Oncologic Outcomes Of Definitive Treatments For Low- & Intermediate-Risk Prostate Cancer After A Period Of Active Surveillance

Transcription

1 Accepted Manuscript Oncologic Outcomes Of Definitive Treatments For Low- & Intermediate-Risk Prostate Cancer After A Period Of Active Surveillance Michael J. Whalen, Jamie S. Pak, Danny Lascano, David Ahlborn, Justin T. Matulay, James M. McKiernan, Mitchell C. Benson, Sven Wenske PII: DOI: Reference: CLGC 950 S (17) /j.clgc To appear in: Clinical Genitourinary Cancer Received Date: 29 July 2017 Revised Date: 26 September 2017 Accepted Date: 9 October 2017 Please cite this article as: Whalen MJ, Pak JS, Lascano D, Ahlborn D, Matulay JT, McKiernan JM, Benson MC, Wenske S, Oncologic Outcomes Of Definitive Treatments For Low- & Intermediate-Risk Prostate Cancer After A Period Of Active Surveillance, Clinical Genitourinary Cancer (2017), doi: /j.clgc This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

2 ONCOLOGIC OUTCOMES OF DEFINITIVE TREATMENTS FOR LOW- & INTERMEDIATE-RISK PROSTATE CANCER AFTER A PERIOD OF ACTIVE SURVEILLANCE Michael J. Whalen 1, Jamie S. Pak 2, 3 Danny Lascano 4, David Ahlborn 3, Justin T. Matulay 2, James M. McKiernan 2, Mitchell C. Benson 2, Sven Wenske 2 1 Dept of Urology, George Washington University School of Medicine and Health Sciences, Washington, DC 2 Dept of Urology, Columbia University Medical Center, New York, NY 3 Icahn School of Medicine at Mount Sinai Hospital, New York, NY 4 Westchester Medical Center, Valhalla, NY Abstract: 239 words Word count: 3,270 with headings Running Title: Prostate Cancer Treatment after Active Surveillance Key words: prostate cancer, active surveillance, definitive treatment, outcomes, PSA velocity Corresponding author: Michael J. Whalen, MD Department of Urology George Washington University School of Medicine and Health Sciences 2150 Pennsylvania Ave NW, Suite Washington, DC Phone: (202) Fax: (202) mjw2117@gmail.com 1

3 ABSTRACT We identified n=237 patient with low- and intermediate-risk prostate cancer enrolled in active surveillance at our institution who later progressed to definitive treatment. We compared outcomes of the different treatments after a median followup of 71.7 months using multivariable regression and Kaplan Meier survival analysis. Radiation therapy was associated with worse overall mortality compared to other treatments, which may influence decisionmaking for patients who progress on active surveillance. Background: To compare oncologic outcomes of different definitive treatment (DT) modalities in a cohort of prostate cancer (PCa) patients after active surveillance (AS). Methods: We identified 237 patients with NCCN low- and intermediate-risk prostate cancer diagnosed from who did not undergo immediate definitive treatment within 12mos of diagnosis (i.e. AS patients as well as watchful waiting and those refusing DT). Charts were examined for clinical/pathologic data and type of DT: surgery (RP), radiation including brachytherapy--(xrt), cryotherapy, and androgen deprivation therapy monotherapy (ADT). The impact of DT on oncologic outcomes of biochemical recurrence (BCR), metastasis, disease-specific (DSS), and overall survival (OS) was examined with the Cox proportional hazards model, along with the Kaplan Meier method and log-rank test. Results: After median time on AS of 63.4 months, 40% of patients underwent DT: 47% XRT, 28% RP, 14% ADT, and 11% cryotherapy. On multivariable analysis, the use of XRT predicted higher BCR (HR 6.1, p=0.001) and worse overall mortality (HR 2.1, p=0.03) compared to other treatments controlling for age, Charlson Comorbidity Index (CCI), stage, Gleason score, and NCCN risk category. Median follow-up was 71.7 months. On Kaplan-Meier analysis, 10- year OS was superior for RP vs. XRT among patients with PSA velocity >2.0 ng/ml/yr. Conclusions: 2

4 Low and intermediate risk prostate cancer patients who progress to definitive treatment after active surveillance may be inadequately treated with radiation therapy compared to other definitive treatment modalities, especially when pre-treatment PSA velocity is >2ng/mL/yr. Clinical Practice Points: Active Surveillance for prostate cancer has gained widespread acceptance with short- to medium-term follow-up in published series. It is unknown whether any of the established definitive treatment modalities for prostate cancer (i.e. radical prostatectomy, radiation, cryotherapy, or androgen deprivation therapy) provide any oncologic superiority for patients who progress to treatment after a period of active surveillance. Under-sampling error from prostate biopsy and the ability of surgery to provide definitive stage and grade diagnosis versus the use of neoadjuvant androgen deprivation therapy with radiotherapy that may provide superior control of locally advanced disease and nodal micrometastasis represent unique advantages that may translate into superior outcomes, especially for men who have had a significant period of time without definitive treatment of their active malignancy. To address this unanswered question in the literature, we identified n=237 patient with low- and intermediate-risk prostate cancer enrolled in active surveillance at our institution who later progressed to definitive treatment. We compared outcomes of the different treatments after a median follow-up of 71.7 months using multivariable regression with the Cox Proportional Hazards Model and the Kaplan-Meier method. Radiation therapy was associated with worse biochemical recurrence and overall mortality compared to other treatments, after controlling for age, Charlson Comorbidity Index (CCI), stage, Gleason score, and NCCN risk category. For patients with PSA velocity >2ng/mL/yr, 10 year overall survival after radical prostatectomy was superior to radiation therapy. These findings may influence the decision-making for patients who progress on active surveillance and remain suitable candidates for radical surgery. 3

5 INTRODUCTION The majority of men diagnosed with prostate cancer (PCa) today are over 65 years old and harbor low- to intermediate-risk disease.[1] Due to growing concern of over-treating these men, active surveillance (AS) with selective delayed intervention is becoming increasingly common,[2-4] with several institutions publishing long-term outcomes.[5-10] Though survival outcomes are encouraging, a substantial proportion of patients eventually undergo treatment. These rates range 24%-41% after 5 years[5-12] and 37%-59% after 10 years on AS.[7,8,11-12] This highlights the need to investigate the efficacy of different definitive treatments (DT) in this setting, especially since the majority of studies report progression-to-treatment as an endpoint in itself without describing the treatment paradigms of those who progress. Prior studies have indicated that rates of adverse pathologic features after radical prostatectomy for men progressing to Gleason 7 disease after a period of active surveillance were similar to men with de novo Gleason 7 disease, suggesting similar efficacy of surgery in both scenarios.[13] In clinical practice, the choice of DT depends on multiple factors, including age, presence of comorbidities, and patient preference. The aim of this study was to compare the oncologic and survival outcomes of different DT for PCa in a cohort of lowand intermediate-risk patients after a period of AS. MATERIALS AND METHODS A retrospective review of the IRB-approved single institution database was conducted to identify 237 patients diagnosed with National Comprehensive Cancer Network (NCCN) low- and intermediate-risk PCa from who did not undergo DT within 12 months of diagnosis (i.e. AS patients plus watchful waiting and those refusing DT). Low-risk was defined as ct1c/t2a, PSA 10ng/mL, or Gleason score (GS) 6 this category also included very-low risk patients by definition; intermediate-risk was defined as ct2b, PSA 10-20ng/mL, or GS 7.[14] Given that NCCN high-risk patients are not suitable AS candidates, they were excluded from our cohort. Charts were examined for clinical/pathologic data, including age, Charlson Comorbidity Index (CCI), prostatic specific antigen (PSA) at diagnosis, PSA velocity (PSAv) clinical stage, GS at treatment, NCCN risk category at diagnosis, time on AS, number of AS biopsies, progression-to-treatment, and reason for progression (PSA doubling, GS upgrading defined as an increase in total GS, or new primary GS 4). The CCI is a validated tool for the classification and quantification of comorbidities to predict both short- and long-term mortality within 4

6 10 years.[15] Staging was performed according to 1997, 2002, and 2010 TNM classification systems as appropriate depending on year of diagnosis.[16] DT type and year of treatment were recorded along with the reason for progression when discernable from the chart. When ADT was performed concomitantly with XRT, the treatment was classified as XRT. Clinicopathologic features were compared between treated patients and those on continued AS using Student s T test for continuous variables and the Chi-square test for categorical variables. Comparisons were also made among each DT type using one-way ANOVA with the Bonferroni correction. Of note, negative PSAv was encoded as zero. Oncologic outcomes included biochemical recurrence (BCR), metastasis, disease-specific survival (DSS), and overall survival (OS). Oncologic outcomes time was measured from the date of definitive treatment. BCR was defined as two consecutive PSA values 0.2 ng/ml after RP, and PSA nadir ng/ml for XRT, primary ADT, and cryosurgery ( Phoenix Criteria, given that 90% of the cryotherapy cases were performed after these criteria were published in 2005).[17] The Cox proportional hazards model was used to examine the impact of treatment vs. continued AS on outcomes, along with DT type. Variables significant on univariable analysis were included in the multivariable model. Each oncologic outcome was also examined with the Kaplan-Meier Method and log-rank test to determine 10-year freedom from BCR, DSS, and OS. Statistical analysis was performed with Stata 11.0SE (StataCorp LP, College Station, Texas, USA) and significance was defined as p Clinical Treatment Paradigms The schedule for patient visits, digital rectal exam, and serum PSA measurements was at the individual clinician s discretion. The decision for DT was based on clinician judgment and patient preferences without established disease progression criteria such as PSA kinetics, repeat biopsy tumor volume, or repeat biopsy GS upgrading. Explicit mention of any of these parameters as the trigger for intervention was recorded. RP was performed as either open retropubic radical prostatectomy or robotic-assisted with lymph node dissection at clinician s discretion. XRT included all modalities of radiation except stereotactic body radiation therapy (brachytherapy, external beam radiotherapy, 3D-conformal radiotherapy, and intensity modulated radiotherapy); data on specific XRT treatment or dose was not available. Only ADT monotherapy was classified as ADT; of note no patient had metastatic disease upon progression-to-treatment. Cryotherapy was performed as bilateral whole- 5

7 gland cryoablation as primary treatment; no patient had unilateral focal treatment or salvage cryoablation. Additional treatment upon BCR was based on original DT in concert with clinician judgment within standards of care. Those patients treated with XRT were offered ADT with variable triggers for therapy initiation. ADT regimen was chosen at the practitioner s discretion and was heterogeneous in the study cohort. Testosterone levels were monitored according to physician preference, but this data is not included here. RESULTS Of the 255 patients on AS at our institution, a total of 237 met criteria for inclusion in the study (Figure 1). Median follow-up was 71.7±51.8 months. Clinical stage T1 comprised 86% of the cohort and 74% were NCCN low-risk (Table 1). Treated patients had significantly higher PSAv (p=0.007), were more likely to be stage ct2 (p=0.009), and were more likely to have pre-treatment GS 7 (p<0.001). The incidence of PSA doubling was similar between treated and untreated groups. Patients with GS upgrading during AS were more likely to be treated (p<0.001). A total of 95 patients (40%) were treated after median time on AS of 63.4 months (mean 78 months, range months): 48% XRT (8.9% with concomitant ADT), 29% RP, 14% ADT, and 11% cryotherapy (Figure 1). The clinical and pathological features by treatment type are shown in Table 2. Mean duration of AS per treatment type (Table 2)were: 95.8mos for XRT, 52.6mos for RP, 34.0mos for cryotherapy, and 106.5mos for ADT. Duration of AS for XRT was significantly longer than that for RP by one-way ANOVA by the Bonferroni correction (p =0.028)--RP vs. cryo p=0.027, cryo vs. ADT, p= RP patients were younger than patients treated with other modalities (Table 2), but had similar CCI. Clinical T2 patients were more likely to receive XRT than ct1 patients (75% vs. 0-15%). While PSA values at diagnosis for RP and XRT patients were lower than patients treated with ADT, PSA values at treatment initiation were similar among groups. A total of 54 patients (57%) were treated on account of documented PSA doubling or GS upgrading. Regarding oncologic outcomes, Kaplan-Meier analysis showed 10-year BCR rate for XRT of 46.7% compared to 24.8% for RP. Comparison of the BCR-free survival curves for XRT vs. RP during the follow-up period was insignificant by the log-rank test (p=0.99) (Figure 3). Cryotherapy had the highest rate of BCR (50%, 6

8 p=0.03 vs. RP, p=0.003 vs. XRT), and ADT 10-yr BCR was 43.4% (p>0.05 vs. other DTs). A total of six patients (2.5%) experienced metastasis (Figure 1). The 10-year metastasis-free-survival was comparable between treated and untreated patients (60% vs. 100%, p=0.14), and no significant differences were found when comparing DT groups. As only 2 patients (0.8%) experienced DSM in this cohort, Kaplan-Meier analysis for this outcome was not meaningful. The 10-yr OS rates for treated vs. untreated patients were 78.8% vs. 72.9% (p=0.5), respectively. The 10-year OS rates by DT were: RP 95.8%, cryotherapy 75%, ADT 57%, and XRT 42.3% (p=0.05) (Figure 2A). Logrank tests for RP vs. all other comparisons were not significant (Figure 2A; XRT vs. ADT: p =0.98; XRT vs. cryo: p=0.63). Sub-analysis to determine the influence of DT type on BCR and OS based on NCCN risk group and GS 6 vs. 7 did not demonstrate any additional findings. Subclassification by PSAv (Figure 2B) revealed that 10-year OS was superior for RP vs. XRT only among patients with PSAv >2.0 ng/ml/yr (Figure 2C). On univariable analysis, predictors of BCR included GS at treatment, GS upgrading, total number of biopsies, XRT, cryotherapy, and ADT. GS upgrading, XRT, cryotherapy, and ADT were all significant on multivariable analysis (Table 3A). No predictive factors for metastasis were identified on Cox proportional hazards. There were too few disease-specific mortality events to make analysis of predictors meaningful. For overall mortality, age (HR 1.1), CCI (HR 1.3), clinical stage (HR 2.1), receipt of DT (HR 1.9), diagnosis PSA (HR 1.1), XRT (HR 2.2, p=0.011), time on AS (HR 0.99), and total number of AS biopsies (HR 0.56) were significant predictors on univariable analysis. On multivariable analysis, diagnosis PSA, CCI, total number of AS biopsies, and XRT were significant (Table 3B). DISCUSSION Though AS for PCa has become increasingly popular there is no standard definition of disease progression requiring DT, as it is predicated upon the unique entry criteria of each institution s AS protocol.[2-4] For most cohorts, GS progression to 7 and increase in tumor volume by biopsy parameters beyond NCCN very-low risk disease have been the most common triggers for treatment.[19] As evidenced by contemporary AS cohorts, disease progression and need for DT are to be expected as natural consequences of AS.[5-12] The present study revealed median AS duration of about five years with progression-to-treatment of 40%. 7

9 There are a few reasons why treatment after active surveillance may represent a different clinical scenario than treatment early after cancer diagnosis. Under-sampling error from prostate biopsy and the ability of surgery to provide definitive stage and grade diagnosis versus the use of neoadjuvant androgen deprivation therapy with radiotherapy that may provide superior control of locally advanced disease and nodal micrometastasis represent unique advantages that may translate into superior outcomes, especially for men who have had a significant period of time without definitive treatment of their active malignancy. Interestingly, it was seen that patients undergoing XRT had definitive treatment after a longer period of active surveillance. Although these durations were significantly different and may at first appear to bias the results, as noted in Table 2 the clinical Gleason score at treatment and NCCN risk category distributions were not significantly different among the treatment groups. Furthermore, oncologic endpoints were measured from the date of definitive treatment, not including the time on AS. The durations on AS are a result of actual clinical practice decisions; the longer duration for AS prior to XRT likely reflects the older age of these patients and the clinical trends to offer these patients radiation rather than surgery. Few retrospective studies have compared the outcomes of different DT-modalities in AS cohorts. Tosoian and colleagues report of the Johns Hopkins AS cohort revealed 4.2% of men who underwent RP after AS experienced BCR versus 14.6% of men who underwent XRT after median follow-up of 2 years. Of note, while patients receiving RP were significantly younger at enrollment and had lower median PSA before treatment, the proportion of patients with GS 7 was not different between treatment groups.[8] Conversely, Klotz and colleagues reported no difference in BCR-free survival between those undergoing RP versus XRT after AS in the Toronto cohort.[7] In accordance with the Johns Hopkins study,[8] we report a higher 10-year BCR rate of 46.7% in patients undergoing XRT versus 24.8% undergoing RP after a period of AS, although this was not statistically significant in our cohort, likely owing to the small sample size and few absolute number of events that occurred for biochemical recurrence (4 in the RP group and 9 in the XRT group). Although there was a greater proportion of ct2 patients in the XRT group, clinical stage was not significantly predictive of BCR on multivariable analysis; in addition, pretreatment GS was comparable across DT-modalities. This suggests that patients who undergo XRT after AS may not be adequately treated for their PCa. As there were only two DSMs in our cohort, concurrent with other reports of 8

10 very low DSM in AS cohorts,[5,7,18] we cannot comment on significant predictors of this outcome. It remains to be seen whether BCR in AS cohorts is an appropriate surrogate for long-term outcomes. Interestingly, we also discovered that use of XRT was independently associated with worse OS. We do not necessarily attribute this finding to the modality of DT. Rather, as XRT is more commonly performed in men who are elderly or too ill to undergo surgery, this substantiates clinicians ability to select the appropriate therapy for patients in light of their life expectancy. Though we did not find comparatively higher CCI in XRT patients, they were indeed seven years older on average and were on AS longer than RP patients. The correlation of increasing number of AS biopsies with improved OS may be a reflection of good selection for more intensive monitoring by the clinician. Since treated and untreated groups had similar numbers of AS biopsies performed, there is likely no detection bias that would have contributed to the untreated group having occult higher-risk disease. In fact, although most clinical features were comparable between groups, the treated group actually had a higher proportion of ct2, elevated pre-treatment GS, and higher PSAv. Despite these higher-risk features, OS was comparable between the groups, further suggesting that physicians were effective at choosing candidates for DT. Regarding the efficacy of other DT compared to RP, our findings of increased BCR rates for ADT compared to RP may be expected as ADT represents a non-curative therapy with PCa eventually becoming castration-resistant. ADT monotherapy has been associated with greater disease specific mortality (HR 12.5) and overall mortality (HR 2.98) compared to RP in a retrospective cohort study of 3,248 patients from the SEER- Medicare data with about three years of follow-up.[19] Despite the lack of evidence of the benefits of primary ADT, a report of the CaPSURE database revealed that 14.4% of clinically localized PCa patients received primary ADT from [20] Also, higher BCR rates after cryotherapy may be a consequence of residual prostatic tissue (not necessarily malignant) resulting from the urethral warming device, which hinders meaningful comparisons to RP and XRT series.[16] Contemporary reports of 5-year BCR rates after cryotherapy range from 11-35% and are dependent on disease risk category.[21,22] With only a few studies comparing different DT modalities after a period of AS, comparisons of first-line treatment modalities in localized prostate cancer are informative. Although no prospective, randomized trials exist, 9

11 a recent systematic review by Sun and colleagues identified several retrospective studies favoring RP over XRT for DSS and OS with a low strength of evidence (Table 4).[23] While it appeared that the XRT group had more highrisk disease in the study by Kibel et al., after statistical adjustment for this disparity the 10-yr disease-specific survival was still superior for RP. (2.9% vs. 1.8%).[24] Similarly, Cooperberg and colleagues reported absolute differences of disease-specific mortality to be small for men with low-risk disease but substantially increased for men with intermediate- and high-risk disease.[25] More recently, another systematic review and meta-analysis by Wallis et al examined a total of 19 studies with low to moderate risk of bias, totaling n=118,830 patients and demonstrated that both prostate-cancer specific mortality (adjusted HR 2.08) and overall mortality (adjusted HR 1.63) were higher for patients treated with XRT vs. RP. [26] The best randomized evidence for evaluating surgery vs. radiation in the primary setting comes from the recently published Prostate Testing for Cancer and Treatment (ProTeCT) trial.[27] A total of n=1643 men were randomized to active surveillance, RP or XRT. After median 10 years follow-up, rates of clinical progression, development of metastatic disease, prostate-cancer mortality, and allcause mortality were comparable between RP and XRT. The absolute rates of treatment failure appeared different (4.6% for RP vs. 14% for XRT), although this comparison was not explored statistically, making the interpretation of these findings unclear...[27] The PREFERE trial began in 2012 and has randomized ~7600 men with low- /intermediate risk PCa to RP, XRT, brachytherapy, or AS,[28] The role of PSA velocity in stratifying men for risk of BCR in men who progress to treatment after AS has investigated in a few reports. It has been established that PSA velocity >2ng/mL/yr in the year prior to surgery is associated with higher cancer-specific death within five years of surgery.[29] Furthermore, of men with PSADT >4 years who progressed to treatment after AS, 0% experienced biochemical recurrence.[30] We found that stratifying by PSAv >2ng/mL/yr showed improved overall survival outcomes for men treated with RP vs. XRT. The implications of this finding regarding patient selection for treatment after AS require further study. Also, the role of the commercially available genomic classifiers on selecting the most suitable treatment after AS also remains to be explored. Our study has a few limitations. First, our institution did not employ strict entry criteria, surveillance protocol, nor triggers for definitive intervention during the period of study, so caution must be exercised in 10

12 comparing our AS cohort with those of other academic institutions. By our eligibility criteria, men refusing first-line DT and who underwent watchful waiting were also included, likely including men with disease characteristics that would preclude them from contemporary AS protocols. Rather than defining this approach as expectant management or watchful waiting, given that the majority of patients underwent at least two prostate biopsies they were considered to be enrolled in active surveillance. Furthermore, it has been reported that intermediate risk men on active surveillance are at higher risk of adverse outcomes and are thus not candidates for AS at several institutions.[7] These patients were included to mimic actual clinical practice. Our current institutional active surveillance protocol mirrors that of the Hopkins cohort.[8] An additional limitation of the study is the lack of explicit reason for progression-to-treatment for 45% of the patients and absence of biopsy core data. The 57% who progressed based on PSA doubling or GS upgrading may be therefore underreported. Additional reasons for treatment that were not specified may include patient preference/anxiety, PSA kinetics other than doubling, change in DRE, or progression of local symptoms (i.e. hematuria, bladder outlet obstruction, hydronephrosis). Although not included in the multivariable model (since the reason for progression being anxiety could not be fully corroborated by retrospective chart review), post-hoc exploratory analysis revealed progression by PSA/Gleason score vs. patient preference to be predictive of increased rate of BCR on Cox proportional hazards (HR 3.4, p=0.016) and by Kaplan-Meier analysis (p=0.01 by log rank test) (Supplementary Figure 1). The improved pathologic outcomes at RP after progression due to patient preference rather than clinical parameters has been previously published.[31] Although ADT is indeed a non-curative treatment for PCa, its use in the current context is as DT for men with progressive disease not suitable for other interventions. Also information on adjuvant or salvage therapies for those patients with BCR was not available. Radiation protocols and doses were not captured and were likely heterogeneous. Although year of surgery was included to mitigate the impact of changing radiation efficacy over time in multivariable analysis, this limitation still exists. Also, we could have inaccurately characterized the severity of baseline comorbid conditions or suboptimally weighted these conditions to derive the CCI. CONCLUSIONS 11

13 After a period of AS, patients who eventually underwent XRT had higher rates of BCR at 10 years compared to those who underwent other treatment modalities. In addition, higher patient CCI, diagnosis PSA, total number of AS biopsies, and the use of XRT were independently associated with worse OS. Our findings suggest that in patients on AS, XRT was appropriately selected for patients with lower life expectancy, but may represent undertreatment for some men, especially those with PSA velocity >2ng/mL/yr while on active surveillance. Longer follow-up will allow correlation with other endpoints of interest, namely DSS. 12

14 DISCLOSURES Conflicts of Interest: None. There were no outside sources of funding for the study. 13

15 ACKNOWLEDGEMENTS None. 14

16 REFERENCES 1. Cooperberg MR, Moul JW, Carroll PR. The changing face of prostate cancer. J Clin Oncol. 2005;23: Filson CP, Schroeck FR, Ye Z, Wei JT, Hollenbeck BK, Miller DC. Variation in use of active surveillance among men undergoing treatment for early stage prostate cancer. J Urol. 2014;192: Cooperberg MR, Broering JM, Kantoff PW, Carroll PR. Contemporary trends in low risk prostate cancer: risk assessment and treatment. J Urol. 2007;178:S Weiner AB, Patel SG, Etzioni R, Eggener SE. National trends in the management of low- and intermediaterisk prostate cancer in the United States. J Urol. doi: /j.juro Eggener SE, Mueller A, Berglund RK, Ayyathurai R, Soloway C, Soloway MS, et al. A multi-institutional evaluation of active surveillance for low risk prostate cancer. J Urol. 2013;189:S Welty CJ, Cowan JE, Nguyen H, Shinohara K, Perez N, Greene KL, et al. Extended followup and risk factors for disease reclassification in a large active surveillance cohort for localized prostate cancer. J Urol. 2015;193: Klotz L, Vesprini D, Sethukavalan P, Jethava V, Zhang L, Jain S, et al. Long-term follow-up of a large active surveillance cohort of patients with prostate cancer. J Clin Oncol. 2015;33: Tosoian JJ, Trock BJ, Landis P, Feng Z, Epstein JI, Partin AW, et al. Active surveillance program for prostate cancer: an update of the Johns Hopkins experience. J Clin Oncol. 2011;29: Selvadurai ED, Singhera M, Thomas K, Mohammed K, Woode-Amissah R, Horwich A, et al. Mediumterm outcomes of active surveillance for localized prostate cancer. Eur Urol. 2013;64: Loeb S, Folkvaljon Y, Makarov DV, Bratt O, Bill-Axelson A, Stattin P. Five-year nationwide follow-up study of active surveillance for prostate cancer. Eur Urol. 2015;67: Patel MI, DeConcini DT, Lopez-Corona E, Ohori M, Wheeler T, Scardino PT. An analysis of men with clinically localized prostate cancer who deferred definitive therapy. J Urol. 2004;171: Godtman RA, Holmberg E, Khatami A, Stranne J, Hugosson J. Outcome following active surveillance of men with screen-detected prostate cancer. Results from the Göteberg randomized population based prostate cancer screening trial. Eur Urol. 2013;63:

17 13. Satkunasivam R, Kulkarni GS, Zlotta AR, et al. Pathological, oncologic and functional outcomes of radical prostatectomy following active surveillance. J Urol, 2013: 190(1): D Amico AV, Whittington R, Malkowicz SB, Cote K, Loffredo M, Schultz D, et al. Biochemical outcome after radical prostatectomy or external beam radiation therapy for patients with clinically localized prostate carcinoma in the prostate specific antigen era. Cancer. 2002; 95: Charlson ME, Pompei P, Ales KL, McKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: Development and validation. J Chron Dis. 1987;40: Prostate. In: Edge SB, Byrd DR, Compton CC, et al., eds.: AJCC Cancer Staging Manual. 7th ed. New York, NY: Springer, 2010, pp Babaian RJ, Donnelly B, Bahn D, Baust JG, Dineen M, Ellis D, et al. Best Practice Statement on Cryosurgery for the treatment of localized prostate cancer. J Urol. 2008;180: Dall Era MA, Albertsen PC, Bangma C, Carroll PR, Carter HB, Cooperberg MR, et al. Active surveillance for prostate cancer: a systematic review of the literature. Eur Urol. 2012;62: Liu J, Shi L, Sartor O, Culbertson R. Androgen-deprivation therapy versus radical prostatectomy as monotherapy among clinically localized prostate cancer patients. Onco Targets Ther. 2013;6: Cooperberg MR, Broering JM, Carroll PR. Time trends and local variation in primary treatment of localized prostate cancer. J Clin Oncol. 2010;28: Bahn DK, Lee R, Badalament A, Kumar A, Greski J, Chernick M. Targeted cryoablation of the prostate: 7- year outcomes in the primary treatment of prostate cancer. Urology. 2002;60: Donnelly BJ, Saliken JC, Ernst DS, Ali-Ridha N, Brasher PM, Robinson JW, et al. Prospective trial of cryosurgical ablation of the prostate: five-year results. Urology. 2002;60: Sun F, Oyesanmi O, Fontanarosa J, Reston J, Guzzo J, Schoelles K. Therapies for clinically localized prostate cancer: Update of a 2008 systematic review [Internet]. Comparative Effectiveness Reviews. 2014;146: Kibel AS, Ciezki JP, Klein EA, Reddy CA, Lubahn JD, Haslag-Minoff J, et al. Survival among men with clinically localized prostate cancer treated with radical prostatectomy or radiation therapy in the prostate specific antigen era. J Urol. 2012;187:

18 25. Cooperberg MR, Vickers AJ, Broering JM, Carroll PR. Comparative risk-adjusted mortality outcomes after primary surgery, radiotherapy, or androgen-deprivation therapy for localized prostate cancer. Cancer. 2010;116; Wallis CJD, Saskin R, Choo R, et al. Surgery versus radiotherapy for clinically-localized prostate cancer: A systematic review and meta-analysis, Eur Urol, (1): Hamdy FC, Donovan JL, Lane JA, et al. 10-year outcomes after monitoring, surgery, or radiotherapy for localized prostate cáncer. NEJM, 2016; 375(15): [accessed April 8th, 2017] 29. D'Amico AV, Chen MH, Roehl KA, Catalona WJ. Preoperative PSA velocity and the risk of death from prostate cancer after radical prostatectomy. N Engl J Med 2004; 351: Khatami A, Gunnar A, Damber JE, Lilja H, Lodding P, Hugosson J. PSA doubling time predicts the outcome after active surveillance in screening-detected prostate cancer: results from the European randomized study of screening for prostate cancer, Sweden section. Int J Cancer. 2007; 120: Bul M, Zhu X, Rannikko A, Staerman F, Valdagni R, Pickles T, et al. Radical prostatectomy for low-risk prostate cancer following initial active surveillance: results from a prospective observational study. Eur Urol, 2012; 62(2): Stattin P, Holmberg E, Johansson JE, Holmberg L, Adolfsson J, Hugosson J, National Prostate Cancer Register (NPCR) of Sweden. Outcomes in localized prostate cancer: National Prostate Cancer Register of Sweden follow-up study. J Natl Cancer Inst. 2010;102: Ferrer M, Guedea F, Suarez JF, de Paula B, Macias V, Marino A, et al. Quality of life impact of treatments for localized prostate cancer: cohort study with a 5 year follow-up. Radiother Oncol. 2013;108: Hoffman RM, Koyama T, Fan KH, Albertsen PC, Barry MJ, Goodman M, et al. Mortality after radical prostatectomy or external beam radiotherapy for localized prostate cancer. Natl Cancer Inst. 2013;105: DeGroot JM, Brundage MD, Lam M, Rohland SL, Heaton J, Mackillop WJ, et al. Prostate cancer-specific survival differences in patients treated by radical prostatectomy versus curative radiotherapy. Can Urol Assoc J. 2013;7:E

19 Figure Legends Figure 1. CONSORT diagram of the cohort. Figure 2A. Kaplan Meier curve for OS stratified by DT-modality after AS. Figure 2B. OS stratified by PSA velocity >2ng/mL/yr. Figure 2C. OS for RP vs. XRT among patients with PSAv >2ng/mL/yr Figure 3. BCR-free survival for RP vs. XRT. Supplementary Figure 1. BCR-free survival by Reason for Progression to Treatment Table 1. Patient clinical and pathological features by definitive treatment after active surveillance Table 2. Patient clinical and pathological features by treatment type after active surveillance Table 3A. Multivariable analysis for BCR Table 3B. Multivariable analysis for OM Table 4. Prior retrospective studies comparing different primary treatment modalities for prostate cancer. Supplementary Table. Cox proportional Hazards models and Schoenfeld residuals to test the proportional Hazards Assumption 18

20 Variable Overall Treated Untreated p n=237 n=95 n=142 Age, yrs 69.8± ± ± CCI 7±2 7.1±2 6.7± PSA at diagnosis, ng/ml 5.6± ± ± PSA pre-treatment/terminal, ng/ml 8.5± ± ± PSAv, ng/ml/yr 3.79± ± ± * Clinical stage, no. (%) 0.009** ct1 ct2 205 (86.1) 33 (13.9) 75 (79.0) 20 (21.0) 130 (91.0) 13 (9.0) Clinical GS at diagnosis, no. (%) (83.5) 39 (16.5) 78 (82.1) 17 (17.9) 120 (84.5) 22 (15.5) NCCN risk category, no. (%) 0.52 Low Intermediate 175 (73.8) 62 (26.2) 68 (71.6) 27 (28.4) 107 (75.4) 35 (24.6) Clinical GS at treatment, no. (%) <0.001** (65.4) 75 (31.7) 7 (3.0) 47 (49.5) 41 (43.2) 7 (7.3) 108 (76.1) 34 (23.9) 0 Disease progression, no. (%) PSA doubling GS upgrading 40 (16.9) 58 (24.5) 15 (15.8) 43 (45.3) 25 (17.6) 15 (10.6) 0.71 <0.001** Total # of AS biopsies 2.3± ± ± Pathological GS, no. (%) (29.6) 17 (63.0) 2 (7.4) N/A N/A N/A -- Table 1. Patient clinical and pathological features by definitive treatment after active surveillance. CCI = Charlson Comorbidity Index; PSA = prostate specific antigen; PSAv = PSA velocity; GS = Gleason Score; NCCN = National Comprehensive Cancer Network (note that NCCN risk category is at diagnosis); AS = active surveillance; *Significant by Student s T test; **Significant by Chi square test

21 Definitive Treatment Type Variable RP XRT Cryo ADT P n=27 n=45 n=10 n=13 Age, yrs 64.0± ± ± ± * CCI 6.6± ± ± ± PSA at diagnosis, ng/ml 5.0± ± ± ± * PSA pre-treatment/terminal, ng/ml 7.9± ± ± ± PSAv, ng/ml/yr 7.5± ±53 1.9± ± Clinical stage, no. (%) 0.035** ct1 ct2 24 (89.0) 3 (11.0) 30 (66.7) 15 (33.3) 10 (100) 0 (0) 11 (84.6) 2 (15.4) Clinical GS at diagnosis, no. (%) (85.2) 4 (14.8) 36 (80.0) 9 (20.0) 8 (80.0) 2 (20.0) 11 (39.3) 17 (60.7) NCCN risk category, no. (%) 0.65 Low Intermediate 21 (77.8) 6 (22.2) 31 (68.9) 14 (31.1) 8 (80.0) 2 (20.0) 8 (61.5) 5 (38.5) Clinical GS at treatment, no. (%) (51.9) 11 (40.7) 2 (7.4) 22 (48.9) 21 (46.7) 2 (4.4) 2 (20.0) 5 (50.0) 3 (30.0) 9 (69.2) 4 (30.8) 0 (0) AS time, months 52.6± ± ± ±47.6 <0.001* Total # of AS biopsies 2.8± ± ± ±1.2 NS Table 2. Patient clinical and pathological features by treatment type after active surveillance. CCI = Charlson Comorbidity Index; PSA = prostate specific antigen; PSAv = PSA velocity; GS = Gleason Score; NCCN = National Comprehensive Cancer Network (note that NCCN risk category is at diagnosis). AS = active surveillance *Significant by one-way Analysis of Variance (ANOVA); **Significant by Chi square test. NS = Non-significant for all ANOVA comparisons using Bonferroni correction.

22 Variable HR 95% CI p Pre-treatment GS GS upgrading Total # of AS Bx PSA velocity XRT Cryo <0.001 ADT <0.001 Table 3A. Multivariable analysis for Biochemical Recurrence (BCR). HR = hazard ratio; Std Err. = standard error; 95% CI = 95% confidence interval. GS = Gleason score; AS Bx = active surveillance biopsies; PSA = prostate specific antigen; XRT = radiation therapy; Cryo = cryotherapy; ADT = androgen deprivation therapy Variable HR 95% CI p Age CCI Clinical Stage PSA at diagnosis Total # of AS Bx XRT Table 3B. Multivariable analysis for Overall Mortality. HR = hazard ratio; Std Err. = standard error; 95% CI = 95% confidence interval. CCI = Charlson Comorbidity Index; PSA = prostate specific antigen; AS Bx = active surveillance biopsies; XRT = radiation therapy

23 Study D Amico, et al. (2002)[13] Kibel, et al. (2012)[24] Cooperberg, et al. (2010)[25] Stattin, et al. (2010)[32] Ferrer, et al. (2013)[33] Hoffman, et al. (2013)[34] DeGroot, et al. (2013)[35] N (study pd) 2,635 ( 88-00) 10,429 ( 95-05) 7,538 ( 87-07) 6,849 ( 97-02) 704 ( 03-05) 1,655 ( 94-10) 2,213 ( 90-98) Risk Group Low & Int All All Low & Int Low & Int All Low & Int FU 4 yr 5.6yr yr 8.2yr 6yr 15yr 4.3yr Outcom e 8-yr PSA survival 10-yr CSS, OS CSM & ACM CSM & ACM BCR, OS CSM & ACM CSM Difference Favor Notes Low: 88% vs. 78% Int: 79% vs 65% RP 2.9% vs. 1.8% RP CSM: XRT HR 2.2 vs. RP ACM: XRT HR 1.6 vs RP CSM: RP ARR 0.72 ACM: RP ARR 0.71 BCR: XRT HR 2.1 vs RP OS: not significant CSM: RP HR 0.35 ACM: RP HR 0.6 Low: not significant Int: not significant RP RP None RP None Table 4. Prior retrospective studies comparing different primary treatment modalities for prostate cancer. FU = follow up; Int = intermediate; PSA = prostate specific antigen; CSS = cancer-specific survival; OS = overall survival; CSM = cancer-specific mortality; ACM = all-cause mortality; BCR = biochemical recurrence; XRT = radiotherapy; RP = radical prostatectomy; HR = hazard ratio; Gy = gray; CAPRA = Cancer of the prostate risk assessment EBRT = external beam radiotherapy; RRP = radical retropubic prostatectomy XRT 66Gy median XRT 74Gy median HR adjusted for Kattan & CAPRA XRT dose unk RRP vs EBRT XRT dose unk XRT dose unk

24

25

26

27

28

29 A. Variable HR 95% CI p RP <0.001 XRT <0.001 B Cryo <0.001 ADT <0.001 B. Variable Rho Chi2 p RP XRT Cryo ADT Global Test C. Variable HR 95% CI p RP XRT Cryo ADT D. Variable Rho Chi2 p RP XRT Cryo ADT Global Test Supplementary Table. Cox proportional Hazards models and Schoenfeld residuals to test the proportional Hazards Assumption A. Cox proportional Hazards model for Biochemical Recurrence (BCR)- free Survival including treatment type as covariates and B. Schoenfeld residuals demonstrating that proportional hazards assumption has not been violated (Global test, p>0.05). C. Cox proportional

30 Hazards model for Overall Survival including treatment type as covariates and B. Schoenfeld residuals demonstrating that proportional hazards assumption has not been violated (p>0.05). RP = radical prostatectomy; XRT = radiation therapy; cryo = cryotherapy; ADT = androgen deprivation therapy. Chi2 = Chi squared test; Rho = Spearman's rank correlation coefficient. P <0.05 = significant.

31