Strahlenther Onkol (2015) 191:801–809 DOI 10.1007/s00066-015-0872-9
O r i g i n a l A rt i c l e
Elective pelvic versus prostate bed-only salvage radiotherapy following radical prostatectomy A propensity score-matched analysis Changhoon Song · Hyun-Cheol Kang · Jae-Sung Kim · Keun-Yong Eom · In Ah Kim · Jin-Beom Chung · Sung Kyu Hong · Seok-Soo Byun · Sang Eun Lee
Received: 16 August 2014 / Accepted: 23 June 2015 / Published online: 10 July 2015 © Springer-Verlag Berlin Heidelberg 2015
Abstract Purpose To compare the impact of elective whole pelvic radiotherapy (WPRT) versus prostate bed-only radiotherapy (PBRT) on biochemical relapse-free survival (bRFS) in prostate cancer patients treated with salvage radiotherapy following radical prostatectomy (RP). Patients and methods In our database, 163 lymph nodenegative prostate cancer patients who had undergone salvage radiotherapy (SRT) for biochemical relapse after RP between September 2004 and April 2012 were identified. PBRT was administered to 134 patients (the PBRT group), while the remaining 29 patients (the WPRT group) received WPRT.
Changhoon Song and Hyun-Cheol Kang contributed equally to this work. J.-S. Kim, M.D., Ph.D. () · C. Song, M.D. · H.-C. Kang, M.D. · K.-Y. Eom, M.D. · I. A. Kim, M.D., PhD. · J.-B. Chung, Ph.D. Department of Radiation Oncology, Seoul National University College of Medicine, Seoul National University Bundang Hospital, 82, Gumi-ro 173 beon-gil, Bundang-gu, Seongnam-si, 463–707, Seongnam, Korea/Republic of Korea e-mail: [email protected] S. K. Hong, M.D., Ph.D. · S.-S. Byun, M.D., Ph.D. · S. E. Lee, M.D., Ph.D. Department of Urology, Seoul National University College of Medicine, Seoul National University Bundang Hospital, Seongnam, Korea H.-C. Kang, M.D. Department of Radiation Oncology, Dongnam Institute of Radiological and Medical Sciences, Busan, Korea
Results Median follow-up was 57 months (range 18– 122 months). In the propensity score-matched cohort, the 4-year bRFS of the WPRT group was significantly higher compared to the PBRT group (63.1 vs. 43.4 %, p = 0.034). Subgroup analysis showed that the bRFS of patients who had two or more risk factors (seminal vesicle invasion, Roach score for lymph node invasion ≥ 45 %, and number of harvested lymph nodes ≤ 5) and were treated with WPRT was significantly improved compared to those who received PBRT (hazard ratio, HR 0.33; 95 % confidence interval, CI 0.13–0.83; p = 0.018). Conclusion Elective WPRT for SRT may improve bRFS in patients with unfavorable risk factors. These results need to be confirmed by a prospective randomized trial. Keywords Prostate cancer · Survival · Risk factors · Lymph nodes · Androgen deprivation therapy Gezielte Becken- versus Salvage-Bestrahlung des Prostatabetts nach radikaler Prostatektomie Eine ergebnisorientierte Analyse Zusammenfassung Ziel Vergleich der Auswirkungen der gezielten Beckenstrahlentherapie (WPRT) gegenüber der Prostatastrahlentherapie (PBRT) hinsichtlich des biochemischen rezidivfreien Überlebens (bRFS) bei Prostatakarzinompatienten, die nach radikaler Prostatektomie (RP) mit einer SalvageStrahlentherapie (SRT) behandelt wurden. Patienten und Methoden Aus unserer Datenbank wurden 163 lymphknotennegative Patienten mit Prostatakrebs identifiziert, die sich nach RP zwischen September 2004 und April 2012 wegen einem biochemischen Rückfall einer ret-
13
802
tenden SRT unterzogen hatten. PBRT wurde 134 Patienten verabreicht (PBRT-Gruppe), während die übrigen 29 Patienten eine WPRT erhielten (WPRT-Gruppe). Ergebnisse Die mediane Nachbeobachtungszeit betrug 57 Monate (Spanne: 18–122 Monate). In der auf den Propensity-Score abgestimmten Kohorte war das 4-JahresbRFS der WPRT-Gruppe im Vergleich zur PBRT-Gruppe signifikant höher (63,1 vs. 43,4 %; p = 0,034). Eine Untergruppenanalyse zeigte, dass sich das bRFS der Patienten, die zwei oder mehr Risikofaktoren hatten (Samenblasenbefall, Roach-Score für Lymphknotenbefall ≥ 45 % und Zahl der entfernten Lymphknoten ≤ 5) und mit WPRT behandelt wurden, im Vergleich zu diejenigen, die PBRT erhalten hatten, deutlich verbessert (Hazard-Ratio 0,33; 95 %-Konfidenzintervall 0,13–0,83; p = 0,018). Fazit Eine gezielte WPRT-Strahlentherapie kann das bRFS bei Patienten mit ungünstigen Risikofaktorenverbessern. Diese Ergebnisse müssen in einer prospektiven randomisierten Studie bestätigt werden. Schlüsselwörter Prostatakrebs · Überleben · Risikofaktoren · Lymphknoten · Androgendeprivationstherapie Approximately 30 % of patients undergoing radical prostatectomy (RP) experience subsequent biochemical relapse [1–4]. For these patients, salvage radiotherapy (SRT) is commonly performed. SRT has been shown to improve biochemical control and survival of these patients in several retrospective studies [5–9]. However, the appropriate radiation target volume, i.e., whether to use whole pelvic radiotherapy (WPRT) or prostate bed-only radiotherapy (PBRT), has not been well established in this setting. Theoretically, elective WPRT improves outcomes of lymph node-negative patients by sterilizing the locoregional lymph nodes, which may harbor occult disease. Imaging studies have poor sensitivities for detecting lymph node metastasis in prostate cancer, even with multiparametric magnetic resonance imaging (MRI) and positronemission tomography/computed tomography (PET/CT) with 11C-labelled choline derivatives [10–13]. Furthermore, pelvic lymph node dissection (PLND) during RP—which is the most accurate method for determining nodal metastasis and also a potential treatment for nodal metastasis—has not been frequently performed since 2000, even in patients with unfavorable features [14, 15]. Therefore, the role of elective WPRT in the salvage setting might need to be reevaluated within the contemporary surgical trend. To our knowledge, only three retrospective studies have compared WPRT and PBRT in the salvage setting [16–18]. However, due to the significant imbalance of clinicopathologic prognostic factors between the two groups, the role
13
C. Song et al.
of WPRT was not evaluated appropriately. Needless to say, the indication for WPRT as SRT has not been accurately defined. Bearing in mind these considerations, we performed a propensity score-matched analysis to adjust for clinicopathologic prognostic factors that are associated with the use of WPRT. We have compared biochemical relapsefree survival (bRFS) and adverse events between WPRT and PBRT among lymph node-negative prostate cancer patients treated with SRT between 2004 and 2012, during which time contemporary treatment strategies were in use. We have endeavored to identify subpopulations of patients that may benefit most from WPRT. Patients and methods Patients After approval by the institutional review board, a retrospective review was performed by collecting data from between September 2004 and April 2012 from our institution's database. A total of 187 consecutive prostate cancer patients underwent SRT for post-RP biochemical relapse. For the purpose of this study, 16 patients with pathologic lymph node metastases at the time of RP were excluded. An additional 8 patients were excluded because of SRT with palliative intent (n = 7) and loss to follow-up (n = 1), leaving 163 patients as the final cohort included in the present study. Of these, 29 patients had received WPRT (the WPRT group) and 134 patients had received PBRT (the PBRT group). Salvage treatment was indicated in cases of a persistently elevated serum prostatespecific antigen (PSA) after RP (PSA ≥ 0.1 ng/mL) or an elevation of serum PSA levels ≥ 0.1 ng/mL after achieving 20 ng/mL, a pathologic Gleason score of ≥ 8, or extracapsular extension (pT3), which has been reported to confer a > 20 % likelihood of pelvic lymph node involvement from Partin tables [17, 23]. The other subgroup delineation stratifies patients according to the pre-SRT PSA level ≥ 0.4 ng/mL [16]. Lastly, the impact of WPRT was evaluated according to several risk factors. These risk factors were preoperative PSA level, pathologic Gleason scores, Roach score for lymph node invasion, extracapsular extension, seminal vesicle invasion,
13
804
surgical margin statuses, PSA doubling time before SRT, interval from RP to biochemical relapse, number of harvested lymph nodes, and the use of ADT. The statistical significance of differences was assessed using the chi-squared or Fisher’s exact test for categorical data, and the t-test or Mann–Whitney U test, for continuous data. The bRFS curves were generated using the Kaplan– Meier method, and a univariate survival comparison was performed using the log-rank test. Hazard ratios (HRs), 95 % confidence intervals (CIs) and p-values with an indicator for WPRT were estimated using proportional hazards regression models. All analyses were carried out with statistical program R (R Foundation for Statistical Computing, Vienna, Austria; http://www.r-project.org/). All reported p-values are two-sided, with p 20 ng/mL (51.7 vs. 35.1 %), a pathologic Gleason score ≥ 8 (69.0 vs. 28.4 %), extracapsular extension (96.6 vs. 68.7 %), seminal vesicle invasion (72.4 vs. 30.6 %), post-RP persistently elevated PSA (89.7 vs. 41.0 %), pre-SRT PSA ≥ 0.4 ng/mL (82.8 vs. 53.0 %), and use of ADT (89.7 vs. 60.4 %). The time to biochemical relapse after RP was significantly shorter in the WPRT group (median 5.4 months vs. 9.5 months). The number of harvested lymph nodes and PSA doubling time were lower in the WPRT group than in the PBRT group, with a trend toward significance (p = 0.188 and p = 0.057, respectively). For the entire cohort, the 4-year bRFS rates were 63.1 % (95 % CI 40.2–79.3) in the WPRT group and 56.4 % (95 % CI 46.3–65.2) in the PBRT group (p = 0.257; Fig. 2a). Propensity score-matched cohort Propensity score matching resulted in 29 patients in WPRT the group and 58 patients in the PBRT group, i.e., 87 patients
13
C. Song et al.
in total. Except a pathologic Gleason score of ≥ 8, no characteristics of patient, tumor, or treatment were significantly different between groups of the matched cohort (Table 1), indicating that the matching procedure worked well. Use of WPRT resulted in improved 4-year bRFS rates compared to PBRT (63.1 %, 95 % CI 40.2–79.3 vs. 43.4 %, 95 % CI 28.2–57.6; p = 0.034; Fig. 2b). Subgroup analyses Patients stratified by “a priori risk group” When all 163 patients were stratified by “a priori risk group”, 29 patients were considered as low risk and 134 patients as high risk. WPRT conferred no benefit on either group. For the high-risk group, the 4-year bRFS rates of patients who received WPRT and PBRT were 61.4 (95 % CI 37.9–78.3) and 56.3 % (95 % CI 45.2–66.1), respectively (p = 0.339; Fig. 3). Patients stratified by pre-SRT PSA level ≥ 0.4 ng/mL When patients were grouped by pre-SRT PSA level, 69 patients had pre-SRT PSA level
O r i g i n a l A rt i c l e
Elective pelvic versus prostate bed-only salvage radiotherapy following radical prostatectomy A propensity score-matched analysis Changhoon Song · Hyun-Cheol Kang · Jae-Sung Kim · Keun-Yong Eom · In Ah Kim · Jin-Beom Chung · Sung Kyu Hong · Seok-Soo Byun · Sang Eun Lee
Received: 16 August 2014 / Accepted: 23 June 2015 / Published online: 10 July 2015 © Springer-Verlag Berlin Heidelberg 2015
Abstract Purpose To compare the impact of elective whole pelvic radiotherapy (WPRT) versus prostate bed-only radiotherapy (PBRT) on biochemical relapse-free survival (bRFS) in prostate cancer patients treated with salvage radiotherapy following radical prostatectomy (RP). Patients and methods In our database, 163 lymph nodenegative prostate cancer patients who had undergone salvage radiotherapy (SRT) for biochemical relapse after RP between September 2004 and April 2012 were identified. PBRT was administered to 134 patients (the PBRT group), while the remaining 29 patients (the WPRT group) received WPRT.
Changhoon Song and Hyun-Cheol Kang contributed equally to this work. J.-S. Kim, M.D., Ph.D. () · C. Song, M.D. · H.-C. Kang, M.D. · K.-Y. Eom, M.D. · I. A. Kim, M.D., PhD. · J.-B. Chung, Ph.D. Department of Radiation Oncology, Seoul National University College of Medicine, Seoul National University Bundang Hospital, 82, Gumi-ro 173 beon-gil, Bundang-gu, Seongnam-si, 463–707, Seongnam, Korea/Republic of Korea e-mail: [email protected] S. K. Hong, M.D., Ph.D. · S.-S. Byun, M.D., Ph.D. · S. E. Lee, M.D., Ph.D. Department of Urology, Seoul National University College of Medicine, Seoul National University Bundang Hospital, Seongnam, Korea H.-C. Kang, M.D. Department of Radiation Oncology, Dongnam Institute of Radiological and Medical Sciences, Busan, Korea
Results Median follow-up was 57 months (range 18– 122 months). In the propensity score-matched cohort, the 4-year bRFS of the WPRT group was significantly higher compared to the PBRT group (63.1 vs. 43.4 %, p = 0.034). Subgroup analysis showed that the bRFS of patients who had two or more risk factors (seminal vesicle invasion, Roach score for lymph node invasion ≥ 45 %, and number of harvested lymph nodes ≤ 5) and were treated with WPRT was significantly improved compared to those who received PBRT (hazard ratio, HR 0.33; 95 % confidence interval, CI 0.13–0.83; p = 0.018). Conclusion Elective WPRT for SRT may improve bRFS in patients with unfavorable risk factors. These results need to be confirmed by a prospective randomized trial. Keywords Prostate cancer · Survival · Risk factors · Lymph nodes · Androgen deprivation therapy Gezielte Becken- versus Salvage-Bestrahlung des Prostatabetts nach radikaler Prostatektomie Eine ergebnisorientierte Analyse Zusammenfassung Ziel Vergleich der Auswirkungen der gezielten Beckenstrahlentherapie (WPRT) gegenüber der Prostatastrahlentherapie (PBRT) hinsichtlich des biochemischen rezidivfreien Überlebens (bRFS) bei Prostatakarzinompatienten, die nach radikaler Prostatektomie (RP) mit einer SalvageStrahlentherapie (SRT) behandelt wurden. Patienten und Methoden Aus unserer Datenbank wurden 163 lymphknotennegative Patienten mit Prostatakrebs identifiziert, die sich nach RP zwischen September 2004 und April 2012 wegen einem biochemischen Rückfall einer ret-
13
802
tenden SRT unterzogen hatten. PBRT wurde 134 Patienten verabreicht (PBRT-Gruppe), während die übrigen 29 Patienten eine WPRT erhielten (WPRT-Gruppe). Ergebnisse Die mediane Nachbeobachtungszeit betrug 57 Monate (Spanne: 18–122 Monate). In der auf den Propensity-Score abgestimmten Kohorte war das 4-JahresbRFS der WPRT-Gruppe im Vergleich zur PBRT-Gruppe signifikant höher (63,1 vs. 43,4 %; p = 0,034). Eine Untergruppenanalyse zeigte, dass sich das bRFS der Patienten, die zwei oder mehr Risikofaktoren hatten (Samenblasenbefall, Roach-Score für Lymphknotenbefall ≥ 45 % und Zahl der entfernten Lymphknoten ≤ 5) und mit WPRT behandelt wurden, im Vergleich zu diejenigen, die PBRT erhalten hatten, deutlich verbessert (Hazard-Ratio 0,33; 95 %-Konfidenzintervall 0,13–0,83; p = 0,018). Fazit Eine gezielte WPRT-Strahlentherapie kann das bRFS bei Patienten mit ungünstigen Risikofaktorenverbessern. Diese Ergebnisse müssen in einer prospektiven randomisierten Studie bestätigt werden. Schlüsselwörter Prostatakrebs · Überleben · Risikofaktoren · Lymphknoten · Androgendeprivationstherapie Approximately 30 % of patients undergoing radical prostatectomy (RP) experience subsequent biochemical relapse [1–4]. For these patients, salvage radiotherapy (SRT) is commonly performed. SRT has been shown to improve biochemical control and survival of these patients in several retrospective studies [5–9]. However, the appropriate radiation target volume, i.e., whether to use whole pelvic radiotherapy (WPRT) or prostate bed-only radiotherapy (PBRT), has not been well established in this setting. Theoretically, elective WPRT improves outcomes of lymph node-negative patients by sterilizing the locoregional lymph nodes, which may harbor occult disease. Imaging studies have poor sensitivities for detecting lymph node metastasis in prostate cancer, even with multiparametric magnetic resonance imaging (MRI) and positronemission tomography/computed tomography (PET/CT) with 11C-labelled choline derivatives [10–13]. Furthermore, pelvic lymph node dissection (PLND) during RP—which is the most accurate method for determining nodal metastasis and also a potential treatment for nodal metastasis—has not been frequently performed since 2000, even in patients with unfavorable features [14, 15]. Therefore, the role of elective WPRT in the salvage setting might need to be reevaluated within the contemporary surgical trend. To our knowledge, only three retrospective studies have compared WPRT and PBRT in the salvage setting [16–18]. However, due to the significant imbalance of clinicopathologic prognostic factors between the two groups, the role
13
C. Song et al.
of WPRT was not evaluated appropriately. Needless to say, the indication for WPRT as SRT has not been accurately defined. Bearing in mind these considerations, we performed a propensity score-matched analysis to adjust for clinicopathologic prognostic factors that are associated with the use of WPRT. We have compared biochemical relapsefree survival (bRFS) and adverse events between WPRT and PBRT among lymph node-negative prostate cancer patients treated with SRT between 2004 and 2012, during which time contemporary treatment strategies were in use. We have endeavored to identify subpopulations of patients that may benefit most from WPRT. Patients and methods Patients After approval by the institutional review board, a retrospective review was performed by collecting data from between September 2004 and April 2012 from our institution's database. A total of 187 consecutive prostate cancer patients underwent SRT for post-RP biochemical relapse. For the purpose of this study, 16 patients with pathologic lymph node metastases at the time of RP were excluded. An additional 8 patients were excluded because of SRT with palliative intent (n = 7) and loss to follow-up (n = 1), leaving 163 patients as the final cohort included in the present study. Of these, 29 patients had received WPRT (the WPRT group) and 134 patients had received PBRT (the PBRT group). Salvage treatment was indicated in cases of a persistently elevated serum prostatespecific antigen (PSA) after RP (PSA ≥ 0.1 ng/mL) or an elevation of serum PSA levels ≥ 0.1 ng/mL after achieving 20 ng/mL, a pathologic Gleason score of ≥ 8, or extracapsular extension (pT3), which has been reported to confer a > 20 % likelihood of pelvic lymph node involvement from Partin tables [17, 23]. The other subgroup delineation stratifies patients according to the pre-SRT PSA level ≥ 0.4 ng/mL [16]. Lastly, the impact of WPRT was evaluated according to several risk factors. These risk factors were preoperative PSA level, pathologic Gleason scores, Roach score for lymph node invasion, extracapsular extension, seminal vesicle invasion,
13
804
surgical margin statuses, PSA doubling time before SRT, interval from RP to biochemical relapse, number of harvested lymph nodes, and the use of ADT. The statistical significance of differences was assessed using the chi-squared or Fisher’s exact test for categorical data, and the t-test or Mann–Whitney U test, for continuous data. The bRFS curves were generated using the Kaplan– Meier method, and a univariate survival comparison was performed using the log-rank test. Hazard ratios (HRs), 95 % confidence intervals (CIs) and p-values with an indicator for WPRT were estimated using proportional hazards regression models. All analyses were carried out with statistical program R (R Foundation for Statistical Computing, Vienna, Austria; http://www.r-project.org/). All reported p-values are two-sided, with p 20 ng/mL (51.7 vs. 35.1 %), a pathologic Gleason score ≥ 8 (69.0 vs. 28.4 %), extracapsular extension (96.6 vs. 68.7 %), seminal vesicle invasion (72.4 vs. 30.6 %), post-RP persistently elevated PSA (89.7 vs. 41.0 %), pre-SRT PSA ≥ 0.4 ng/mL (82.8 vs. 53.0 %), and use of ADT (89.7 vs. 60.4 %). The time to biochemical relapse after RP was significantly shorter in the WPRT group (median 5.4 months vs. 9.5 months). The number of harvested lymph nodes and PSA doubling time were lower in the WPRT group than in the PBRT group, with a trend toward significance (p = 0.188 and p = 0.057, respectively). For the entire cohort, the 4-year bRFS rates were 63.1 % (95 % CI 40.2–79.3) in the WPRT group and 56.4 % (95 % CI 46.3–65.2) in the PBRT group (p = 0.257; Fig. 2a). Propensity score-matched cohort Propensity score matching resulted in 29 patients in WPRT the group and 58 patients in the PBRT group, i.e., 87 patients
13
C. Song et al.
in total. Except a pathologic Gleason score of ≥ 8, no characteristics of patient, tumor, or treatment were significantly different between groups of the matched cohort (Table 1), indicating that the matching procedure worked well. Use of WPRT resulted in improved 4-year bRFS rates compared to PBRT (63.1 %, 95 % CI 40.2–79.3 vs. 43.4 %, 95 % CI 28.2–57.6; p = 0.034; Fig. 2b). Subgroup analyses Patients stratified by “a priori risk group” When all 163 patients were stratified by “a priori risk group”, 29 patients were considered as low risk and 134 patients as high risk. WPRT conferred no benefit on either group. For the high-risk group, the 4-year bRFS rates of patients who received WPRT and PBRT were 61.4 (95 % CI 37.9–78.3) and 56.3 % (95 % CI 45.2–66.1), respectively (p = 0.339; Fig. 3). Patients stratified by pre-SRT PSA level ≥ 0.4 ng/mL When patients were grouped by pre-SRT PSA level, 69 patients had pre-SRT PSA level
