- Email: [email protected]
Population-based Comparative Effectiveness of Salvage Radical Prostatectomy vs Cryotherapy
Radical Prostatectomy Population-based Comparative Effectiveness of Salvage Radical Prostatectomy vs Cryotherapy David F. Friedlander, Xiangmei Gu, Sandip M. Prasad, Stuart R. Lipsitz, Paul L. Nguyen, Quoc-Dien Trinh, Maxine Sun, and Jim C. Hu OBJECTIVE METHODS
RESULTS
CONCLUSION
To characterize population-based practice patterns, disease-specific and overall mortality, and cost associated with salvage cryotherapy (SCT) vs salvage radical prostatectomy (SRP). We retrospectively identified 440 men who failed primary radiation therapy and subsequently underwent SCT (n ¼ 341, 77.5%) or SRP (n ¼ 99, 22.5%) between 1992 and 2009 from Surveillance, Epidemiology, and End ResultseMedicare linked data. Propensity score analyses were used to compare overall and prostate cancer-specific mortality and associated Medicare expenditures for SRP vs SCT. Men undergoing SCT were more likely to be white (P <.001), less likely to be high school graduates (P ¼ .008), and experienced shorter median time from diagnosis to salvage therapy (44.1 vs 60.1, P <.001) and from primary radiotherapy to salvage therapy (38.7 vs 55.8 months, P <.001). In adjusted analyses, overall mortality was higher (21.6 vs 6.1 deaths/100 person years, P <.001) for SRP vs SCT. There was a trend for higher prostate cancer-specific death rates with SRP vs SCT (6.5 vs 1.4 deaths/100 person years, P ¼ .061). Medicare expenditures for SRP vs SCT were more than 2-fold higher ($19,543 vs $8,088, P <.001). SRP vs SCT is associated with higher overall mortality and greater health care expenditures. However, longer follow-up is needed to assess long-term functional outcomes and cancer control. UROLOGY 83: 653e657, 2014. 2014 Elsevier Inc.
P
rostate cancer is the most prevalent solid organ tumor among US men, with an estimated incidence of 241,740 cases and 28,170 prostate cancer-specific deaths in 2012.1 Owing to stage migration observed over the past 3 decades, most incident prostate cancers present as localized disease.2 Despite the popularity of traditional radiation therapies (external-beam and brachytherapy) as primary treatment options for localized disease,3 63% of men will experience biochemical recurrence within 10 years of radiotherapy,4
Financial Disclosure: Jim C. Hu receives salary support from a Department of Defense Prostate Cancer Physician Training Award (W81XWH-08-1-0283). The remaining authors declare that they have no relevant financial interests. Jim C. Hu 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. From the Division of Urology, Brigham and Women’s Hospital, Boston, MA; the Center for Surgery and Public Health, Brigham and Women’s Hospital, Boston, MA; the Department of Urology, Medical University of South Carolina, Charleston, SC; the Section of Urology, Ralph A. Johnson VA Medical Center, Charleston, SC; the Department of Radiation Oncology, Lank Center for Genitourinary Oncology, Dana Farber Cancer Institute, Boston, MA; the Division of Urologic Surgery, Center for Surgery and Public Health, Brigham and Women’s Hospital, Boston, MA; the Department of Public Health, Faculty of Medicine, University of Montreal, Montreal, Quebec, Canada; and the Department of Urology, David Geffen School of Medicine at University of California Los Angeles, Los Angeles, CA Reprint requests: Jim C. Hu, M.D., M.P.H., Department of Urology, David Geffen School of Medicine at University of California Los Angeles, 924 Westwood Blvd, STE 1000, Los Angeles, CA 90024. E-mail: [email protected] Submitted: August 9, 2013, accepted (with revisions): November 18, 2013
ª 2014 Elsevier Inc. All Rights Reserved
and it is estimated that 25%-32% will experience local failure.5 Salvage radical prostatectomy (SRP) and salvage cryotherapy (SCT) are performed with curative intent for clinically localized radiorecurrent prostate cancer, with 5-year progression-free survival approaching 60% for both.6,7 Although SRP historically has a longer follow-up with acceptable oncologic outcomes, it is accompanied by significant morbidity owing to radiation-induced fibrosis and tissue-plane obliteration.8 Conversely, SCT (particularly third-generation technology) has gained popularity, possibly because of improved technique and fewer complications.9 Although several systematic reviews have assessed SRP and SCT morbidity and survival,6,10 there is a dearth of directly comparative studies. Moreover, consensus regarding the optimal management of primary radiotherapy failures remains elusive.11 Through our population-based approach that provides insight beyond referral centers in which salvage therapy outcomes largely originate, we sought to compare the use, patterns of care, outcomes, and costs of SRP vs SCT.
MATERIALS AND METHODS Data Our study was approved by the University of California, Los Angeles Review Board; patient data were deidentified, and the 0090-4295/14/$36.00 http://dx.doi.org/10.1016/j.urology.2013.11.019
653
requirement for consent was waived. We used Surveillance, Epidemiology, and End Results (SEER)eMedicare data for analysis, which currently comprises a linkage of cancer registry data from 20 SEER regions with Medicare administrative data. The Medicare program provides benefits to 97% of Americans aged 65 years.12 We identified 345,803 men aged 65 years who were diagnosed with prostate cancer between 1992 and 2007, with Medicare follow-up through 2009. After excluding men not enrolled in both Medicare Part A and B or who were enrolled in a Medicare health maintenance organization because of unreliable claims submissions during the study period, 103,508 men treated with primary radiotherapy were subsequently identified. Of this group, 430 were found to have received SRP (n ¼ 99) or SCT (n ¼ 341). Patients were grouped into SRP or SCT cohorts based on International Classification of Diseases, Ninth Revision (ICD-9), Physicians Current Procedural Terminology Coding System, 4th edition, and Healthcare Common procedure Coding System codes (60.62, C2618, 55840, 55842, 55845, and 55873), with SRP defined as surgery after primary radiotherapy (external beam radiotherapy, brachytherapy, and/or intensity-modulated radiotherapy). We excluded perineal and minimally invasive radical prostatectomy, as these were uncommon in the salvage setting, totaling 25 procedures. We restricted our cohort to men with prostate cancer diagnosed as their only cancer.
Outcomes We examined overall and prostate cancer-specific mortality after SRP vs SCT. In addition, we characterized Medicare expenditures associated with each treatment.
Control Variables Age was obtained from the Medicare file; race, census tract measures of median household income and high school education, region, population density (urban vs rural), and marital status were obtained from SEER registry data. Comorbidity was assessed using the Klabunde modification of the Charlson index during the year before surgery.13 ICD-9 codes were used to identify disease categories, whereas Physicians Current Procedural Terminology Coding System, 4th edition and Healthcare Common Procedure Coding System code sets were used to identify medical, surgical, and diagnostic services. To increase specificity, only imaging studies designated with a corresponding ICD-9 code for prostate cancer were included. Adjuvant androgen deprivation therapy (ADT) was defined as ADT use anytime between 6 months before and after primary therapy. “Additional ADT” was defined as ADT use at 2 years after SCT and use anytime after SRP.
experiment. Propensity score adjustment was performed using a logistic regression model to calculate the probability of undergoing SRP vs SCT on the basis of described covariates and then weighting the data on the basis of the inverse propensity of being in either of the treatment groups.17 After adjustment, covariate balance was assessed. All tests were considered statistically significant at a ¼ 0.05. All analyses were performed with SAS version 9.2 (SAS Institute, Cary, NC). Because of confidentiality, values <11 may not be reported directly or in a derivable way. Therefore, for outcomes with <11 patients, we present odds ratios.
RESULTS Among men experiencing radiotherapy failure, 341 (77.5%) underwent SCT and 99 (22.5%) SRP. Median follow-up for SRP vs SCT was 30 (interquartile range 1844.4) vs 15 (interquartile range 4.8-33.6) months after salvage therapy. Although men undergoing SRP vs SCT experienced longer median time from primary to salvage therapy (55.8 vs 38.7 months, P <.001) and from diagnosis to salvage therapy (60.1 vs 44.1 months, P <.001), they experienced similar median time from prostate cancer diagnosis to primary therapy (3.0 months for both, P ¼ .132). Men receiving SCT vs SRP (Table 1) were more likely to be white (P <.001) and to live in areas with <85% high school graduation rates (P ¼ .008). In addition, men undergoing SCT vs SRP were more likely to have received previous ADT (50.4% vs 45.5%, P ¼ .001) and primary brachytherapy (43.7% vs 24.2%, P ¼ .001). Propensity adjusted analyses is presented in Table 2. Overall survival was reduced (21.57 vs 6.14 deaths/100 person years, P <.001) for SRP vs SCT. Similarly, there was a trend for higher disease-specific mortality for SRP vs SCT (6.54 vs 1.37 deaths/100 person years, P ¼ .061). Health care expenditures in the 3 months before salvage therapy were higher for SRP vs SCT (median $8,416 vs $2,363, P <.001). Similarly, SRP costs were 2fold higher than SCT (median $19,543 vs $8,088, P <.001). SRP vs SCT patients were more likely to require inpatient care in the 6 months surrounding surgery (58.6% vs 32.0%), at a median cost of $23,321 vs $10,497 (P <.001).
COMMENT Expenditures To best attribute the costs associated with each surgical setting, we assessed Medicare payments within 3 months of salvage therapy, which represents the traditional global payment period.14
Statistical Analysis Unadjusted analysis using the Pearson chi-square statistic was performed to compare demographic and tumor characteristics for SRP vs SCT.15 Adjusted analyses were performed with weighted propensity scoring.16 Propensity score methods control for all observed confounding factors that might influence cohort assignment and outcome using a single composite measure, balancing patient characteristics as would occur in a randomized 654
The optimal management of radiorecurrent prostate cancer remains controversial, particularly for younger men or those with life expectancy >10 years.11 Current salvage treatment options include ADT, brachytherapy, cryoablative therapy, high-intensity focused ultrasound, and radical prostatectomy. Although SRP has the longest documented follow-up with acceptable oncologic outcomes, its use has been limited by greater perioperative morbidity, including rectal injury, urethrocutaneous fistula, lymphedema, urinary incontinence, and anastomotic stricture.18,19 The use of SCT has increased in recent years owing to comparable oncologic outcomes and superior complication rates.5 However, in the UROLOGY 83 (3), 2014
Table 1. Demographics of study population Independent Variables n (%) Year of diagnosis 2002-2005 2006-2009 Age at diagnosis 65-69 70-74 75þ Race Nonwhite White Median income <$35,000 $35,000-$44,000 $45,000-$59,999 $60,000 Education (%) <85 85 Population density Urban Region Northeast South Midwest West Marital status Not married Married Clinical stage T1 T2-T3 Grade Well differentiated Poorly/ undifferentiated ADT before salvage therapy Primary therapy Brachytherapy External beam Combination Median primary to salvage therapy time interval (mo)
Salvage Radical Salvage Cryotherapy Prostatectomy n ¼ 99 P Value n ¼ 341 110 (32.26) 231 (67.74)
30 (30.30) 69 (69.70)
.713
144 (42.23) 112 (32.84) 85 (24.93)
40 (40.40) 40 (40.40) 19 (19.19)
.302
82 (24.05) 259 (75.95)
* *
109 92 73 66
(26.26) (20.20) (25.25) (28.28)
.135
154 (45.29) 186 (54.71)
30 (30.30) 69 (69.70)
.008
308 (90.32)
*
.901
76 81 52 132
(32.06) (27.06) (21.47) (19.41)
(22.29) (23.75) (15.25) (38.71)
26 20 25 28
<.001
21 16 23 39
Events per 100 Person Years
Salvage Radical Salvage Cryotherapy Prostatectomy n ¼ 99 P Value n ¼ 341
Unadjusted Overall mortality Prostate cancer-specific mortality Additional ADT Adjusted Overall mortality Prostate cancer-specific mortality Additional ADT
5.76 1.33
22.54 6.17
<.001 .058
6.22
9.49
.162
6.14 1.37
21.57 6.54
<.001 .061
6.3
8.79
.273
Abbreviation as in Table 1.
(21.21) (16.16) (23.23) (39.39)
.177
54 (16.93) 265 (83.07)
16 (17.39) 76 (82.61)
.917
140 (42.55) 189 (57.45)
43 (44.79) 53 (55.21)
.697
207 (61.79)
68 (70.10)
.134
128 (38.21)
29 (29.90)
172 (50.44)
45 (45.45)
.001
149 (43.70) 146 (42.82) 46 (13.49) 38.7
24 (24.24) 51 (51.52) 24 (24.24) 55.8
.001 <.001
ADT, androgen deprivation therapy. * Race and population density for salvage radical prostatectomy are not shown, as there are <11 patients within a category.
absence of prospective perioperative and survival data,7 direct comparison of SRP and SCT is difficult. To our knowledge, our study is the first to directly compare the mortality outcomes between salvage therapies for locally recurrent prostate cancer after initial radiation therapy failure, and our study has several important findings. First, we observed a trend in overall and prostate cancer-specific death that was more common among SRP vs SCT men. Although previous investigators have focused on traditional markers of treatment failure such as UROLOGY 83 (3), 2014
Table 2. Unadjusted and propensity score adjusted outcomes of salvage cryotherapy vs radical prostatectomy
postsalvage therapy biopsy positivity and prostate-specific antigen (PSA) levels, we examine overall and cancerspecific mortality. Our findings differ from those reported in a retrospective single-center study by Pisters et al20, who demonstrated that SRP yielded better overall survival but comparable disease-specific survival when compared with SCT. However, tumor grade was significantly higher in their SCT cohort, potentially confounding their results given the association between high tumor grade and PSA failure after SCT.21 Moreover, their analysis was limited to men with a presalvage PSA level of <10 ng/mL and history of previous radiation therapy alone, making direct comparison with our study difficult given the absence of presalvage therapy PSA values. To the best of our knowledge, previous studies did not use the Phoenix definition of PSA nadir plus 2 ng/mL. This might account for the significantly higher recurrence rates reported by Pisters et al relative to a recent retrospective analysis of SCT conducted by Williams et al.22 This discrepancy demonstrates the importance of establishing standardized definitions of treatment failure. Second, salvage therapy was rarely used among men who failed primary radiotherapy, with just 0.4% subsequently undergoing SRP or SCT. This finding is lower than salvage therapy rates noted in previous populationbased studies.4 As noted by Jang et al23, physician perceptions regarding optimal primary therapy for localized prostate cancer vary widely across specialties and geographic regions. It can be inferred that a similar level of equivocality surrounds salvage therapies and might contribute to lower rates of utilization. Furthermore, Jang et al observed increasing frequency of conservative management with age among men seen exclusively by urologists. The SEER cohort consists of men older than 65 years, and the infrequency of salvage therapy seen in our study might reflect higher rates of conservative management in this population. Finally, Mouraviev et al24 note in a recent review that life expectancy >10 years might confer a greater likelihood of favorable biochemical disease-free survival in the setting of salvage 655
therapy, and it is conceivable that our low salvage therapy utilization rate is the result of low 10-year survival rate projections within the SEER population. Third, men undergoing SRP vs SCT experienced longer median time from primary to salvage therapy and from diagnosis to salvage therapy, although time from diagnosis to primary therapy was similar to SCT men. These findings might represent greater reluctance to perform SRP, which is more fraught with complications and technical challenges as compared with SCT.5 Previous reports have postulated that serum PSA level of <10 ng/mL, Gleason score of 8, and clinical stage T1c or T2 before salvage therapy are predictors of improved biochemical recurrence-free survival after salvage therapy.9 These findings suggest that salvage therapy is likely to be more effective in the setting of less advanced radiorecurrent disease. Although we control for differences in primary to salvage therapy time intervals experienced by our 2 treatment cohorts, our adjusted analysis might not account for potential differences in disease severity secondary to longer primary to salvage therapy intervals, as assessed by possible PSA value differences between our 2 treatment groups. In addition, a longer primary to salvage interval might contribute to greater overall morbidity and therefore have a deleterious effect on lifestyle, both of which were not controlled for in our analysis. Finally, health care expenditures for SRP vs SCT were more than 2-fold higher, with a difference of approximately $11,000. This finding is interesting given higher salvage ADT use among SCT men, as previous studies have shown significant added cost associated with salvage ADT.25 Differences in cost between SCT and SRP might be partially explained by the fact that SRP patients were more likely to require inpatient care in the 6 months surrounding salvage therapy. Although this figure might not capture all payments associated with complications beyond 3 months, it includes the costs of postoperative clinic visits, emergency room visits, readmission, and additional surgical or radiologic procedures. Nonetheless, selective referral to experienced SRP surgeons might reduce the significant morbidity and costs observed in our study.26 Our findings must be interpreted in the context of the study design. First, analyses were restricted to Medicare beneficiaries older than 65 years residing in SEER regions, and our findings might not be applicable to younger men. Similarly, there might be selection bias for younger men to undergo SRP because of less comorbidities and postoperative incontinence relative to older men.27 Because our study of Medicare beneficiaries omits these men, the recorded overall and disease-specific mortality might consequently reflect the inherently more difficult nature of SRP in older individuals. Neither clinical staging data nor biopsy Gleason scores after failure of primary radiotherapy (before salvage therapy) were available for analysis. Moreover, recurrence PSA values were unavailable before salvage therapy. Previous studies have shown that 656
lower pre-SRP PSA levels confer improved oncologic outcomes.28 Similarly, higher SCT failure rates are experienced by men with postradiation PSA levels >10 ng/mL or PSA doubling times of 16 months.5 Taken together in the absence of presalvage therapy PSA values, it is impossible to know whether the unfavorable subgroups noted by Chade, Pisters, and Finley are disproportionately over or underrepresented in either of our study cohorts. Similarly, because our regression does not account for the temporality of salvage ADT, our results are subject to the confounding influence of hormonal therapy.29 In addition, in the absence of data from validated quality of life instruments, we were unable to compare functional outcomes after SRP vs SCT. Finally, as with any adjusted analysis, propensity score methods do not control for unmeasured confounders, such as lifestyle factors that might effect overall survival and possess other limitations.30
CONCLUSION Although men undergoing SRP vs SCT are less likely to require salvage ADT, they incur more total health care expenditures and experience a higher rate of overall death, even after adjusting for differences in follow-up and time between primary to salvage therapy. Patients and providers alike should consider these populationbased findings when discussing salvage options for radiorecurrent prostate cancer. References 1. Siegel R, Naishadham D, Jemal A. Cancer statistics, 2012. CA: A Cancer J Clin. 2012;62:10-29. 2. Galper SL, Chen MH, Catalona WJ, et al. Evidence to support a continued stage migration and decrease in prostate cancer specific mortality. J Urol. 2006;175:907-912. 3. Cooperberg MR, Broering JM, Carroll PR. Time trends and local variation in primary treatment of localized prostate cancer. J Clin Oncol. 2010;28:1117-1123. 4. Agarwal PK, Sadetsky N, Konety BR, et al. Cancer of the Prostate Strategic Urological Research Endeavor (CaPSURE) treatment failure after primary and salvage therapy for prostate cancer: likelihood, patterns of care, and outcomes. Cancer. 2008;112: 307-314. 5. Finley DS, Pouliot F, Miller DC, Belldegrun AS. Primary and salvage cryotherapy for prostate cancer. Urol Clin North Am. 2010; 37:67-82. 6. Finley DS, Belldegrun AS. Salvage cryotherapy for radiationrecurrent prostate cancer: outcomes and complications. Curr Urol Rep. 2011;12:209-215. 7. Nguyen PL, D’Amico AV, Lee AK, et al. Patient selection, cancer control, and complications after salvage local therapy for postradiation prostate-specific antigen failure. Cancer. 2007;110:14171428. 8. Cox JM, Busby JE. Salvage therapy for prostate cancer recurrence after radiation therapy. Curr Urol Rep. 2009;10:199-205. 9. Kimura M, Mouraviev V, Tsivian M, et al. Current salvage methods for recurrent prostate cancer after failure of primary radiotherapy. BJU Int. 2010;105:191-201. 10. Heidenreich A, Richter S, Thuer D, et al. Prognostic parameters, complications, and oncologic and functional outcome of salvage radical prostatectomy for locally recurrent prostate cancer after 21stcentury radiotherapy. Eur Urol. 2010;57:437-443.
UROLOGY 83 (3), 2014
11. Stephen J. Radiorecurrent prostate Cancer: an emerging and largely mismanaged epidemic. Eur Urol. 2011;60:411-412. 12. Warren JL, Klabunde CN, Schrag D, et al. Overview of the SEERMedicare data: content, research applications, and generalizability to the United States elderly population. Med Care. 2002;40(8 Suppl):IV-3-IV18. 13. Klabunde CN, Potosky AL, Legler JM, et al. Development of a comorbidity index using physician claims data. J Clin Epidemiol. 2000;53:1258-1267. 14. Prasad SM, Gu X, Lavelle R, et al. Comparative effectiveness of perineal versus retropubic and minimally invasive radical prostatectomy. J Urol. 2011;185:111-115. 15. Rao JNK, Scott AJ. The analysis of categorical-data from complex sample-surveys e Chi-squared tests for goodness of fit and independence in 2-way tables. JASA. 1981;76:221-230. 16. Rubin DB. Estimating causal effects from large data sets using propensity scores. Ann Intern Med. 1997;127:757-763. 17. Robins JM, Hernan MA, Brumback B. Marginal structural models and causal inference in epidemiology. Epidemiology. 2000;11: 550-560. 18. Rainwater LM, Zincke H. Radical prostatectomy after radiation therapy for cancer of the prostate: feasibility and prognosis. J Urol. 1998;140:1455-1459. 19. Vaidya A, Soloway MS. Salvage radical prostatectomy for radiorecurrent prostate cancer: morbidity revisited. J Urol. 2000;164: 1998-2001. 20. Pisters LL, Leibovici D, Blute M, et al. Locally recurrent prostate cancer after initial radiation therapy: a comparison of salvage radical prostatectomy versus cryotherapy. J Urol. 2009;182: 517-525.
UROLOGY 83 (3), 2014
21. Pisters LL, Perrotte P, Scott SM, et al. Patient selection for salvage cryotherapy for locally recurrent prostate cancer after radiation therapy. J Clin Oncol. 1999;17:2514-2520. 22. Williams AK, Martınez CH, Lu C, et al. Disease-free survival following salvage cryotherapy for biopsy-proven radio-recurrent prostate cancer. Eur Urol. 2011;60:405-410. 23. Jang TL, Bekelman JE, Liu Y, et al. Physician visits prior to treatment for clinically localized prostate cancer. Arch Intern Med. 2010; 170:440-450. 24. Mouraviev V, Spiess PE, Jones JS. Salvage cryoablation for locally recurrent prostate cancer following primary radiotherapy. Eur Urol. 2012;61:1204-1211. 25. Krupski TL, Foley KA, Baser O, et al. Health care cost associated with prostate cancer, androgen deprivation therapy and bone complications. J Urol. 2007;178:1423-1428. 26. Williams SB, Gu X, Lipsitz SR, et al. Utilization and expense of adjuvant cancer therapies following radical prostatectomy. Cancer. 2011;117:4846-4854. 27. Anderson CB, Kaufman MR, Dietrich MS, et al. Recovery of urinary function after radical prostatectomy: identification of trajectory cluster groups. J Urol. 2012;187:1346-1351. 28. Chade DC, Shariat SF, Cronin AM, et al. Salvage radical prostatectomy for radiation-recurrent prostate cancer: a multi-institutional collaboration. Eur Urol. 2011;60:205-210. 29. D’Amico AV, Schultz D, Loffredo M, et al. Biochemical outcome following external beam radiation therapy with or without androgen suppression therapy for clinically localized prostate cancer. JAMA. 2000;284:1280-1283. 30. Glenny AM, Altman DG, Song F, et al. Indirect comparisons of competing interventions. Health Technol Assess. 2005;9:1-134; iii-iv.
657
RESULTS
CONCLUSION
To characterize population-based practice patterns, disease-specific and overall mortality, and cost associated with salvage cryotherapy (SCT) vs salvage radical prostatectomy (SRP). We retrospectively identified 440 men who failed primary radiation therapy and subsequently underwent SCT (n ¼ 341, 77.5%) or SRP (n ¼ 99, 22.5%) between 1992 and 2009 from Surveillance, Epidemiology, and End ResultseMedicare linked data. Propensity score analyses were used to compare overall and prostate cancer-specific mortality and associated Medicare expenditures for SRP vs SCT. Men undergoing SCT were more likely to be white (P <.001), less likely to be high school graduates (P ¼ .008), and experienced shorter median time from diagnosis to salvage therapy (44.1 vs 60.1, P <.001) and from primary radiotherapy to salvage therapy (38.7 vs 55.8 months, P <.001). In adjusted analyses, overall mortality was higher (21.6 vs 6.1 deaths/100 person years, P <.001) for SRP vs SCT. There was a trend for higher prostate cancer-specific death rates with SRP vs SCT (6.5 vs 1.4 deaths/100 person years, P ¼ .061). Medicare expenditures for SRP vs SCT were more than 2-fold higher ($19,543 vs $8,088, P <.001). SRP vs SCT is associated with higher overall mortality and greater health care expenditures. However, longer follow-up is needed to assess long-term functional outcomes and cancer control. UROLOGY 83: 653e657, 2014. 2014 Elsevier Inc.
P
rostate cancer is the most prevalent solid organ tumor among US men, with an estimated incidence of 241,740 cases and 28,170 prostate cancer-specific deaths in 2012.1 Owing to stage migration observed over the past 3 decades, most incident prostate cancers present as localized disease.2 Despite the popularity of traditional radiation therapies (external-beam and brachytherapy) as primary treatment options for localized disease,3 63% of men will experience biochemical recurrence within 10 years of radiotherapy,4
Financial Disclosure: Jim C. Hu receives salary support from a Department of Defense Prostate Cancer Physician Training Award (W81XWH-08-1-0283). The remaining authors declare that they have no relevant financial interests. Jim C. Hu 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. From the Division of Urology, Brigham and Women’s Hospital, Boston, MA; the Center for Surgery and Public Health, Brigham and Women’s Hospital, Boston, MA; the Department of Urology, Medical University of South Carolina, Charleston, SC; the Section of Urology, Ralph A. Johnson VA Medical Center, Charleston, SC; the Department of Radiation Oncology, Lank Center for Genitourinary Oncology, Dana Farber Cancer Institute, Boston, MA; the Division of Urologic Surgery, Center for Surgery and Public Health, Brigham and Women’s Hospital, Boston, MA; the Department of Public Health, Faculty of Medicine, University of Montreal, Montreal, Quebec, Canada; and the Department of Urology, David Geffen School of Medicine at University of California Los Angeles, Los Angeles, CA Reprint requests: Jim C. Hu, M.D., M.P.H., Department of Urology, David Geffen School of Medicine at University of California Los Angeles, 924 Westwood Blvd, STE 1000, Los Angeles, CA 90024. E-mail: [email protected] Submitted: August 9, 2013, accepted (with revisions): November 18, 2013
ª 2014 Elsevier Inc. All Rights Reserved
and it is estimated that 25%-32% will experience local failure.5 Salvage radical prostatectomy (SRP) and salvage cryotherapy (SCT) are performed with curative intent for clinically localized radiorecurrent prostate cancer, with 5-year progression-free survival approaching 60% for both.6,7 Although SRP historically has a longer follow-up with acceptable oncologic outcomes, it is accompanied by significant morbidity owing to radiation-induced fibrosis and tissue-plane obliteration.8 Conversely, SCT (particularly third-generation technology) has gained popularity, possibly because of improved technique and fewer complications.9 Although several systematic reviews have assessed SRP and SCT morbidity and survival,6,10 there is a dearth of directly comparative studies. Moreover, consensus regarding the optimal management of primary radiotherapy failures remains elusive.11 Through our population-based approach that provides insight beyond referral centers in which salvage therapy outcomes largely originate, we sought to compare the use, patterns of care, outcomes, and costs of SRP vs SCT.
MATERIALS AND METHODS Data Our study was approved by the University of California, Los Angeles Review Board; patient data were deidentified, and the 0090-4295/14/$36.00 http://dx.doi.org/10.1016/j.urology.2013.11.019
653
requirement for consent was waived. We used Surveillance, Epidemiology, and End Results (SEER)eMedicare data for analysis, which currently comprises a linkage of cancer registry data from 20 SEER regions with Medicare administrative data. The Medicare program provides benefits to 97% of Americans aged 65 years.12 We identified 345,803 men aged 65 years who were diagnosed with prostate cancer between 1992 and 2007, with Medicare follow-up through 2009. After excluding men not enrolled in both Medicare Part A and B or who were enrolled in a Medicare health maintenance organization because of unreliable claims submissions during the study period, 103,508 men treated with primary radiotherapy were subsequently identified. Of this group, 430 were found to have received SRP (n ¼ 99) or SCT (n ¼ 341). Patients were grouped into SRP or SCT cohorts based on International Classification of Diseases, Ninth Revision (ICD-9), Physicians Current Procedural Terminology Coding System, 4th edition, and Healthcare Common procedure Coding System codes (60.62, C2618, 55840, 55842, 55845, and 55873), with SRP defined as surgery after primary radiotherapy (external beam radiotherapy, brachytherapy, and/or intensity-modulated radiotherapy). We excluded perineal and minimally invasive radical prostatectomy, as these were uncommon in the salvage setting, totaling 25 procedures. We restricted our cohort to men with prostate cancer diagnosed as their only cancer.
Outcomes We examined overall and prostate cancer-specific mortality after SRP vs SCT. In addition, we characterized Medicare expenditures associated with each treatment.
Control Variables Age was obtained from the Medicare file; race, census tract measures of median household income and high school education, region, population density (urban vs rural), and marital status were obtained from SEER registry data. Comorbidity was assessed using the Klabunde modification of the Charlson index during the year before surgery.13 ICD-9 codes were used to identify disease categories, whereas Physicians Current Procedural Terminology Coding System, 4th edition and Healthcare Common Procedure Coding System code sets were used to identify medical, surgical, and diagnostic services. To increase specificity, only imaging studies designated with a corresponding ICD-9 code for prostate cancer were included. Adjuvant androgen deprivation therapy (ADT) was defined as ADT use anytime between 6 months before and after primary therapy. “Additional ADT” was defined as ADT use at 2 years after SCT and use anytime after SRP.
experiment. Propensity score adjustment was performed using a logistic regression model to calculate the probability of undergoing SRP vs SCT on the basis of described covariates and then weighting the data on the basis of the inverse propensity of being in either of the treatment groups.17 After adjustment, covariate balance was assessed. All tests were considered statistically significant at a ¼ 0.05. All analyses were performed with SAS version 9.2 (SAS Institute, Cary, NC). Because of confidentiality, values <11 may not be reported directly or in a derivable way. Therefore, for outcomes with <11 patients, we present odds ratios.
RESULTS Among men experiencing radiotherapy failure, 341 (77.5%) underwent SCT and 99 (22.5%) SRP. Median follow-up for SRP vs SCT was 30 (interquartile range 1844.4) vs 15 (interquartile range 4.8-33.6) months after salvage therapy. Although men undergoing SRP vs SCT experienced longer median time from primary to salvage therapy (55.8 vs 38.7 months, P <.001) and from diagnosis to salvage therapy (60.1 vs 44.1 months, P <.001), they experienced similar median time from prostate cancer diagnosis to primary therapy (3.0 months for both, P ¼ .132). Men receiving SCT vs SRP (Table 1) were more likely to be white (P <.001) and to live in areas with <85% high school graduation rates (P ¼ .008). In addition, men undergoing SCT vs SRP were more likely to have received previous ADT (50.4% vs 45.5%, P ¼ .001) and primary brachytherapy (43.7% vs 24.2%, P ¼ .001). Propensity adjusted analyses is presented in Table 2. Overall survival was reduced (21.57 vs 6.14 deaths/100 person years, P <.001) for SRP vs SCT. Similarly, there was a trend for higher disease-specific mortality for SRP vs SCT (6.54 vs 1.37 deaths/100 person years, P ¼ .061). Health care expenditures in the 3 months before salvage therapy were higher for SRP vs SCT (median $8,416 vs $2,363, P <.001). Similarly, SRP costs were 2fold higher than SCT (median $19,543 vs $8,088, P <.001). SRP vs SCT patients were more likely to require inpatient care in the 6 months surrounding surgery (58.6% vs 32.0%), at a median cost of $23,321 vs $10,497 (P <.001).
COMMENT Expenditures To best attribute the costs associated with each surgical setting, we assessed Medicare payments within 3 months of salvage therapy, which represents the traditional global payment period.14
Statistical Analysis Unadjusted analysis using the Pearson chi-square statistic was performed to compare demographic and tumor characteristics for SRP vs SCT.15 Adjusted analyses were performed with weighted propensity scoring.16 Propensity score methods control for all observed confounding factors that might influence cohort assignment and outcome using a single composite measure, balancing patient characteristics as would occur in a randomized 654
The optimal management of radiorecurrent prostate cancer remains controversial, particularly for younger men or those with life expectancy >10 years.11 Current salvage treatment options include ADT, brachytherapy, cryoablative therapy, high-intensity focused ultrasound, and radical prostatectomy. Although SRP has the longest documented follow-up with acceptable oncologic outcomes, its use has been limited by greater perioperative morbidity, including rectal injury, urethrocutaneous fistula, lymphedema, urinary incontinence, and anastomotic stricture.18,19 The use of SCT has increased in recent years owing to comparable oncologic outcomes and superior complication rates.5 However, in the UROLOGY 83 (3), 2014
Table 1. Demographics of study population Independent Variables n (%) Year of diagnosis 2002-2005 2006-2009 Age at diagnosis 65-69 70-74 75þ Race Nonwhite White Median income <$35,000 $35,000-$44,000 $45,000-$59,999 $60,000 Education (%) <85 85 Population density Urban Region Northeast South Midwest West Marital status Not married Married Clinical stage T1 T2-T3 Grade Well differentiated Poorly/ undifferentiated ADT before salvage therapy Primary therapy Brachytherapy External beam Combination Median primary to salvage therapy time interval (mo)
Salvage Radical Salvage Cryotherapy Prostatectomy n ¼ 99 P Value n ¼ 341 110 (32.26) 231 (67.74)
30 (30.30) 69 (69.70)
.713
144 (42.23) 112 (32.84) 85 (24.93)
40 (40.40) 40 (40.40) 19 (19.19)
.302
82 (24.05) 259 (75.95)
* *
109 92 73 66
(26.26) (20.20) (25.25) (28.28)
.135
154 (45.29) 186 (54.71)
30 (30.30) 69 (69.70)
.008
308 (90.32)
*
.901
76 81 52 132
(32.06) (27.06) (21.47) (19.41)
(22.29) (23.75) (15.25) (38.71)
26 20 25 28
<.001
21 16 23 39
Events per 100 Person Years
Salvage Radical Salvage Cryotherapy Prostatectomy n ¼ 99 P Value n ¼ 341
Unadjusted Overall mortality Prostate cancer-specific mortality Additional ADT Adjusted Overall mortality Prostate cancer-specific mortality Additional ADT
5.76 1.33
22.54 6.17
<.001 .058
6.22
9.49
.162
6.14 1.37
21.57 6.54
<.001 .061
6.3
8.79
.273
Abbreviation as in Table 1.
(21.21) (16.16) (23.23) (39.39)
.177
54 (16.93) 265 (83.07)
16 (17.39) 76 (82.61)
.917
140 (42.55) 189 (57.45)
43 (44.79) 53 (55.21)
.697
207 (61.79)
68 (70.10)
.134
128 (38.21)
29 (29.90)
172 (50.44)
45 (45.45)
.001
149 (43.70) 146 (42.82) 46 (13.49) 38.7
24 (24.24) 51 (51.52) 24 (24.24) 55.8
.001 <.001
ADT, androgen deprivation therapy. * Race and population density for salvage radical prostatectomy are not shown, as there are <11 patients within a category.
absence of prospective perioperative and survival data,7 direct comparison of SRP and SCT is difficult. To our knowledge, our study is the first to directly compare the mortality outcomes between salvage therapies for locally recurrent prostate cancer after initial radiation therapy failure, and our study has several important findings. First, we observed a trend in overall and prostate cancer-specific death that was more common among SRP vs SCT men. Although previous investigators have focused on traditional markers of treatment failure such as UROLOGY 83 (3), 2014
Table 2. Unadjusted and propensity score adjusted outcomes of salvage cryotherapy vs radical prostatectomy
postsalvage therapy biopsy positivity and prostate-specific antigen (PSA) levels, we examine overall and cancerspecific mortality. Our findings differ from those reported in a retrospective single-center study by Pisters et al20, who demonstrated that SRP yielded better overall survival but comparable disease-specific survival when compared with SCT. However, tumor grade was significantly higher in their SCT cohort, potentially confounding their results given the association between high tumor grade and PSA failure after SCT.21 Moreover, their analysis was limited to men with a presalvage PSA level of <10 ng/mL and history of previous radiation therapy alone, making direct comparison with our study difficult given the absence of presalvage therapy PSA values. To the best of our knowledge, previous studies did not use the Phoenix definition of PSA nadir plus 2 ng/mL. This might account for the significantly higher recurrence rates reported by Pisters et al relative to a recent retrospective analysis of SCT conducted by Williams et al.22 This discrepancy demonstrates the importance of establishing standardized definitions of treatment failure. Second, salvage therapy was rarely used among men who failed primary radiotherapy, with just 0.4% subsequently undergoing SRP or SCT. This finding is lower than salvage therapy rates noted in previous populationbased studies.4 As noted by Jang et al23, physician perceptions regarding optimal primary therapy for localized prostate cancer vary widely across specialties and geographic regions. It can be inferred that a similar level of equivocality surrounds salvage therapies and might contribute to lower rates of utilization. Furthermore, Jang et al observed increasing frequency of conservative management with age among men seen exclusively by urologists. The SEER cohort consists of men older than 65 years, and the infrequency of salvage therapy seen in our study might reflect higher rates of conservative management in this population. Finally, Mouraviev et al24 note in a recent review that life expectancy >10 years might confer a greater likelihood of favorable biochemical disease-free survival in the setting of salvage 655
therapy, and it is conceivable that our low salvage therapy utilization rate is the result of low 10-year survival rate projections within the SEER population. Third, men undergoing SRP vs SCT experienced longer median time from primary to salvage therapy and from diagnosis to salvage therapy, although time from diagnosis to primary therapy was similar to SCT men. These findings might represent greater reluctance to perform SRP, which is more fraught with complications and technical challenges as compared with SCT.5 Previous reports have postulated that serum PSA level of <10 ng/mL, Gleason score of 8, and clinical stage T1c or T2 before salvage therapy are predictors of improved biochemical recurrence-free survival after salvage therapy.9 These findings suggest that salvage therapy is likely to be more effective in the setting of less advanced radiorecurrent disease. Although we control for differences in primary to salvage therapy time intervals experienced by our 2 treatment cohorts, our adjusted analysis might not account for potential differences in disease severity secondary to longer primary to salvage therapy intervals, as assessed by possible PSA value differences between our 2 treatment groups. In addition, a longer primary to salvage interval might contribute to greater overall morbidity and therefore have a deleterious effect on lifestyle, both of which were not controlled for in our analysis. Finally, health care expenditures for SRP vs SCT were more than 2-fold higher, with a difference of approximately $11,000. This finding is interesting given higher salvage ADT use among SCT men, as previous studies have shown significant added cost associated with salvage ADT.25 Differences in cost between SCT and SRP might be partially explained by the fact that SRP patients were more likely to require inpatient care in the 6 months surrounding salvage therapy. Although this figure might not capture all payments associated with complications beyond 3 months, it includes the costs of postoperative clinic visits, emergency room visits, readmission, and additional surgical or radiologic procedures. Nonetheless, selective referral to experienced SRP surgeons might reduce the significant morbidity and costs observed in our study.26 Our findings must be interpreted in the context of the study design. First, analyses were restricted to Medicare beneficiaries older than 65 years residing in SEER regions, and our findings might not be applicable to younger men. Similarly, there might be selection bias for younger men to undergo SRP because of less comorbidities and postoperative incontinence relative to older men.27 Because our study of Medicare beneficiaries omits these men, the recorded overall and disease-specific mortality might consequently reflect the inherently more difficult nature of SRP in older individuals. Neither clinical staging data nor biopsy Gleason scores after failure of primary radiotherapy (before salvage therapy) were available for analysis. Moreover, recurrence PSA values were unavailable before salvage therapy. Previous studies have shown that 656
lower pre-SRP PSA levels confer improved oncologic outcomes.28 Similarly, higher SCT failure rates are experienced by men with postradiation PSA levels >10 ng/mL or PSA doubling times of 16 months.5 Taken together in the absence of presalvage therapy PSA values, it is impossible to know whether the unfavorable subgroups noted by Chade, Pisters, and Finley are disproportionately over or underrepresented in either of our study cohorts. Similarly, because our regression does not account for the temporality of salvage ADT, our results are subject to the confounding influence of hormonal therapy.29 In addition, in the absence of data from validated quality of life instruments, we were unable to compare functional outcomes after SRP vs SCT. Finally, as with any adjusted analysis, propensity score methods do not control for unmeasured confounders, such as lifestyle factors that might effect overall survival and possess other limitations.30
CONCLUSION Although men undergoing SRP vs SCT are less likely to require salvage ADT, they incur more total health care expenditures and experience a higher rate of overall death, even after adjusting for differences in follow-up and time between primary to salvage therapy. Patients and providers alike should consider these populationbased findings when discussing salvage options for radiorecurrent prostate cancer. References 1. Siegel R, Naishadham D, Jemal A. Cancer statistics, 2012. CA: A Cancer J Clin. 2012;62:10-29. 2. Galper SL, Chen MH, Catalona WJ, et al. Evidence to support a continued stage migration and decrease in prostate cancer specific mortality. J Urol. 2006;175:907-912. 3. Cooperberg MR, Broering JM, Carroll PR. Time trends and local variation in primary treatment of localized prostate cancer. J Clin Oncol. 2010;28:1117-1123. 4. Agarwal PK, Sadetsky N, Konety BR, et al. Cancer of the Prostate Strategic Urological Research Endeavor (CaPSURE) treatment failure after primary and salvage therapy for prostate cancer: likelihood, patterns of care, and outcomes. Cancer. 2008;112: 307-314. 5. Finley DS, Pouliot F, Miller DC, Belldegrun AS. Primary and salvage cryotherapy for prostate cancer. Urol Clin North Am. 2010; 37:67-82. 6. Finley DS, Belldegrun AS. Salvage cryotherapy for radiationrecurrent prostate cancer: outcomes and complications. Curr Urol Rep. 2011;12:209-215. 7. Nguyen PL, D’Amico AV, Lee AK, et al. Patient selection, cancer control, and complications after salvage local therapy for postradiation prostate-specific antigen failure. Cancer. 2007;110:14171428. 8. Cox JM, Busby JE. Salvage therapy for prostate cancer recurrence after radiation therapy. Curr Urol Rep. 2009;10:199-205. 9. Kimura M, Mouraviev V, Tsivian M, et al. Current salvage methods for recurrent prostate cancer after failure of primary radiotherapy. BJU Int. 2010;105:191-201. 10. Heidenreich A, Richter S, Thuer D, et al. Prognostic parameters, complications, and oncologic and functional outcome of salvage radical prostatectomy for locally recurrent prostate cancer after 21stcentury radiotherapy. Eur Urol. 2010;57:437-443.
UROLOGY 83 (3), 2014
11. Stephen J. Radiorecurrent prostate Cancer: an emerging and largely mismanaged epidemic. Eur Urol. 2011;60:411-412. 12. Warren JL, Klabunde CN, Schrag D, et al. Overview of the SEERMedicare data: content, research applications, and generalizability to the United States elderly population. Med Care. 2002;40(8 Suppl):IV-3-IV18. 13. Klabunde CN, Potosky AL, Legler JM, et al. Development of a comorbidity index using physician claims data. J Clin Epidemiol. 2000;53:1258-1267. 14. Prasad SM, Gu X, Lavelle R, et al. Comparative effectiveness of perineal versus retropubic and minimally invasive radical prostatectomy. J Urol. 2011;185:111-115. 15. Rao JNK, Scott AJ. The analysis of categorical-data from complex sample-surveys e Chi-squared tests for goodness of fit and independence in 2-way tables. JASA. 1981;76:221-230. 16. Rubin DB. Estimating causal effects from large data sets using propensity scores. Ann Intern Med. 1997;127:757-763. 17. Robins JM, Hernan MA, Brumback B. Marginal structural models and causal inference in epidemiology. Epidemiology. 2000;11: 550-560. 18. Rainwater LM, Zincke H. Radical prostatectomy after radiation therapy for cancer of the prostate: feasibility and prognosis. J Urol. 1998;140:1455-1459. 19. Vaidya A, Soloway MS. Salvage radical prostatectomy for radiorecurrent prostate cancer: morbidity revisited. J Urol. 2000;164: 1998-2001. 20. Pisters LL, Leibovici D, Blute M, et al. Locally recurrent prostate cancer after initial radiation therapy: a comparison of salvage radical prostatectomy versus cryotherapy. J Urol. 2009;182: 517-525.
UROLOGY 83 (3), 2014
21. Pisters LL, Perrotte P, Scott SM, et al. Patient selection for salvage cryotherapy for locally recurrent prostate cancer after radiation therapy. J Clin Oncol. 1999;17:2514-2520. 22. Williams AK, Martınez CH, Lu C, et al. Disease-free survival following salvage cryotherapy for biopsy-proven radio-recurrent prostate cancer. Eur Urol. 2011;60:405-410. 23. Jang TL, Bekelman JE, Liu Y, et al. Physician visits prior to treatment for clinically localized prostate cancer. Arch Intern Med. 2010; 170:440-450. 24. Mouraviev V, Spiess PE, Jones JS. Salvage cryoablation for locally recurrent prostate cancer following primary radiotherapy. Eur Urol. 2012;61:1204-1211. 25. Krupski TL, Foley KA, Baser O, et al. Health care cost associated with prostate cancer, androgen deprivation therapy and bone complications. J Urol. 2007;178:1423-1428. 26. Williams SB, Gu X, Lipsitz SR, et al. Utilization and expense of adjuvant cancer therapies following radical prostatectomy. Cancer. 2011;117:4846-4854. 27. Anderson CB, Kaufman MR, Dietrich MS, et al. Recovery of urinary function after radical prostatectomy: identification of trajectory cluster groups. J Urol. 2012;187:1346-1351. 28. Chade DC, Shariat SF, Cronin AM, et al. Salvage radical prostatectomy for radiation-recurrent prostate cancer: a multi-institutional collaboration. Eur Urol. 2011;60:205-210. 29. D’Amico AV, Schultz D, Loffredo M, et al. Biochemical outcome following external beam radiation therapy with or without androgen suppression therapy for clinically localized prostate cancer. JAMA. 2000;284:1280-1283. 30. Glenny AM, Altman DG, Song F, et al. Indirect comparisons of competing interventions. Health Technol Assess. 2005;9:1-134; iii-iv.
657