Brachytherapy 14 (2015) 502e510
Combination therapy improves prostate cancer survival for patients with potentially lethal prostate cancer: The impact of Gleason pattern 5 Adam L. Liss1, Eyad I. Abu-Isa2, Maha S. Jawad1,5, Felix Y. Feng1, Sean M. Vance1,6, Raymond J. Winfield3, Vrinda Narayana2, Howard M. Sandler4, P. William McLaughlin1,2, Daniel A. Hamstra1,* 1
Department of Radiation Oncology, University of Michigan, Ann Arbor, MI 2 Department of Radiation Oncology, Providence Hospital, Novi, MI 3 Department of Urology, Providence Hospital, Novi, MI 4 Department of Radiation Oncology, Cedars-Sinai Medical Center, Los Angeles, CA
ABSTRACT
PURPOSE: To investigate the impact of Gleason pattern 5 (GP5) prostate cancer after either external beam radiotherapy (EBRT) or the combination of EBRT with low-dose rate brachytherapy boost (combo). METHODS AND MATERIALS: Between 1998 and 2008, 467 patients with National Comprehensive Cancer Network high-risk prostate cancer were treated with EBRT (n 5 326) or combo (low-dose rate to 90e108 Gy using I-125 followed by EBRT) (n 5 141). Freedom from biochemical failure, freedom from metastasis (FFM), cancer-specific survival (CSS), and overall survival were evaluated. RESULTS: Combo patients were younger (66 vs. 72 years, p ! 0.001) and had fewer comorbidities (Charlson comorbidity index 3.7 vs. 4.4, p ! 0.001). EBRT patients had higher tumor stages (T3e4: 30% vs. 21%, p 5 0.03) and lower Gleason scores (8e10: 61% vs. 75%, p 5 0.01). Androgen deprivation therapy use was similar between cohorts (85% vs. 87%, p 5 0.5), but EBRT patients had longer androgen deprivation therapy use (median 14 vs. 12 months, p 5 0.05). GP5 predicted worse FFM ( p ! 0.001, hazard ratio [HR] 3.3, 95% confidence interval [CI]1.8e6.2]) and CSS ( p ! 0.001, HR 5.9, 95% CI 2.7e12.9) for the EBRT group, but not for the combo group ( p 5 0.86, HR 0.48, 95% CI 0.1e2.4 for metastasis and p 5 0.5, HR 1.6, 95% CI 0.33e8.0 for CSS). In those with GP5 (n 5 143), combo was associated with improved outcomes in all endpoints. On univariate analysis, 5-year outcomes for combo vs. EBRT were as follows: freedom from biochemical failure 89% vs. 65%, FFM 89% vs. 67%, CSS 93% vs. 78%, and overall survival 88% vs. 67% ( p ! 0.05 for all). CONCLUSION: Combo was associated with improved outcomes for men with GP5 prostate cancer. This highlights the importance of local therapy, especially in patients with the highest pathologic grade disease. Ó 2015 American Brachytherapy Society. Published by Elsevier Inc. All rights reserved.
Keywords:
Gleason pattern 5; Combination therapy; High-risk prostate cancer
Introduction Received 26 December 2014; received in revised form 16 February 2015; accepted 18 February 2015. Disclosure Statement: Dr. DAH received was a consultant for Medivation, Myriad and Varian. The other authors of the manuscript declare no actual or potential conflicts of interest. Dr. FYF received grant funding from Varian and from Medivation/Astellas and has served on advisory boards for Medivation/Astellas. * Corresponding author. University Hospital, 1500 E Medical Center Drive, UHB2C490, Ann Arbor, MI 48109. Tel.: 734-936-4300; fax: 734763-7371. E-mail address: [email protected] (D.A. Hamstra). 5 Department of Radiation Oncology, William Beaumont Hospital, 3601 W 13 Mile Rd, Royal Oak, MI 48073. 6 Department of Radiation Oncology, Henry Ford Hospital, 2799 W. Grand Blvd, Detroit, MI 48202.
Prostate cancer remains the most common cancer diagnosis in men, with an estimated 220,800 new diagnoses in 2015, and is the second most common cause of cancer death in men with 27,540 estimated cancer deaths (1). Although risk stratification can help define the high-risk subgroup that is most likely to have recurrence after initial therapy, improving outcomes and decreasing mortality among the high-risk patients remains an unsolved challenge. For many men with high-risk prostate cancer, radiation, surgery, or androgen deprivation therapy (ADT) alone are insufficient treatment (2e4). Combining treatment modalities has been shown to yield improved results (4, 5). It
1538-4721/$ - see front matter Ó 2015 American Brachytherapy Society. Published by Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.brachy.2015.02.389
A.L. Liss et al. / Brachytherapy 14 (2015) 502e510
is well established that the addition of ADT to conventional-dose (!72 Gy) external beam radiation therapy (EBRT) improves overall survival (OS) in randomized studies (2, 6, 7). Nevertheless, despite combined EBRT and ADT, clinical disease-free survival for men with high-risk disease remains less than optimal (6). Further stratifying patients with high-risk prostate cancer can identify groups of men with an even worse prognosis. Patients with higher Gleason scores (GS) have been shown to have an increased risk of recurrence and decreased prostate cancerespecific survival after treatment (8e10). The presence of Gleason pattern 5 (GP5) prostate cancer has been correlated with an increased risk of recurrence and metastasis after prostatectomy and salvage RT, definitive external beam radiation, and brachytherapy (11e13). Studies of high-risk prostate cancer have identified the presence of GP5 as the strongest prognostic factor of all clinical endpoints (11, 12). The significance of GP5 for patients treated with EBRT and low-dose rate (LDR) brachytherapy boost (combo) has not been fully examined. It was our hypothesis that the enhancement in local radiation dose associated with combo treatment might offer superior local control over EBRT alone. Therefore, we sought to investigate if GP5 would predict for improved clinical outcomes for patients treated with combo as compared with dose-escalated EBRT.
503 125
permanent interstitial brachytherapy implant, using I for all implants (90e108 Gy), followed by EBRT using three-dimensional conformal or intensity modulated radiotherapy with MRI and CT planning. The EBRT dose was calculated with consideration of the implant dosimetry to achieve a total external beam equivalent dose to the prostate plus 0.5-cm margin of 90 Gy (prostate O105 Gy) and 45 Gy to the pelvic lymph nodes. The median postimplant day 21 D90 was 113% of the prescription dose and the median postimplant V100 was 94.8%. The prostate, seminal vesicles, and pelvic lymph nodes were treated using EBRT in the combo patients as well, after LDR implant. ADT was prescribed based on the discretion of the treating physician. Endpoints
Methods and materials
Outcomes measured consisted of freedom from biochemical failure (FFBF), freedom from metastasis (FFM), cancer-specific survival (CSS), and OS. Time to biochemical failure (BF), distant metastasis (DM), and CSS were calculated from the initiation of therapy (either start of ADT if given neoadjuvantly or start of RT). BF was defined per the Phoenix definition as a serum PSA level at least 2 ng/mL greater than the posttreatment PSA nadir (14). DM was defined as the presence of clinical, radiographic, or pathologic evidence of metastatic disease. Cancer-specific mortality was defined as a death attributed to prostate cancer, or any death in a patient after metastasis or the development of castrate-resistant prostate cancer.
Patient selection
Statistical analysis
Through an institutional review boardeapproved retrospective analysis, we identified 467 patients who received definitive radiation with or without ADT for high-risk prostate cancer (defined by the National Comprehensive Cancer Network [NCCN] as T3eT4, GS 8e10, or prostate-specific antigen [PSA] O20 ng/mL). Patients were considered in the GP5 subgroup if GP5 was present as either the primary or secondary Gleason pattern in at least one core. Tertiary pattern 5 was not routinely collected during this period and was not included. All patients studied received treatment from 1998 to 2008 at the University of Michigan or at regional practices affiliated with the University of Michigan and staffed by University of Michigan physicians and physics faculty. Treatment consisted of either EBRT (n 5 326) or combo (n 5 141), with most patients also receiving ADT.
Differences between categorical variable frequencies were tested by c2 or Fisher exact test, whereas differences between continuous variables were determined via one-way analysis of variance. Univariate survival analyses were performed using the log-rank test and Cox proportional hazards models were used for multivariate analyses.
Treatment EBRT treatment consisted of minimum planning target volume coverage of 75 Gy (range, 75.0e79.2 Gy) using three-dimensional conformal or intensity modulated radiotherapy with CT-based planning and conventional fractions of 1.8 to 2.0 Gy daily. Targets for patients treated with EBRT consisted of the prostate, seminal vesicles, and pelvic lymph nodes. Combo therapy consisted of an LDR
Results Patient characteristics The median age of all patients was 69 years, with combo patients younger as compared with those treated with EBRT (median 66 vs. 72 years; p ! 0.001, Table 1). Those treated with LDR brachytherapy also had less comorbid illness (mean age-adjusted Charlson comorbidity index of 3.7) as compared those treated with EBRT (mean Charlson comorbidity index 4.4, p ! 0.001). PSA levels were similar in both cohorts of patients with a mean PSA of 25.9 ng/mL (median 20.4 ng/mL) in the combo group and 27.4 ng/mL (median 13.5 ng/mL) in the EBRT group ( p 5 0.64). Patients in the EBRT cohort were more likely to have T3 or T4 prostate cancer as compared with patients treated with combo therapy (30% vs. 21%, p 5 0.03). Patients treated
504
A.L. Liss et al. / Brachytherapy 14 (2015) 502e510
Table 1 Patient characteristics Characteristic Number Median followup, mo Median age, y Mean CAPRA score Mean age-adjusted Charlson CMI Mean PSA (median), ng/mL T-stage T1eT2a T2beT2c T3eT4 Gleason score 2e6 7 8 (no GP5) 8e10 (with GP5) ADT use Overall Use O2 y Median duration (IQR) Median duration no GP5 (IQR) Median duration with GP5 (IQR)
Univariate analyses Combo
EBRT
p Value
141 57 66 5.8 3.7
326 70 72 5.9 4.4
!0.1 !0.001 0.5 !0.001
25.9 (20.4)
27.4 (13.5)
0.64
60% 19% 21%
46% 24% 30%
0.03
4% 21% 41% 34%
12% 26% 32% 29%
87% 13% 12 mo (9e18) 12 mo (11.5e18)
85% 33% 14 mo (6e24) 12 mo (6.1e24)
0.5
12 mo (9e13)
23.9 mo (9e27)
0.008
0.3
0.05 0.5
CAPRA 5 cancer of the prostate risk assessment score; CMI 5 comorbidity index; EBRT 5 external beam radiotherapy; GP5 5 Gleason pattern 5; ADT 5 androgen deprivation therapy; IQR 5 interquartile range; mo 5 months; PSA 5 prostate-specific antigen.
with combo, on the other hand, were more likely to have GS 8 to 10 compared with the EBRT patient cohort (75% vs. 61%, p 5 0.01). There was no difference in the number of patients with GP5 between treatment groups (29% EBRT vs. 34% combo, p 5 0.3). To attempt to adjust for these clinical risk differences, the cancer of the prostate risk assessment score was calculated for both groups of patients yielding a score between 0 and 10, and there was no overall prognostic difference between treatment groups, with an average score of 5.9 for the EBRT group and 5.8 for the combo group ( p 5 0.5) (15). For treatment characteristics, there was no difference in the overall rate of ADT use between groups, with 85% of men in the EBRT group and 87% of men in the combo group receiving ADT ( p 5 0.5). However, men in the combo therapy group tended to have a shorter duration of ADT use (median 12 months, interquartile range [IQR] 9e18 months, 13% use greater than 2 years) compared with the EBRT group (median 14 months, IQR 6e24 months, 33% use greater than 2 years, p 5 0.05 for duration). Patients with GP5 in the EBRT cohort also had longer use of ADT (median 23.9 months, IQR 9e27 months) as compared with patients with GP5 in the combo cohort (median 12 months, IQR 9e13 months; p 5 0.008). Finally, median followup was shorter in those treated with combo at 57 months vs. 70 months for EBRT ( p ! 0.1).
There was a significant improvement in BF for NCCN high-risk patients treated with combo therapy as compared with the EBRT cohort ( p 5 0.046, hazard ratio [HR] 0.59, 95% confidence interval [CI] 0.38e0.92) (Fig. 1a). The 5year rate of BF was 30% (3% [S.E.M.]) with EBRT and 20% (6%) with combo. There was, however, no statistically significant difference in FFM ( p 5 0.38, HR 0.75, 95% CI 0.4e1.4) or CSS ( p 5 0.23, HR 0.63, 95% CI 0.3e1.2) for all NCCN high-risk patients in the combo cohort as compared with the EBRT cohort (Figs. 1b and 1c). The 5-year FFM for those with NCCN high-risk disease was 83% (2%) with EBRT and 87% (5%) with combo therapy and the 5-year CSS for those with NCCN high-risk disease was 93% (2%) with EBRT and 94% (2%) with combo therapy. Although a statistically significant difference in OS was noted among the NCCN highrisk patients in the combo cohort as compared with the EBRT cohort ( p 5 0.02), this likely is resultant from the younger age and decreased comorbidities of the patients in the combo cohort (Fig. 1d).
GP5 endpoint analysis Given previous results suggesting that Gleason score is the most prognostic factor for clinical outcome after radiotherapy, all 467 patients were stratified by Gleason score to assess endpoints by increasing GS as follows: GS 2 to 6, GS 7, GS 4 þ 4 5 8, and any GP5 (12, 13). A significantly higher rate of metastasis was noted for patients with GP5 as compared with all others in the study ( p ! 0.0001, HR 2.6, 95% CI 1.5e4.7). Similarly, a significantly higher rate of prostate cancer death was noted for patients with GP5 as compared with other GS ( p ! 0.0001, HR 4.3, 95% CI 2.2e8.6). In addition, there was no difference in DM ( p O 0.05) or CSS ( p O 0.1) as stratified by Gleason score when GP5 patients were excluded. When patients were stratified by both GS and treatment cohort, the increased risk of metastasis and prostate cancer death associated with GP5 was not noted in the combo cohort, but was demonstrated only for patients in the EBRT cohort (Fig. 2). Specifically, the rate of metastasis at 5 years for patients with GP5 treated with EBRTwas 33%, and this was substantially greater when compared with all other GS in those treated with EBRT ( p ! 0.0001, HR 3.3, 95% CI 1.8e6.2) (Fig. 2a). In comparison, the 5-year rate of metastasis for GP5 patients treated with combo was 11% without a statistically significant difference compared with other GS when treated with combo ( p 5 0.86, HR 0.48, 95% CI 0.1e2.4) (Fig. 2b). The rate of prostate cancer death at 5 years for patients with GP5 in the EBRT cohort was 22%, whereas for patients treated with combo therapy the rate of prostate cancer death was only 7%. The HR for prostate cancer death for GP5 patients in the EBRT group was 5.9 ( p ! 0.0001, 95% CI 2.7e12.9) (Fig. 2c). The HR for prostate cancer death for
A.L. Liss et al. / Brachytherapy 14 (2015) 502e510
505
Combo vs. EBRT for All High-risk Patients b
Freedom From Biochemical Failure
Freedom From Metastasis
100
100
80
80
Survival Probability (%)
Survival Probability (%)
a
60
40
p=0.046 (HR: 0.59 [95% CI: 0.38-0.92]) EBRT EBRT + LDR
20
0
60
40 p=0.38 (HR: 0.75 [95%CI: 0.41-1.4]) EBRT EBRT + LDR
20
0 0
12
24
36
48
60
72
84
96
108
0
120
12
24
Time (months)
d
Cancer Specific Survival
60
72
84
96
72
84
96
Overall Survival
100
100
80
80
60
40 p=0.23 (HR: 0.63 [95%CI: 0.33-1.2]) EBRT EBRT + LDR
20
48
Time (months)
Survival Probability (%)
Survival Probability (%)
c
36
0
60
40 p=0.02 (HR: 0.55 [95%CI: 0.37-0.83]) EBRT EBRT + LDR
20
0 0
12
24
36
48
60
72
84
96
0
12
24
Time (months)
36
48
60
Time (months)
Fig. 1. Freedom from biochemical failure, freedom from metastasis, cancer-specific survival, and overall survival for all National Comprehensive Cancer Network high-risk patients are depicted above following univariate analysis (aed). Combo therapy was associated with a statistically significant improvement in overall survival and freedom from biochemical failure at 5 years as compared with external beam radiotherapy. There was a trend toward improvement in cancer-specific survival and freedom from metastasis, but this did not reach statistical significance.
GP5 patients in the combo group was not significant at 1.6 ( p 5 0.89, 95% CI 0.33e8.0) (Fig. 2d). As an additional analysis of the role of GP5 in our study, we separated all 467 NCCN high-risk patients into two groups: patients without GP5 (n 5 324) and patients with GP5 (n 5 143). For patients with GP5, there was a statistically significant advantage with respect to FFBF, FFM, and CSS at 5 years for those patients treated with combination therapy as compared with EBRT (Fig. 3). Specifically, the 5-year FFBF for those with GP5 was 65% (6%) with EBRT and 89% (6%) with combo, translating into an approximate 60% relative risk reduction ( p 5 0.046, HR 0.37, 95% CI 0.17e0.78) (Fig. 3a). The 5-year FFM for those with GP5 was 67% (5%) with EBRT and 89% (5%) with combo, translating into an approximate 80% relative risk reduction ( p 5 0.01, HR 0.12, 95% CI 0.05e0.27) (Fig. 3b). The 5-year CSS for those with GP5 was 78% (5%) with EBRT and 93% (5%) with combo, translating into an approximate 75% relative risk reduction
( p 5 0.012, HR 0.24, 95% CI 0.11e0.54) (Fig. 3c). Finally, the 5-year OS for those with GP5 was 67% with EBRT and 88% with combo ( p 5 0.002, HR 0.28, 95% CI 0.16e0.52), although a likely explanation for this finding is that patients treated with combination therapy were younger with fewer medical comorbidities. Nevertheless, the 15% difference in CSS at 5 years might be expected to make up a substantial portion of the 21% difference observed in OS. Further, when excluding patients with GP5, there was no significant difference in any of the 5year endpoints for the patients with NCCN high-risk prostate cancer based on radiation treatment (Table 2). Multivariate analysis Given that both GP5 and brachytherapy boost seemed to have a significant impact on clinical outcome, we performed a multivariate analysis of the patients with GS 8 to 10 (Table 3). Variables included the following: PSA, GS,
506
A.L. Liss et al. / Brachytherapy 14 (2015) 502e510
EBRT and Combo Cohorts Stratified by Gleason Score a
b
Rate of Metastasis (EBRT alone)
Rate of Metastasis (Combo) 60
p
Combination therapy improves prostate cancer survival for patients with potentially lethal prostate cancer: The impact of Gleason pattern 5 Adam L. Liss1, Eyad I. Abu-Isa2, Maha S. Jawad1,5, Felix Y. Feng1, Sean M. Vance1,6, Raymond J. Winfield3, Vrinda Narayana2, Howard M. Sandler4, P. William McLaughlin1,2, Daniel A. Hamstra1,* 1
Department of Radiation Oncology, University of Michigan, Ann Arbor, MI 2 Department of Radiation Oncology, Providence Hospital, Novi, MI 3 Department of Urology, Providence Hospital, Novi, MI 4 Department of Radiation Oncology, Cedars-Sinai Medical Center, Los Angeles, CA
ABSTRACT
PURPOSE: To investigate the impact of Gleason pattern 5 (GP5) prostate cancer after either external beam radiotherapy (EBRT) or the combination of EBRT with low-dose rate brachytherapy boost (combo). METHODS AND MATERIALS: Between 1998 and 2008, 467 patients with National Comprehensive Cancer Network high-risk prostate cancer were treated with EBRT (n 5 326) or combo (low-dose rate to 90e108 Gy using I-125 followed by EBRT) (n 5 141). Freedom from biochemical failure, freedom from metastasis (FFM), cancer-specific survival (CSS), and overall survival were evaluated. RESULTS: Combo patients were younger (66 vs. 72 years, p ! 0.001) and had fewer comorbidities (Charlson comorbidity index 3.7 vs. 4.4, p ! 0.001). EBRT patients had higher tumor stages (T3e4: 30% vs. 21%, p 5 0.03) and lower Gleason scores (8e10: 61% vs. 75%, p 5 0.01). Androgen deprivation therapy use was similar between cohorts (85% vs. 87%, p 5 0.5), but EBRT patients had longer androgen deprivation therapy use (median 14 vs. 12 months, p 5 0.05). GP5 predicted worse FFM ( p ! 0.001, hazard ratio [HR] 3.3, 95% confidence interval [CI]1.8e6.2]) and CSS ( p ! 0.001, HR 5.9, 95% CI 2.7e12.9) for the EBRT group, but not for the combo group ( p 5 0.86, HR 0.48, 95% CI 0.1e2.4 for metastasis and p 5 0.5, HR 1.6, 95% CI 0.33e8.0 for CSS). In those with GP5 (n 5 143), combo was associated with improved outcomes in all endpoints. On univariate analysis, 5-year outcomes for combo vs. EBRT were as follows: freedom from biochemical failure 89% vs. 65%, FFM 89% vs. 67%, CSS 93% vs. 78%, and overall survival 88% vs. 67% ( p ! 0.05 for all). CONCLUSION: Combo was associated with improved outcomes for men with GP5 prostate cancer. This highlights the importance of local therapy, especially in patients with the highest pathologic grade disease. Ó 2015 American Brachytherapy Society. Published by Elsevier Inc. All rights reserved.
Keywords:
Gleason pattern 5; Combination therapy; High-risk prostate cancer
Introduction Received 26 December 2014; received in revised form 16 February 2015; accepted 18 February 2015. Disclosure Statement: Dr. DAH received was a consultant for Medivation, Myriad and Varian. The other authors of the manuscript declare no actual or potential conflicts of interest. Dr. FYF received grant funding from Varian and from Medivation/Astellas and has served on advisory boards for Medivation/Astellas. * Corresponding author. University Hospital, 1500 E Medical Center Drive, UHB2C490, Ann Arbor, MI 48109. Tel.: 734-936-4300; fax: 734763-7371. E-mail address: [email protected] (D.A. Hamstra). 5 Department of Radiation Oncology, William Beaumont Hospital, 3601 W 13 Mile Rd, Royal Oak, MI 48073. 6 Department of Radiation Oncology, Henry Ford Hospital, 2799 W. Grand Blvd, Detroit, MI 48202.
Prostate cancer remains the most common cancer diagnosis in men, with an estimated 220,800 new diagnoses in 2015, and is the second most common cause of cancer death in men with 27,540 estimated cancer deaths (1). Although risk stratification can help define the high-risk subgroup that is most likely to have recurrence after initial therapy, improving outcomes and decreasing mortality among the high-risk patients remains an unsolved challenge. For many men with high-risk prostate cancer, radiation, surgery, or androgen deprivation therapy (ADT) alone are insufficient treatment (2e4). Combining treatment modalities has been shown to yield improved results (4, 5). It
1538-4721/$ - see front matter Ó 2015 American Brachytherapy Society. Published by Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.brachy.2015.02.389
A.L. Liss et al. / Brachytherapy 14 (2015) 502e510
is well established that the addition of ADT to conventional-dose (!72 Gy) external beam radiation therapy (EBRT) improves overall survival (OS) in randomized studies (2, 6, 7). Nevertheless, despite combined EBRT and ADT, clinical disease-free survival for men with high-risk disease remains less than optimal (6). Further stratifying patients with high-risk prostate cancer can identify groups of men with an even worse prognosis. Patients with higher Gleason scores (GS) have been shown to have an increased risk of recurrence and decreased prostate cancerespecific survival after treatment (8e10). The presence of Gleason pattern 5 (GP5) prostate cancer has been correlated with an increased risk of recurrence and metastasis after prostatectomy and salvage RT, definitive external beam radiation, and brachytherapy (11e13). Studies of high-risk prostate cancer have identified the presence of GP5 as the strongest prognostic factor of all clinical endpoints (11, 12). The significance of GP5 for patients treated with EBRT and low-dose rate (LDR) brachytherapy boost (combo) has not been fully examined. It was our hypothesis that the enhancement in local radiation dose associated with combo treatment might offer superior local control over EBRT alone. Therefore, we sought to investigate if GP5 would predict for improved clinical outcomes for patients treated with combo as compared with dose-escalated EBRT.
503 125
permanent interstitial brachytherapy implant, using I for all implants (90e108 Gy), followed by EBRT using three-dimensional conformal or intensity modulated radiotherapy with MRI and CT planning. The EBRT dose was calculated with consideration of the implant dosimetry to achieve a total external beam equivalent dose to the prostate plus 0.5-cm margin of 90 Gy (prostate O105 Gy) and 45 Gy to the pelvic lymph nodes. The median postimplant day 21 D90 was 113% of the prescription dose and the median postimplant V100 was 94.8%. The prostate, seminal vesicles, and pelvic lymph nodes were treated using EBRT in the combo patients as well, after LDR implant. ADT was prescribed based on the discretion of the treating physician. Endpoints
Methods and materials
Outcomes measured consisted of freedom from biochemical failure (FFBF), freedom from metastasis (FFM), cancer-specific survival (CSS), and OS. Time to biochemical failure (BF), distant metastasis (DM), and CSS were calculated from the initiation of therapy (either start of ADT if given neoadjuvantly or start of RT). BF was defined per the Phoenix definition as a serum PSA level at least 2 ng/mL greater than the posttreatment PSA nadir (14). DM was defined as the presence of clinical, radiographic, or pathologic evidence of metastatic disease. Cancer-specific mortality was defined as a death attributed to prostate cancer, or any death in a patient after metastasis or the development of castrate-resistant prostate cancer.
Patient selection
Statistical analysis
Through an institutional review boardeapproved retrospective analysis, we identified 467 patients who received definitive radiation with or without ADT for high-risk prostate cancer (defined by the National Comprehensive Cancer Network [NCCN] as T3eT4, GS 8e10, or prostate-specific antigen [PSA] O20 ng/mL). Patients were considered in the GP5 subgroup if GP5 was present as either the primary or secondary Gleason pattern in at least one core. Tertiary pattern 5 was not routinely collected during this period and was not included. All patients studied received treatment from 1998 to 2008 at the University of Michigan or at regional practices affiliated with the University of Michigan and staffed by University of Michigan physicians and physics faculty. Treatment consisted of either EBRT (n 5 326) or combo (n 5 141), with most patients also receiving ADT.
Differences between categorical variable frequencies were tested by c2 or Fisher exact test, whereas differences between continuous variables were determined via one-way analysis of variance. Univariate survival analyses were performed using the log-rank test and Cox proportional hazards models were used for multivariate analyses.
Treatment EBRT treatment consisted of minimum planning target volume coverage of 75 Gy (range, 75.0e79.2 Gy) using three-dimensional conformal or intensity modulated radiotherapy with CT-based planning and conventional fractions of 1.8 to 2.0 Gy daily. Targets for patients treated with EBRT consisted of the prostate, seminal vesicles, and pelvic lymph nodes. Combo therapy consisted of an LDR
Results Patient characteristics The median age of all patients was 69 years, with combo patients younger as compared with those treated with EBRT (median 66 vs. 72 years; p ! 0.001, Table 1). Those treated with LDR brachytherapy also had less comorbid illness (mean age-adjusted Charlson comorbidity index of 3.7) as compared those treated with EBRT (mean Charlson comorbidity index 4.4, p ! 0.001). PSA levels were similar in both cohorts of patients with a mean PSA of 25.9 ng/mL (median 20.4 ng/mL) in the combo group and 27.4 ng/mL (median 13.5 ng/mL) in the EBRT group ( p 5 0.64). Patients in the EBRT cohort were more likely to have T3 or T4 prostate cancer as compared with patients treated with combo therapy (30% vs. 21%, p 5 0.03). Patients treated
504
A.L. Liss et al. / Brachytherapy 14 (2015) 502e510
Table 1 Patient characteristics Characteristic Number Median followup, mo Median age, y Mean CAPRA score Mean age-adjusted Charlson CMI Mean PSA (median), ng/mL T-stage T1eT2a T2beT2c T3eT4 Gleason score 2e6 7 8 (no GP5) 8e10 (with GP5) ADT use Overall Use O2 y Median duration (IQR) Median duration no GP5 (IQR) Median duration with GP5 (IQR)
Univariate analyses Combo
EBRT
p Value
141 57 66 5.8 3.7
326 70 72 5.9 4.4
!0.1 !0.001 0.5 !0.001
25.9 (20.4)
27.4 (13.5)
0.64
60% 19% 21%
46% 24% 30%
0.03
4% 21% 41% 34%
12% 26% 32% 29%
87% 13% 12 mo (9e18) 12 mo (11.5e18)
85% 33% 14 mo (6e24) 12 mo (6.1e24)
0.5
12 mo (9e13)
23.9 mo (9e27)
0.008
0.3
0.05 0.5
CAPRA 5 cancer of the prostate risk assessment score; CMI 5 comorbidity index; EBRT 5 external beam radiotherapy; GP5 5 Gleason pattern 5; ADT 5 androgen deprivation therapy; IQR 5 interquartile range; mo 5 months; PSA 5 prostate-specific antigen.
with combo, on the other hand, were more likely to have GS 8 to 10 compared with the EBRT patient cohort (75% vs. 61%, p 5 0.01). There was no difference in the number of patients with GP5 between treatment groups (29% EBRT vs. 34% combo, p 5 0.3). To attempt to adjust for these clinical risk differences, the cancer of the prostate risk assessment score was calculated for both groups of patients yielding a score between 0 and 10, and there was no overall prognostic difference between treatment groups, with an average score of 5.9 for the EBRT group and 5.8 for the combo group ( p 5 0.5) (15). For treatment characteristics, there was no difference in the overall rate of ADT use between groups, with 85% of men in the EBRT group and 87% of men in the combo group receiving ADT ( p 5 0.5). However, men in the combo therapy group tended to have a shorter duration of ADT use (median 12 months, interquartile range [IQR] 9e18 months, 13% use greater than 2 years) compared with the EBRT group (median 14 months, IQR 6e24 months, 33% use greater than 2 years, p 5 0.05 for duration). Patients with GP5 in the EBRT cohort also had longer use of ADT (median 23.9 months, IQR 9e27 months) as compared with patients with GP5 in the combo cohort (median 12 months, IQR 9e13 months; p 5 0.008). Finally, median followup was shorter in those treated with combo at 57 months vs. 70 months for EBRT ( p ! 0.1).
There was a significant improvement in BF for NCCN high-risk patients treated with combo therapy as compared with the EBRT cohort ( p 5 0.046, hazard ratio [HR] 0.59, 95% confidence interval [CI] 0.38e0.92) (Fig. 1a). The 5year rate of BF was 30% (3% [S.E.M.]) with EBRT and 20% (6%) with combo. There was, however, no statistically significant difference in FFM ( p 5 0.38, HR 0.75, 95% CI 0.4e1.4) or CSS ( p 5 0.23, HR 0.63, 95% CI 0.3e1.2) for all NCCN high-risk patients in the combo cohort as compared with the EBRT cohort (Figs. 1b and 1c). The 5-year FFM for those with NCCN high-risk disease was 83% (2%) with EBRT and 87% (5%) with combo therapy and the 5-year CSS for those with NCCN high-risk disease was 93% (2%) with EBRT and 94% (2%) with combo therapy. Although a statistically significant difference in OS was noted among the NCCN highrisk patients in the combo cohort as compared with the EBRT cohort ( p 5 0.02), this likely is resultant from the younger age and decreased comorbidities of the patients in the combo cohort (Fig. 1d).
GP5 endpoint analysis Given previous results suggesting that Gleason score is the most prognostic factor for clinical outcome after radiotherapy, all 467 patients were stratified by Gleason score to assess endpoints by increasing GS as follows: GS 2 to 6, GS 7, GS 4 þ 4 5 8, and any GP5 (12, 13). A significantly higher rate of metastasis was noted for patients with GP5 as compared with all others in the study ( p ! 0.0001, HR 2.6, 95% CI 1.5e4.7). Similarly, a significantly higher rate of prostate cancer death was noted for patients with GP5 as compared with other GS ( p ! 0.0001, HR 4.3, 95% CI 2.2e8.6). In addition, there was no difference in DM ( p O 0.05) or CSS ( p O 0.1) as stratified by Gleason score when GP5 patients were excluded. When patients were stratified by both GS and treatment cohort, the increased risk of metastasis and prostate cancer death associated with GP5 was not noted in the combo cohort, but was demonstrated only for patients in the EBRT cohort (Fig. 2). Specifically, the rate of metastasis at 5 years for patients with GP5 treated with EBRTwas 33%, and this was substantially greater when compared with all other GS in those treated with EBRT ( p ! 0.0001, HR 3.3, 95% CI 1.8e6.2) (Fig. 2a). In comparison, the 5-year rate of metastasis for GP5 patients treated with combo was 11% without a statistically significant difference compared with other GS when treated with combo ( p 5 0.86, HR 0.48, 95% CI 0.1e2.4) (Fig. 2b). The rate of prostate cancer death at 5 years for patients with GP5 in the EBRT cohort was 22%, whereas for patients treated with combo therapy the rate of prostate cancer death was only 7%. The HR for prostate cancer death for GP5 patients in the EBRT group was 5.9 ( p ! 0.0001, 95% CI 2.7e12.9) (Fig. 2c). The HR for prostate cancer death for
A.L. Liss et al. / Brachytherapy 14 (2015) 502e510
505
Combo vs. EBRT for All High-risk Patients b
Freedom From Biochemical Failure
Freedom From Metastasis
100
100
80
80
Survival Probability (%)
Survival Probability (%)
a
60
40
p=0.046 (HR: 0.59 [95% CI: 0.38-0.92]) EBRT EBRT + LDR
20
0
60
40 p=0.38 (HR: 0.75 [95%CI: 0.41-1.4]) EBRT EBRT + LDR
20
0 0
12
24
36
48
60
72
84
96
108
0
120
12
24
Time (months)
d
Cancer Specific Survival
60
72
84
96
72
84
96
Overall Survival
100
100
80
80
60
40 p=0.23 (HR: 0.63 [95%CI: 0.33-1.2]) EBRT EBRT + LDR
20
48
Time (months)
Survival Probability (%)
Survival Probability (%)
c
36
0
60
40 p=0.02 (HR: 0.55 [95%CI: 0.37-0.83]) EBRT EBRT + LDR
20
0 0
12
24
36
48
60
72
84
96
0
12
24
Time (months)
36
48
60
Time (months)
Fig. 1. Freedom from biochemical failure, freedom from metastasis, cancer-specific survival, and overall survival for all National Comprehensive Cancer Network high-risk patients are depicted above following univariate analysis (aed). Combo therapy was associated with a statistically significant improvement in overall survival and freedom from biochemical failure at 5 years as compared with external beam radiotherapy. There was a trend toward improvement in cancer-specific survival and freedom from metastasis, but this did not reach statistical significance.
GP5 patients in the combo group was not significant at 1.6 ( p 5 0.89, 95% CI 0.33e8.0) (Fig. 2d). As an additional analysis of the role of GP5 in our study, we separated all 467 NCCN high-risk patients into two groups: patients without GP5 (n 5 324) and patients with GP5 (n 5 143). For patients with GP5, there was a statistically significant advantage with respect to FFBF, FFM, and CSS at 5 years for those patients treated with combination therapy as compared with EBRT (Fig. 3). Specifically, the 5-year FFBF for those with GP5 was 65% (6%) with EBRT and 89% (6%) with combo, translating into an approximate 60% relative risk reduction ( p 5 0.046, HR 0.37, 95% CI 0.17e0.78) (Fig. 3a). The 5-year FFM for those with GP5 was 67% (5%) with EBRT and 89% (5%) with combo, translating into an approximate 80% relative risk reduction ( p 5 0.01, HR 0.12, 95% CI 0.05e0.27) (Fig. 3b). The 5-year CSS for those with GP5 was 78% (5%) with EBRT and 93% (5%) with combo, translating into an approximate 75% relative risk reduction
( p 5 0.012, HR 0.24, 95% CI 0.11e0.54) (Fig. 3c). Finally, the 5-year OS for those with GP5 was 67% with EBRT and 88% with combo ( p 5 0.002, HR 0.28, 95% CI 0.16e0.52), although a likely explanation for this finding is that patients treated with combination therapy were younger with fewer medical comorbidities. Nevertheless, the 15% difference in CSS at 5 years might be expected to make up a substantial portion of the 21% difference observed in OS. Further, when excluding patients with GP5, there was no significant difference in any of the 5year endpoints for the patients with NCCN high-risk prostate cancer based on radiation treatment (Table 2). Multivariate analysis Given that both GP5 and brachytherapy boost seemed to have a significant impact on clinical outcome, we performed a multivariate analysis of the patients with GS 8 to 10 (Table 3). Variables included the following: PSA, GS,
506
A.L. Liss et al. / Brachytherapy 14 (2015) 502e510
EBRT and Combo Cohorts Stratified by Gleason Score a
b
Rate of Metastasis (EBRT alone)
Rate of Metastasis (Combo) 60
p
