þ7.6%, þ4.1%, þ4.8%, þ3.2%, and þ3.7%, respectively. Overall, the 10-year CSM rate was 14.8%. When patients were stratified according to time to BCR, patients experiencing BCR within 36 months from surgery had higher 10-year CSM rates compared with those experiencing late BCR (19.1% vs. 4.4%; P o 0.001). At multivariate analyses, time to BCR represented an independent predictor of CSM (P o 0.001). Conclusions: Increasing time from surgery is associated with a reduction of the risk of subsequent BCR. Additionally, time to BCR represents a predictor of CSM in these patients. These results might help provide clinicians with better follow-up strategies and more aggressive treatments for early BCR. r 2015 Elsevier Inc. All rights reserved. Keywords: Prostate cancer; Radical prostatectomy; Biochemical recurrence; Cancer-specific mortality; Time to biochemical recurrence

1. Introduction 1

Both authors contributed equally to the manuscript. Corresponding author. Tel.: þ39 02 26437286; fax: þ39 02 26437298. E-mail addresses: [email protected] (A. Briganti) *

http://dx.doi.org/10.1016/j.urolonc.2014.11.018 1078-1439/r 2015 Elsevier Inc. All rights reserved.

Radical prostatectomy (RP) is considered a first-line treatment option for patients with prostate cancer (PCa) [1]. This surgical approach is associated with excellent cancer control

2

A. Briganti et al. / Urologic Oncology: Seminars and Original Investigations ] (2015) 1–7

rates [2–4]. This is true even in the high-risk setting, where patients treated with RP might be more likely to succumb to competing causes of death rather than owing to PCa [5]. In this group, patients at higher risk of death owing to PCa after surgery are those having more aggressive pathological features, longer life expectancy, and better comorbidity profile. However, no data currently exist regarding the patterns of biochemical recurrence (BCR) and their effect on survival in high-risk patients treated with RP alone. Indeed, most studies included mainly patients with favorable disease characteristics [6–10], whereas none of them focused on high-risk PCa. This is crucial, as high-risk patients are those who are more likely to benefit from active treatment [11]. Similarly, the effect of competing causes of mortality on the probability of experiencing cancer-specific mortality (CSM) in these men is not well documented. Exploring these data would be important for 3 main reasons—first, to identify patients at high-risk of dying owing to PCa, who might therefore benefit from multimodal therapies; second, to spare the use of aggressive treatments in patients more likely to die of other causes; and third, to appropriately counsel patients concerning their long-term outcomes. To address these issues, we evaluated the patterns of BCR and their prognostic role on survival in a large contemporary cohort of men with high-risk PCa treated with RP alone. Specifically, we aimed at (1) evaluating the effect of the time elapsed between surgery and patient evaluation on the subsequent risk of BCR and (2) addressing the role of time to BCR on CSM. We relied on competing-risks analyses to account for the risk of dying owing to other causes rather than PCa. Indeed, given the protracted natural history of the disease, the proportion of high-risk patients dying owing to other-cause mortality (OCM) is not negligible [5,12]. 2. Materials and methods 2.1. Study population Overall, 5,334 patients who received RP and pelvic lymph node dissection for nonmetastatic PCa between 1991 and 2011 at 7 worldwide tertiary care centers were considered. All patients had high-risk PCa according to D’Amico classification (clinical stage ZT2c, biopsy Gleason score 8–10, or preoperative prostate-specific antigen [PSA] Z 20 ng/ml) [13]. For the purpose of our analyses, we excluded 1,957 patients who underwent adjuvant radiotherapy or androgen deprivation therapy (ADT). Additional exclusion criteria consisted of unknown comorbidity profile (n ¼ 490), unavailable followup (n ¼ 382), unknown biopsy Gleason score (n ¼ 131), unknown clinical stage (n ¼ 68), unknown pathological characteristics (n ¼ 29), and neoadjuvant ADT (n ¼ 230). This resulted in a final population of 2,065 patients.

stage, biopsy Gleason score, pathological stage and Gleason score, surgical margin status, lymph node invasion (LNI), and time to BCR. The comorbidity profile was assessed with the Charlson Comorbidity Index (CCI) [14]. BCR was defined as 2 consecutive PSA values Z0.2 ng/ml after RP. Deaths owing to PCa were coded as CSM. All other events were considered as OCM. Patients underwent follow-up visits every 3 months during the first year after surgery, and every 6 months thereafter. Vital status and cause of death were identified from death certificates or physician correspondence. 2.3. Statistical analysis Statistical analyses were severalfold. First, we calculated the conditional survival estimates for BCR using the multiplicative law of probability, which states that knowledge of the probability of an event A and event B occurring, and the probability of event A occurring, allows the calculation of the conditional probability of event B occurring, given that event A has occurred: (probability A/B) ¼ (probability of A and B)/ (probability of A). These analyses were repeated after stratifying patients according to pathological Gleason score, pathological stage, surgical margins, and LNI. Second, multivariate Cox regression models were fitted for prediction of BCR. To examine the variation for risk of BCR with time, separate Cox regression models were fitted in patients who did not experience BCR within Z12, 24, 36, 48, and 60 months after RP. Adjustment was made for age, year, pathological stage, nodal status, Gleason score, and surgical margins. Third, we evaluated the effect of time to BCR on CSM in a cohort of patients who experienced BCR (n ¼ 823). Time from surgery to BCR was dichotomized according to the most informative cutoff predicting CSM. This was obtained applying the chisquare test for every possible cutoff value and choosing the lowest P value. Cumulative incidence CSM rates were then generated according to the dichotomized time to BCR and compared with the Gray test [15]. Additionally, to generate a graphical illustration of the 10-year CSM and OCM rates given time to BCR and CCI, we relied on Poisson regression models. Finally, multivariable competing-risks regression models were used to test the effect of time from surgery to BCR on the risk of dying owing to PCa. The competing-risks regression methodology allowed us to account for the effects of OCM, providing the most unbiased estimates of CSM [15–17]. All statistical tests were performed using the R statistical package (version 2.13.1), with a 2-sided significance level set at P o 0.05. 3. Results 3.1. Baseline characteristics

2.2. Prognostic factors and outcome variables All patients had complete clinical and pathological data, including age, year of surgery, preoperative PSA, clinical

Baseline descriptive characteristics are summarized in the Table. When patients were stratified according to BCR after RP, significant differences were recorded with respect to year

A. Briganti et al. / Urologic Oncology: Seminars and Original Investigations ] (2015) 1–7

3

Table Descriptive characteristics of 2,065 patients treated with radical prostatectomy (RP) for high-risk prostate cancer (PCa), stratified according to biochemical recurrence status (BCR defined as 2 consecutive PSA values Z0.2 ng/ml) Overall, n ¼ 2,065

No BCR, n ¼ 1,242 (60.1%)

BCR, n ¼ 823 (39.9%)

Age at surgery, y Mean (median) IQR

64.5 (65) 60–69

64.6 (66) 60–69

64.4 (65) 60–69

0.6

Year of diagnosis (%) 1991–1998 1999–2003 2004–2008 2009–2011

586 480 514 485

243 270 337 392

(19.6) (21.7) (27.1) (31.6)

343 (41.7) 210 (25.5) 177 (21.5) 93 (11.3)

o0.001

Charlson Comorbidity Index 0 1 Z2

1,376 (66.6) 427 (20.7) 262 (12.7)

813 (65.5) 262 (21.1) 167 (13.4)

563 (68.4) 165 (20.0) 95 (11.5)

0.3

Preoperative PSA, ng/ml Mean (median) IQR

16.3 (8.8) 4.4–22.7

14.6 (8.5) 4.6–21.2

18.9 (9.3) 4.0–24.4

0.001

Clinical stage rT2a T2b ZT2c

1,001 (48.5) 211 (10.2) 853 (41.3)

685 (55.2) 90 (7.2) 467 (37.6)

316 (38.4) 121 (14.7) 386 (46.9)

o0.001

Biopsy Gleason score (%) r6 7 Z8

575 (27.8) 534 (25.9) 956 (46.3)

388 (31.2) 261 (21.0) 593 (47.7)

187 (22.7) 273 (33.2) 363 (44.1)

o0.001

Pathological T stage (%) pT2 pT3a pT3b/4

917 (44.4) 679 (32.9) 469 (22.7)

671 (54.0) 393 (31.6) 178 (14.3)

246 (29.9) 286 (34.8) 291 (35.4)

o0.001

N stage (%) pN1

207 (10.0)

98 (7.9)

109 (13.2)

o0.001

Pathological Gleason score (%) r6 7 Z8

377 (18.3) 1,288 (62.4) 400 (19.4)

269 (21.7) 794 (63.9) 179 (14.4)

108 (13.1) 494 (60.0) 221 (26.9)

o0.001

Surgical margins (%) Negative Positive

936 (45.3) 1,129 (54.7)

528 (42.5) 714 (57.5)

408 (49.6) 415 (50.4)

0.002

(28.4) (23.2) (24.9) (23.5)

P value

IQR ¼ interquartile range.

of diagnosis, preoperative PSA, clinical stage, biopsy Gleason score, pathological stage, pathological Gleason score, and surgical margins (all P r 0.002). Median follow-up after surgery was 70 months (mean ¼ 81). Overall, 823 (39.9%) patients experienced BCR after surgery.

respectively (Fig. 1). Similar trends were observed when patients were stratified according to pathological Gleason score, pathological stage, surgical margins, and LNI.

3.2. Conditional BCR-free survival rates

In multivariable Cox regression analyses, pathological Gleason score, pathological stage, and LNI were significant predictors of BCR after surgery (all P r 0.01). The risk of BCR decreased with increasing length of BCR-free survival (Z1, Z2, Z3, Z4, and Z5 y) among patients with Gleason score 8 to 10. Particularly, the risk of BCR for patients with Gleason score 8 to 10 was significantly higher than for

Overall, the 5-year BCR-free survival rate after surgery was 55.2% (95% CI: 55.0%–55.3%). Given the BCR-free survivorship at 1, 2, 3, 4, and 5 years, the 5-year BCR-free survival probabilities were improved by þ7.6 (62.8%), þ4.1 (66.9%), þ4.8 (71.7%), þ3.2 (74.9%), and þ3.7 (78.6%),

3.3. Multivariable regression analyses predicting BCR

4

A. Briganti et al. / Urologic Oncology: Seminars and Original Investigations ] (2015) 1–7

4. Discussion

Fig. 1. The 5-year biochemical recurrence (BCR)–free survival probability, provided the number of years survived after radical prostatectomy in the entire population (n ¼ 2,065).

men with lower grade only within 36 months after surgery (P r 0.03). Conversely, the risk of BCR remained stable among patients with pathological stage pT3a and pT3b/4 (all P r 0.002). Finally, LNI was not associated with increased risk of BCR in patients who survived more than 1 year without experiencing recurrence (all P Z 0.2). 3.4. CSM rates after BCR When focusing exclusively on patients who experienced BCR after surgery (n ¼ 823), the mean (median) time from surgery to BCR was 29.2 (17) months. The overall 5- and 10-year CSM rates were 6.3% (95% CI: 4.6%–8.9%) and 14.8% (95% CI: 16.3%–24.4%), respectively. The most informative cutoff for time from RP to BCR in predicting CSM was 36 months (P o 0.001). Patients were divided in 2 groups according to this cutoff: men experiencing BCR within 36 months from surgery (n ¼ 569, 69.1%) and those who experienced BCR after 36 months from surgery (n ¼ 254, 30.9%). Overall, CSM rates were significantly higher in patients who experienced BCR within 36 months from surgery compared with their counterpart who experienced late BCR (P o 0.001; Fig. 2A). Overall, OCM rates increased with CCI and were significantly higher in patients experiencing late BCR (P o 0.001; Fig. 2B). 3.5. Competing-risks regression analyses In multivariable analyses, time from surgery to BCR represented a significant predictor of CSM (hazard ratio ¼ 0.97, 95% CI: 0.96–0.98; P o 0.001; Supplementary Table). Additionally, pathological Gleason score and pathological stage were associated with higher risk of dying owing to PCa (all P r 0.04). Particularly, patients with Gleason score 8 to 10 had 2.9-fold higher risk of dying owing to PCa as compared with their counterpart with lower grade (P o 0.001). Similarly, men with pT3b/4 disease had 1.7-fold higher risk of CSM compared with individuals with pT2 disease (P ¼ 0.002).

Although several studies have investigated predictors of poor outcomes after RP [3,4,18–26], none of them assessed the patterns of BCR and their effect on CSM in patients with high-risk PCa treated with surgery alone. To address this issue, our study consisted of several steps. First, we evaluated the effect of the time elapsed between RP and patient evaluation on the risk of BCR. Conditional survival analyses allowed us to estimate the 5-year BCR probability, given that a patient had already a BCR-free survivorship of x years after surgery. Second, we evaluated the role of time to BCR on CSM. To account for the risk of dying owing to other causes rather than from PCa, we relied on competingrisks analyses. Indeed, given the protracted natural history of the disease, a not-negligible proportion of patients may not be at considerable risk of CSM, even after recurrence [8,10,19,21,22,24,27]. The results of our study are severalfold. First, we showed that the probability of being free of recurrence substantially increases with increasing BCR-free survivorship time. For example, the probability of 5-year BCR-free survival improved from 62.8% to 78.6% given a 1- and 5-year BCR-free survivorship after surgery, respectively. This was confirmed even after stratifying patients according to pathological Gleason score, tumor stage, and nodal status. When evaluating the predictors of recurrence, patients with higher Gleason score, non–organ-confined disease, and LNI had increased risk of BCR at baseline. However, the presence of a pathologic Gleason score 8 to 10 increased the risk of BCR only within the first 3 years after surgery. After this time point, the risk of subsequent BCR was not significantly higher than for those with lower grade disease. Similarly, LNI conferred an increased risk of BCR only within the first year after RP. These findings confirm that patients with more aggressive disease features at final pathology should be considered at higher risk of recurrence immediately after surgery. However, this risk substantially decreases over time. Our results corroborate previous findings [21,25,26,28]. For example, Loeb et al. [25] recently reported that approximately 80% of all the BCRs after RP occurred within the first 5 years from surgery. Additionally, Tollefson et al. [26] demonstrated that the risk of BCR is inversely proportional to the duration of the BCR-free survival. However, they focused exclusively on patients with low-risk disease. Conversely, our study was able to confirm these results in a large contemporary cohort of patients with high-risk PCa treated with RP without neoadjuvant or adjuvant therapies. The interest in these results stems on the evidence that patients with high-risk disease are those who benefit the most from active treatment [11]. Taken together, these findings highlight that the time elapsed between surgery and patient evaluation should be factored when planning follow-up schedules. Particularly, the frequency of postoperative PSA testing should be individually tailored based not only on patient and disease

A. Briganti et al. / Urologic Oncology: Seminars and Original Investigations ] (2015) 1–7

5

Fig. 2. (A) Cumulative incidence plot depicting CSM rates stratified according to time from surgery to biochemical recurrence (BCR) (o36 vs. Z36 mo) in patients who experienced BCR after surgery (n ¼ 823) and (B) smoothed model-derived 10-year cumulative mortality estimates. (Color version of the figure is available online.)

6

A. Briganti et al. / Urologic Oncology: Seminars and Original Investigations ] (2015) 1–7

characteristics but also on the duration of BCR-free survivorship. Interestingly, even in the high-risk setting, the patterns of recurrence are favorable given a 3-year period of survivorship without BCR. Second, our findings show that individuals who experienced BCR within 3 years from surgery had significantly higher CSM rates compared with those who developed late BCR. This was confirmed at competing-risks regression analyses, where a longer time to BCR was associated with lower risk of CSM, after accounting for the risk of OCM. Death from PCa was a competing cause of death only in men with early BCR with or without favorable comorbidity profile. Early recurrent patients are therefore those in whom early salvage treatment may eventually exert the maximal effect on survival. Conversely, the effect of PCa on patient survival was limited in case of late recurrence. Some of these patients, even when recurring after treatment for highrisk disease, may be affected by slowly progressive relapse. In these men, OCM was indeed the leading cause of death. To the best of our knowledge, this is the first study addressing this issue in a large contemporary cohort of patients with high-risk disease treated with RP alone. Others have addressed the association between time of recurrence and CSM in different patient populations [8,19,21, 22,24,25]. Freedland et al. [22] showed that patients who experienced BCR within 3 years from surgery had an increased risk of PCa-related mortality compared with those who recurred after this time point. Conversely, Boorjian et al. [24] failed to demonstrate an association between time from surgery to BCR and systemic progression or CSM. However, they evaluated a population of patients with more favorable disease characteristics. For example, although we included roughly 20% of patients with pathological Gleason score 8 to 10, this category represented only the 7% of the individuals included in their cohort. Similarly, only 10% of their patients had pT3b disease. The inclusion of individuals with favorable characteristics might be responsible for the lack of significance in the association between time to BCR and CSM observed in the previous study. Indeed, although patients with low stage and grade have a low risk of dying owing to PCa even after tumor recurrence, those with more aggressive disease characteristics represent individuals at higher risk of dying owing to PCa after surgery [8,19,20,22]. Consequently, it might be hypothesized that the detrimental effect of early BCR appears only in these individuals. This hypothesis has been recently confirmed by Bolton et al. [19] who showed that the relationship between time to BCR and the risk of CSM critically depends on disease characteristics at final pathology. Taken together, our observations might have profound implications. Because we demonstrated that not all the patients with high-risk disease share the same prognosis even in the context of recurrent disease, patient selection is crucial to identify which patients with BCR are more likely to die owing to PCa. This would allow also sparing the side effects related to salvage treatments to those individuals at

high risk of OCM. Although patients with more adverse disease characteristics should undergo close follow-up to promptly detect the first signs of recurrence and administer salvage treatments [29], this might not be true for men with less aggressive recurrent disease. Patients who have late recurrence might be indeed counseled regarding their relatively favorable long-term prognosis. Only 4% of these patients will indeed die owing to PCa at 10-year follow-up. On the contrary, approximately 1 of 3 of these individuals will experience OCM. Consequently, comorbidity status should be factored when balancing the benefits and side effects of the administration of local or systemic salvage treatments. This is particularly important when considering the cardiovascular toxicity of systemic treatments such as ADT. Despite several strengths, our investigation is not devoid of limitations. First, it is limited by its retrospective nature. For example, we cannot exclude that individuals with more aggressive characteristics at final pathology received adjuvant treatments and consequently were excluded from the current analyses. However, we decided to exclude from our cohort patients treated with adjuvant therapies to evaluate the natural history of BCR after surgery and to avoid the confounding effect of multimodal treatments on recurrence patterns. Second, we included patients treated over a relatively long time interval. Over time, improvements might have been achieved in diagnostic and surgical techniques. Additionally, the interpretative changes applied to the Gleason grading system over the past decades might have changed the patterns of grade distribution over time [30], potentially leading to different outcomes in particular subgroups of patients. This, together with the lack of central pathological review, might in part limit the generalizability of our findings to contemporary patients with PCa. However, all pathological specimens were evaluated by experienced uropathologists at high-volume referral centers. Third, heterogeneity might exist regarding salvage therapies among several different high-volume centers over such a long study period. Particularly, the timing, dose, and schedule of administration of salvage treatments might have varied over time and according to each treating institution based on disease characteristics, as well as physician attitudes. Despite this, it should be noted that the association between time from surgery to BCR and the risk of dying owing to PCa was maintained even after adjusting our multivariable models for year of diagnosis and for the administration of salvage therapies.

5. Conclusions Our findings indicate that the time elapsed between surgery and patient evaluation without recurrence has a significant effect on the subsequent risk of BCR. Particularly, the probability of recurrence decreases according to the length of BCR-free survival intervals. Moreover, in

A. Briganti et al. / Urologic Oncology: Seminars and Original Investigations ] (2015) 1–7

patients experiencing BCR, a short time between surgery and recurrence increases the risk of CSM. These results might help provide clinicians with better follow-up strategies and more aggressive multimodal treatments for patients experiencing early BCR. Appendix A. Supporting Information

[14]

[15] [16] [17]

Supplementary material cited in this article is available online at http://dx.doi.org/10.1016/j.urolonc.2014.11.018.

References [1] Siegel R, Naishadham D, Jemal A. Cancer statistics, 2013. CA Cancer J Clin 2013;63:11–30. [2] Boorjian SA, Eastham JA, Graefen M, et al. A critical analysis of the long-term impact of radical prostatectomy on cancer control and function outcomes. Eur Urol 2012;61:664–75. [3] Eggener SE, Scardino PT, Walsh PC, et al. Predicting 15-year prostate cancer specific mortality after radical prostatectomy. J Urol 2011;185:869–75. [4] Mullins JK, Feng Z, Trock BJ, Epstein JI, Walsh PC, Loeb S. The impact of anatomical radical retropubic prostatectomy on cancer control: the 30-year anniversary. J Urol 2012;188:2219–24. [5] Briganti A, Spahn M, Joniau S, et al. Impact of age and comorbidities on long-term survival of patients with high-risk prostate cancer treated with radical prostatectomy: a multi-institutional competing-risks analysis. Eur Urol 2013;63:693–701. [6] Antonarakis ES, Chen Y, Elsamanoudi SI, et al. Long-term overall survival and metastasis-free survival for men with prostate-specific antigen-recurrent prostate cancer after prostatectomy: analysis of the Center for Prostate Disease Research National Database. BJU Int 2011;108:378–85. [7] Antonarakis ES, Feng Z, Trock BJ, et al. The natural history of metastatic progression in men with prostate-specific antigen recurrence after radical prostatectomy: long-term follow-up. BJU Int 2012;109:32–9. [8] Freedland SJ, Humphreys EB, Mangold LA, et al. Risk of prostate cancer-specific mortality following biochemical recurrence after radical prostatectomy. J Am Med Assoc 2005;294:433–9. [9] Pound CR, Partin AW, Eisenberger MA, Chan DW, Pearson JD, Walsh PC. Natural history of progression after PSA elevation following radical prostatectomy. J Am Med Assoc 1999;281:1591–7. [10] Ward JF, Blute ML, Slezak J, Bergstralh EJ, Zincke H. The long-term clinical impact of biochemical recurrence of prostate cancer 5 or more years after radical prostatectomy. J Urol 2003;170:1872–6. [11] Abdollah F, Sun M, Thuret R, et al. A competing-risks analysis of survival after alternative treatment modalities for prostate cancer patients: 1988-2006. Eur Urol 2011;59:88–95. [12] Daskivich TJ, Fan KH, Koyama T, et al. Effect of age, tumor risk, and comorbidity on competing risks for survival in a U.S. population-based cohort of men with prostate cancer. Ann Intern Med 2013;158:709–17. [13] D’Amico AV, Whittington R, Malkowicz SB, et al. Biochemical outcome after radical prostatectomy, external beam radiation therapy,

[18]

[19]

[20]

[21]

[22]

[23]

[24]

[25]

[26]

[27]

[28]

[29]

[30]

7

or interstitial radiation therapy for clinically localized prostate cancer. J Am Med Assoc 1998;280:969–74. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chron Dis 1987;40:373–83. Gray R. A class of K-sample tests for comparing the cumulative incidence of a competing risk. Ann Stat 1988;16:1140–54. Jason PF, Gray RJ. A proportional hazards model for the subdistribution of a competing risk. J Am Stat Assoc 1999;94:496–509. Muller H-G, Wang J-L. Hazard rate estimation under random censoring with varying kernels and bandwidths. Biometrics 1994;50: 61–76. Bianco FJ Jr, Scardino PT, Eastham JA. Radical prostatectomy: longterm cancer control and recovery of sexual and urinary function (“trifecta”). Urology 2005;66(Suppl. 5):83–94. Bolton DM, Ta A, Bagnato M, et al. Interval to biochemical recurrence following radical prostatectomy does not affect survival in men with low-risk prostate cancer. World J Urol 2014;32:431–5. Stephenson AJ, Kattan MW, Eastham JA, et al. Prostate cancerspecific mortality after radical prostatectomy for patients treated in the prostate-specific antigen era. J Clin Oncol 2009;27:4300–5. Caire AA, Sun L, Ode O, et al. Delayed prostate-specific antigen recurrence after radical prostatectomy: how to identify and what are their clinical outcomes? Urology 2009;74:643–7. Freedland SJ, Humphreys EB, Mangold LA, Eisenberger M, Partin AW. Time to prostate specific antigen recurrence after radical prostatectomy and risk of prostate cancer specific mortality. J Urol 2006;176(4 Pt 1):1404–8. Zhou P, Chen MH, McLeod D, Carroll PR, Moul JW, D’Amico AV. Predictors of prostate cancer-specific mortality after radical prostatectomy or radiation therapy. J Clin Oncol 2005;23:6992–8. Boorjian SA, Thompson RH, Tollefson MK, et al. Long-term risk of clinical progression after biochemical recurrence following radical prostatectomy: the impact of time from surgery to recurrence. Eur Urol 2011;59:893–9. Loeb S, Feng Z, Ross A, Trock BJ, Humphreys EB, Walsh PC. Can we stop prostate specific antigen testing 10 years after radical prostatectomy? J Urol 2011;186:500–5. Tollefson MK, Blute ML, Rangel LJ, Karnes RJ, Frank I. Lifelong yearly prostate specific antigen surveillance is not necessary for low risk prostate cancer treated with radical prostatectomy. J Urol 2010;184:925–9. Abdollah F, Boorjian S, Cozzarini C, et al. Survival following biochemical recurrence after radical prostatectomy and adjuvant radiotherapy in patients with prostate cancer: the impact of competing causes of mortality and patient stratification. Eur Urol 2013;64: 557–64. Amling CL, Blute ML, Bergstralh EJ, Seay TM, Slezak J, Zincke H. Long-term hazard of progression after radical prostatectomy for clinically localized prostate cancer: continued risk of biochemical failure after 5 years. J Urol 2000;164:101–5. Briganti A, Wiegel T, Joniau S, et al. Early salvage radiation therapy does not compromise cancer control in patients with pT3N0 prostate cancer after radical prostatectomy: results of a match-controlled multiinstitutional analysis. Eur Urol 2012;62:472–87. Delahunt B, Miller RJ, Srigley JR, Evans AJ, Samaratunga H. Gleason grading: past, present and future. Histopathology 2012;60:75–86.