You are here
Continued Benefit to Androgen Deprivation Therapy for Prostate Cancer Patients Treated With Dose-Escalated Radiation Therapy Across Multiple Definitions of High-Risk Disease
International Journal of Radiation Oncology*Biology*Physics, 4, 81, pages e335 - e344
To analyze prognostic factors in patients with high-risk prostate cancer treated with dose-escalated external-beam radiation therapy (EBRT) and androgen deprivation (ADT).
Methods and Materials
Between 1998 and 2008 at the University of Michigan Medical Center, 718 men were consecutively treated with EBRT to at least 75 Gy. Seven definitions of high-risk prostate cancer, applying to 11–33% of patients, were evaluated. Biochemical failure (BF), salvage ADT use, metastatic progression, and prostate cancer–specific mortality (PCSM) were estimated by the Kaplan-Meier method and Cox proportional hazards regression.
Each high-risk definition was associated with increased BF (hazard ratio [HR] 2.8–3.9, p < 0.0001), salvage ADT use (HR 3.9–6.3, p < 0.0001), metastasis (HR 3.7–6.6, p < 0.0001), and PCSM (HR 3.7–16.2, p < 0.0001). Furthermore, an increasing number of high-risk features predicted worse outcome. Adjuvant ADT yielded significant reductions in both metastases (HR 0.19–0.38, p < 0.001) and PCSM (HR 0.38–0.50, p < 0.05) for all high-risk definitions (with the exception of clinical Stage T3–4 disease) but improved BF only for those with elevated Gleason scores (p < 0.03, HR 0.25–0.48). When treated with ADT and dose-escalated EBRT, patients with Gleason scores 8 to 10, without other high-risk features, had 8-year freedom from BF of 74%, freedom from distant metastases of 93%, and cause-specific survival of 92%, with salvage ADT used in 16% of patients.
Adjuvant ADT results in a significant improvement in clinical progression and PCSM across multiple definitions of high-risk disease even with dose-escalated EBRT. There is a subset of patients, characterized by multiple high-risk features or the presence of Gleason Pattern 5, who remain at significant risk for metastasis and PCSM despite current treatment.
Prostate, Androgen deprivation therapy, Salvage therapy, Metastasis, Cause-specific survival.
Prostate-specific antigen (PSA), T stage, and biopsy Gleason score (GS) are used to assess the risk of prostate cancer recurrence (1), (2), (3), and (4). The National Comprehensive Cancer Network (NCCN) guidelines to predict BF define high risk based on the presence of at least one adverse prognostic factor: cT3 or greater, GS 8 to 10, or pretreatment PSA level >20 ng/mL (4) . Biochemical recurrence, however, is a poor surrogate endpoint, which does not necessarily culminate in salvage therapy, distant metastases, or death from prostate cancer (5) .
Randomized trials demonstrated improvements in prostate cancer–specific survival and/or overall survival with the combination of ADT and conventional-dose external-beam radiation therapy (EBRT) in patients with locally advanced prostate cancer (6), (7), (8), (9), and (10). Four randomized trials have also shown an improvement in biochemical control with higher doses of EBRT (74–79.2 Gy vs. 64.8–70.2 Gy) (11), (12), (13), and (14). In the context of dose-escalated EBRT, the relevance of different risk stratification schemes and the impact of ADT remain unclear. Therefore, we analyzed BF, salvage ADT use, the risk of metastatic progression, and prostate cancer–specific mortality (PCSM) using seven different commonly used high-risk definitions, including GS 8 to 10, Gleason Pattern 5, PSA>20 ng/mL, cT3/4, NCCN high risk, the number of high-risk features, and the Kattan recurrence nomogram.
Methods and Materials
At the University of Michigan Medical Center between 1998 and 2008, 718 men with prostate cancer were treated to a planning target volume prescription dose of at least 75 Gy with or without neoadjuvant or adjuvant ADT. Patients with evidence of metastatic disease (not including pelvic lymph nodes) based on computed tomography of the abdomen/pelvis or bone scan were excluded. High-risk patients were identified using seven different definitions: (1) clinical Stage T3–4 (15) , (2) biopsy GS 8–10 (15) , (3) biopsy Gleason Pattern 5 (16) , (4) pretreatment PSA >20 ng/mL (15) , (5) NCCN criteria of cT3–4 or GS 8–10 or pretreatment PSA level >20 ng/mL (4) , (6) number of NCCN high-risk features (17) , and (7) Kattan pretreatment nomogram 5-year recurrence-free probability ≤50% (18) .
The techniques for treatment have been previously described (19) . Briefly, all treatments used computed tomography planning with either three-dimensional conformal radiation therapy (RT) or intensity-modulated RT. The prescribed radiation dose represented the minimum isodose line encompassing 95% of the planning target volume. The prescribed dose levels ranged from 75 to 79.2 Gy using daily fractions of 1.8 to 2.0 Gy. For low-risk patients, the clinical target volume (CTV) included the prostate only, and for intermediate-risk and high-risk patients, the CTV consisted of the prostate and seminal vesicles. Elective pelvic lymph node radiation to 45.0 Gy was routinely used for high-risk patients, four-field box to treat a minipelvic field (with the upper border most often placed at the bottom of the sacroiliac joints), followed by a boost to the CTV using six to eight fields. Per institutional practice, ADT was commonly used for high-risk patients and consisted of either monotherapy with a gonadotropin-releasing hormone analog or combined blockade with a gonadotropin-releasing hormone analog and an antiandrogen.
Patients were routinely followed up at 3- to 6-month intervals for the first 5 years and every 6 to 12 months thereafter. Biochemical failure (BF) was based on the Radiation Therapy Oncology Group (RTOG) Phoenix definition (nadir + 2 ng/mL) (20) . Freedom from distant metastasis (FFM) was defined as the absence of clinical, radiographic, or pathologic evidence of metastatic disease. PCSM was defined as death attributed to prostate cancer and was also presumed in any patient with either castration-resistant prostate cancer or metastatic disease before death. Salvage ADT (defined as the administration of ADT starting more than 1 month after the end of RT) was initiated at the discretion of the treating physician but was typically used to treat BF before evidence of metastatic disease.
The chi-square test and one-way analysis of variance were used to compare pretreatment variables. Survival analyses used the Kaplan-Meier method. To compare estimates between the high-risk and non–high-risk subsets, hazard ratios were calculated using the log-rank test, except in the case of the multipart “number of high-risk features” model, where Cox proportional hazards regression analysis was used. ADT was analyzed with regard to its presence or absence without adjustment for duration. All statistical analysis was performed using MedCalc (v11.5.0, MedCalc Software, Mariakerke, Belgium).
Long-term outcome for the NCCN high-risk cohort
Depending on the definition used for the 718 patients, high-risk groups constituted 11–33% of the cohort ( Table 1 ), with a median follow-up of 69 months. High-risk disease based on the NCCN risk grouping was present in 234 (33%) of patients. In these men, 145 (62%) had one high-risk factor, 63 (27%) had two risk factors, and 26 (11%) had all three risk factors. Both pelvic RT (91%) and ADT (79%) were commonly used in high-risk patients.
|Clinical features||Overall||NCCN high-risk|
|No. of patients||718||234|
|Age (y)||69 (63–74)||71 (64–75)|
|Clinical T stage (AJCC v. 6)|
|PSA level (ng/mL)|
|Median (IQR)||8.4 (5.6–14.0)||20.3 (9.7–30.7)|
|Any Gleason Pattern 5||11%||33%|
|Positive lymph nodes||1%||3%|
|Number of NCCN high-risk features|
5-year progression-free probability
|Median (IQR)||84% (69–90)||60% (44–79)|
|RT dose (Gy)|
|Median (IQR)||77.0 (75.8–77.8)||77.0 (76.2–78.3)|
|Duration (mo) (median, IQR)||7.0 (6.0–24.4)||21.2 (6.1–26.7)|
Abbreviations: NCCN = National Comprehensive Cancer Network; AJCC = American Joint Committee on Cancer; PSA = prostate-specific antigen; IQR = interquartile range; RT = radiation therapy; ADT = androgen deprivation therapy; PFP = Progression Free Probability.
Association of risk factors with recurrence, distant metastases, and death from prostate cancer
The 8-year risk of BF, metastasis, and PCSM for the whole cohort was 29%, 14%, and 8%, respectively. The absence of any NCCN high-risk criteria carried the lowest risk of BF, metastasis, and PCSM (19%, 8%, and 2% at 8 years, respectively). Each of the seven high-risk criteria was associated with a higher rate of BF ( Table 2 ) (HR 2.8–3.9), metastasis ( Table 3 ) (HR 3.7–6.6), and PCSM ( Table 4 ) (HR 3.7–16.2). Higher GS was the most adverse prognostic factor, with GS 8 to 10 yielding a 2.8-fold increased risk of BF, a 3.7-fold increased risk of metastasis, and a sevenfold increased risk of PCSM. The presence of Gleason Pattern 5 carried an even higher risk, with 3.2-fold increased risk of BF, 6.6-fold risk of metastasis, and 16.2-fold increased risk for PCSM. Notably, all of the high-risk definitions carried a similar risk of BF (HR 2.8–3.9), whereas those based on GS carried a higher risk of metastasis and PCSM than definitions based on either T stage or PSA.
|Risk factor||Yes/No||n||Biochemical failure|
|5-year||8-year||p value ∗||HR (95% CI)|
|Overall||—||718||18% (16–20)||29% (27–31)||—||—|
|Gleason score 8–10||Yes||145||37% (32–42)||44% (38–52)||p < 0.0001||2.8 (1.8–4.3)|
|No||573||13% (11–15)||25% (22–28)|
|Gleason Pattern 5||Yes||76||39% (32–46)||50% (42–58)||p < 0.0001||3.2 (1.6-6.1)|
|No||639||15% (13–17)||27% (24–30)|
|PSA >20 ng/mL||Yes||121||40% (25–45)||57% (51–63)||p < 0.0001||3.5 (2.2–5.4)|
|No||595||13% (11–15)||22% (19–25)|
|cT3/4||Yes||81||45% (39–51)||62% (55–69)||p < 0.0001||3.9 (2.3–6.5)|
|No||636||14% (12–16)||23% (20–26)|
|NCCN high risk||Yes||234||33% (29–37)||47% (43–51)||p < 0.0001||3.5 (2.4–4.9)|
|No||483||10% (8–12)||19% (16–22)|
|Number of NCCN high-risk features||1||145||20% (16–24)||33% (28–38)||p = 0.0001 †||2.3 (1.5–3.4)|
|2||64||47% (40–54)||59% (51–67)||p < 0.0001 †||4.8 (3.1–7.3)|
|3||25||64% (54–74)||79% (69–89)||p < 0.0001 †||9.5 (5.5–16.2)|
|Nomogram 5-y PFP ≤50%||Yes||81||42% (36–48)||59% (52–66)||p < 0.0001||3.5 (2.1–6.1)|
|No||635||15% (13–17)||24% (21–27)|
∗ Log-rank test.
† Cox proportional hazards model.
Abbreviations: HR = hazard ration; CI = confidence interval; PSA = prostate-specific antigen; NCCN = National Comprehensive Cancer Network; PFP = Progression Free Probability at 5-years by Kattan Nomogram.
|5-year||8-year||p value ∗||HR (95% CI)|
|Overall||—||718||8% (6–9)||14% (12–16)||—||—|
|Gleason 8–10||Yes||145||22% (20–24)||24% (20–28)||p < 0.0001||3.7 (2.7–7.0)|
|No||573||4% (4–5)||11% (9–13)|
|Gleason Pattern 5||Yes||76||35% (28–42)||39% (31–47)||p < 0.0001||6.6 (2.5–17)|
|No||639||5% (4–6)||11% (9–13)|
|PSA >20 ng/mL||Yes||121||22% (20–24)||32% (27–37)||p < 0.0001||3.8 (2.1–7.0)|
|No||595||5% (4–6)||9% (7–11)|
|cT3/4||Yes||81||31% (26–36)||36% (29–40)||p < 0.0001||4.4 (2.2–8.8)|
|No||636||5% (4–6)||11% (9–13)|
|NCCN high risk||Yes||234||18% (15–21)||25% (21–29)||p < 0.0001||4.1 (2.5–6.7)|
|No||483||3% (2–4)||8% (6–10)|
|Number of NCCN high-risk features||1||145||9% (6–12)||17% (12–22)||p < 0.004 †||2.5 (1.4–4.6)|
|2||64||25% (19–31)||29% (22–36)||p < 0.0001 †||4.4 (2.3-8.0)|
|3||25||51% (40–62)||58% (47–69)||p < 0.0001 †||17.4 (8.9–34)|
|Nomogram 5-year PFP ≤50%||Yes||81||30% (26–35)||37% (31–43)||p < 0.0001||4.9 (2.4–10.1)|
|No||635||5% (4–6)||10% (8–12)|
∗ Log-rank test.
† Cox proportional hazards model.
Abbreviations: HR = hazard ration; CI = confidence interval; PSA = prostate-specific antigen; NCCN = National Comprehensive Cancer Network; PFP = Progression Free Probability at 5-years .
|Risk factor||Yes/No||n||Prostate cancer death|
|5-year||8-year||p value ∗||HR (95% CI)|
|Overall||—||718||3% (2–4)||8% (6–10)||—||—|
|Gleason score 8–10||Yes||145||13% (10–16)||21% (16–26)||p < 0.0001||7.0 (2.9–16.8)|
|No||573||0.5% (0.2–0.9)||5% (3–7)|
|Gleason Pattern 5||Yes||76||25% (19–31)||45% (36–54)||p < 0.0001||16.2 (4.3–61)|
|No||639||1% (0.5–1.5)||4% (3–5)|
|PSA >20 ng/mL||Yes||121||6% (3–9)||19% (14–24)||p < 0.0001||3.8 (1.7–8.4)|
|No||595||2% (1–3)||4% (3–5)|
|cT3/4||Yes||81||10% (6–14)||24% (18–30)||p < 0.0001||3.7 (1.6–8.5)|
|No||636||2% (1–3)||5% (4–6)|
|NCCN high risk||Yes||234||8% (6–10)||18% (14–22)||p < 0.0001||6.9 (3.5–13.4)|
|No||483||0.4% (0.1–0.8)||2% (1–3)|
|Number of NCCN high-risk features||1||145||5% (3–7)||12% (8–16)||p < 0.002||4.8 (1.8–12.6)|
|2||64||8% (4–12)||20% (14–26)||p < 0.0002||6.7 (2.5–18.3)|
|3||25||19% (10–28)||39% (27–51)||p < 0.0001||22.8 (8.5–61)|
|Nomogram 5-year PFP ≤50%||Yes||81||11% (7–15)||29% (22–36)||p < 0.0001||4.8 (2.0–11.3)|
|No||635||2% (1–3)||4% (3–5)|
∗ Log-rank test.
Abbreviations: HR = hazard ration; CI = confidence interval; PSA = prostate-specific antigen; NCCN = National Comprehensive Cancer Network.; PFP = Progression Free Probability at 5-years
For NCCN high-risk patients, the 8-year rates of BF, metastasis, and PCSM were 47%, 25%, and 18%, respectively. Cox regression analysis revealed that an increasing number of high-risk features was associated with worse outcome. The 8-year BF was 33% for one risk factor, 59% with two factors, and 79% with all three factors. Metastasis followed a similar trend, with metastasis 17% for one risk factor, 29% for two factors, and 58% for three factors. In comparison, PCSM occurred in 12%, 20%, and 39% for one, two, and three risk factors, respectively.
Rates of adjuvant and salvage ADT
Given that adjuvant ADT influences outcomes (6), (7), (8), (9), and (10), the rate of ADT use was assessed based on the high-risk definitions ( Table 5 ). All definitions were associated with a greater likelihood of patients receiving adjuvant ADT, which was 79% in NCCN high-risk patients (median, 21 months) compared to 22% of patients not meeting NCCN high-risk criteria (median, 6 months). Among the various high-risk definitions, patients with GS 8 to 10 were most likely to receive adjuvant ADT (87%) for a median 24 months. By contrast, those with either cT3–4 disease or a Kattan recurrence nomogram <50% were least likely to have received adjuvant ADT (65% each) for a median 24 months. When analyzed by the number of risk factors, patients with one NCCN high-risk factor were more likely to receive a shorter course of adjuvant ADT (median, 10 months) than were patients with two or three high-risk features (median, 24 and 26 months, respectively).
|Risk factor||Yes/No||n||Adjuvant ADT||Salvage ADT|
|Percentage ∗||Months †
|If BF % salvage ADT ‡||10-year rate §||HR (95% CI) §|
|Overall||—||718||40%||6.9 (6.0–24.4)||60%||21% (18–24)||—|
|Gleason score 8–10||Yes||145||87%||24.3 (6.3–27.7)||81%||49% (40–58)||3.9 (2.2–7.0)|
|No||573||29%||6.1 (5.0–7.4)||53%||15% (12–18)|
|Gleason Pattern 5||Yes||76||88%||24.4 (7.2–27.7)||92%||47% (39–55)||5.2 (2.3–12.1)|
|No||639||35%||6.3 (5.7–21.2)||56%||18% (15–21)|
|PSA >20 ng/mL||Yes||121||74%||10.8 (6.1–26.5)||78%||47% (41–53)||5.1 (2.9–9.0)|
|No||595||34%||6.3 (5.7–23.6)||53%||15% (12–18)|
|cT3/4||Yes||81||65%||23.5 (6.1–29.5)||84%||62% (54–70)||6.3 (3.2–12.3)|
|No||636||37%||6.3 (6.0–23.4)||52%||13% (11–15)|
|NCCN high risk||Yes||234||79%||21.2 (6.1–26.7)||76%||42% (36–48)||6.2 (3.9–9.8)|
|Number of NCCN high-risk features||1||145||82%||9.5 (6.1–25.2)||61%||22% (17–27)||3.3 (1.9–5.9)|
|2||64||73%||24.0 (6.3–26.8)||87%||63% (54–72)||10.0 (5.8–17.2)|
|3||25||76%||26.2 (6.2–30.6)||88%||74% (62–86)||18.0 (9.3–35)|
|Nomogram 5-year PFP ≤50%||Yes||81||65%||24.4 (6.3–28.8)||85%||58% (49–67)||5.6 (2.8–11.2)|
|No||635||37%||6.4 (6.0–24.4)||53%||15% (12–18)|
∗ Chi-square test used to compare the rate of neoadjuvant/adjuvant ADT use based on presence or absence of the indicated high-risk feature (all with p values <0.0001).
† One-way analysis of variance used to compare the duration of neoadjuvant ADT based on presence or absence of the indicated high-risk feature (all with p values <0.03).
‡ Chi-square test used to compare the likelihood of receiving salvage ADT once the patient had experienced biochemical failure based on the presence or absence of the indicated high-risk feature (all with p values <0.004).
§ For the bivariate definitions of high-risk disease, the log-rank test and Kaplan-Meier analysis were used to estimate the 10-year rates of salvage ADT use (independent of failure status) and hazard ratios for all patients with or without the indicated high-risk feature. For the multipart definition of Number of NCCN High-Risk Features, the hazard ratios are from Cox proportional hazards regression (all with p values <0.0001).
Abbreviations: BF = biochemical failure; ADT = androgen deprivation therapy; HR = hazard ration; CI = confidence interval; PSA = prostate-specific antigen; NCCN = National Comprehensive Cancer Network; PFP = Progression Free Probability at 5-years.
After BF, 60% of patients received salvage ADT, and the presence of high-risk disease by any definition increased the likelihood of salvage ADT after recurrence, which ranged between 61% and 92% of patients defined as high-risk, compared to only 42% of patients without any high-risk features. The 10-year rate of salvage ADT was 21%, which was higher in those meeting any definition of high-risk disease (42–62%) than in those without any high-risk features (10%). The incidence of salvage ADT also increased as the number of risk factors increased (p <0.0001), with a 10-year rate of salvage ADT of 10% in patients without high-risk features, 22% in patients with one risk factor, 63% in patients with two risk factors, and 74% in patients with three risk factors ( Fig. 1 A). Given that currently the majority of high-risk patients are classified via GS 8 to 10 disease, we assessed the rate of salvage ADT in these patients. At 10 years, salvage ADT was used in 16% of patients with GS 8 to 10 disease only, whereas in patients with both GS 8 to 10 and additional high-risk features, this rate was 69–74% (p < 0.0001) ( Fig. 1 B).
Association of adjuvant ADT with recurrence, distant metastases, and death from prostate cancer
The influence of adjuvant ADT on recurrence, metastasis, and PCSM was evaluated for each of the definitions ( Table 6 ). The Kattan nomogram was not assessed because it already incorporates the use of ADT. For those with GS 8 to 10 disease, there was a 50% reduction in the relative risk of BF for those treated with ADT as compared to those treated with RT alone (p < 0.03, HR 0.48), and for Gleason Pattern 5 there was an even larger impact of ADT use (p = 0.001, HR 0.25) on BF. However, none of the other definitions of high-risk disease demonstrated a significant influence of ADT use on BF. By contrast, for metastasis, five of the six definitions demonstrated significant reductions in the risk of metastasis, with 8-year rates of metastasis of 41–75% for EBRT alone and 20–35% for EBRT with adjuvant ADT (HR 0.19–0.38, all p values <0.005). The one exception was the cT3–4 model (HR 0.54, p < 0.1). In addition, four of six definitions demonstrated a reduction in death secondary to prostate cancer with ADT, with rates at 8 years between 30% and 69% without ADT and between 12% and 37% with adjuvant ADT use (HR 0.38–0.50, all p values <0.05). Again, the cT3–4 model did not demonstrate a statistically significant reduction in PCSM with the use of ADT (HR 0.68, p > 0.3), whereas the model evaluating the number of NCCN high-risk features had a trend to improvement with ADT use (HR 0.50, p < 0.06).
|Risk feature||Adjuvant ADT||n||Biochemical failure||Metastasis||Prostate cancer death|
HR (95% CI)
HR (95% CI)
HR (95% CI)
|Gleason 8–10 ∗||Yes||126||34% (29–39)||p < 0.03||20% (16–24)||p < 0.001||18% (13–23%)||p < 0.04|
|No||19||56% (44–68)||0.48 (0.20–1.1)||51% (39–63)||0.31 (0.11–0.89)||37% (25–49)||0.41 (0.14–1.2)|
|Gleason Pattern 5 ∗||Yes||68||41% (32–50)||p = 0.001||35% (27–43)||p < 0.0001||40% (30–50)||p < 0.04|
|No||8||75% (60–90)||0.25 (0.05–1.1)||75% (60–90)||0.19 (0.04–1.0)||69% (51–87)||0.38 (0.11–1.3)|
|PSA >20 ng/mL ∗||Yes||90||40% (34–46)||p = 0.19||27% (20–34)||p < 0.005||12% (7–17)||p < 0.04|
|No||31||43% (34–52)||0.70 (0.39–1.3)||46% (36–56)||0.38 (0.17–0.85)||34% (24–42)||0.43 (0.17–1.0)|
|cT3/4 ∗||Yes||53||43% (35–51)||p = 0.33||31% (23–39)||p < 0.1||19% (12–26)||p > 0.3|
|No||27||49% (30–48)||0.75 (0.40–1.4)||46% (36–56)||0.54 (0.25–1.2)||32% (22–42)||0.68 (0.26–1.7)|
|NCCN high risk ∗||Yes||185||31% (27–35)||p = 0.10||20% (15–25)||p < 0.0004||13% (9–17)||p < 0.04|
|No||49||39% (32–46)||0.68 (0.41–1.1)||41% (33–39)||0.38 (0.19–0.74)||30% (22–38)||0.50 (0.23–1.1)|
|Biochemical failure||Metastasis||Prostate cancer death|
|p value||HR (95% CI)||p value||HR (95% CI)||p value||HR (95% CI)|
|Number of NCCN high-risk features †||Per point||p < 0.0001||2.1 (1.6–2.8)||p < 0.0001||2.6 (1.8–3.9)||p < 0.003||2.1 (1.3–3.5)|
|ADT||p = 0.07||0.65 (0.41–1.0)||p < 0.0004||0.35 (0.31–0.86)||p < 0.06||0.50 (0.24–1.1)|
∗ Log-rank test used to compare influence of ADT use in patients stratified as high risk.
† Cox-proportional hazards models used to evaluate the influence of the use of ADT on the multipart high-risk model Number of NCCN High-Risk Features.
Abbreviations: ADT = androgen deprivation therapy; SEM = standard equivalent of the mean; HR = hazard ration; CI = confidence interval; PSA = prostate-specific antigen; NCCN = National Comprehensive Cancer Network.
The rates of clinical events were evaluated as a function of the number of NCCN high-risk features in patients with and without adjuvant ADT ( Fig. 2 and Table E1 ). In those without any high-risk features, there was no apparent improvement in any endpoint with the addition of ADT ( Table E1 ). However, with or without adjuvant ADT use, an increasing number of high-risk features was associated with worse clinical outcome for all endpoints. For patients treated with dose-escalated RT without ADT, the 8-year FFBF ranged from 62% with one high-risk feature to 17% with all three features (p < 0.0001) ( Fig. 2 A). The use of ADT showed only a modest trend toward improved FFBF ( Fig. 2 B). When patients were treated with EBRT alone, FFM ranged from 67% with one high-risk feature to 18% with all three high-risk features (p < 0.0001) ( Fig. 2 C). With adjuvant ADT, the 8-year FFM ranged from 88% with one high-risk feature to 49% with all three features (p < 0.0001) ( Fig. 2 D). Furthermore, the 8-year cause-specific survival for those treated without ADT was 77% in patients with one high-risk feature and 50% in those with all three features (p < 0.0001) ( Fig. 2 E); whereas in those treated with EBRT and adjuvant ADT, the 10-year cause-specific survival was 86% with one high-risk feature and 65% with all three (p < 0.003) ( Fig. 2 F).
Men with GS 8 to 10 disease now constitute the majority of the high-risk population, and these overall patterns also held true for this subgroup. When treated with combined ADT and EBRT, those with GS 8 to 10 without other high-risk features (i.e., PSA <20 ng/mL and cT1–T2c disease, n = 74) had 8-year FFBF of 74% (SEM 67–81), FFM of 93% (SEM 90–96), and cause-specific survival of 92% (SEM 88–96), with salvage ADT used in 16% (SEM 9–23).
This study demonstrates that commonly used definitions of high-risk prostate cancer are associated with higher rates of adjuvant and salvage ADT use, BF, distant metastasis, and death from prostate cancer after dose-escalated EBRT. Our results also confirm those of a large surgical series (21) indicating that biopsy GS ≥8, pretreatment PSA level >20 ng/mL, and clinical T3–4 disease are predictors of higher rates of secondary cancer therapy, distant metastasis, and prostate cancer death. Compared with non–high-risk patients, those meeting any of the individual high-risk criteria had a fourfold to sixfold increased likelihood of receiving salvage ADT, a fourfold to sevenfold higher probability of metastasis, and a fourfold to 16-fold higher risk of PCSM. Each high-risk definition carried a similar risk of biochemical failure (2.8- to 3.9-fold), but the definitions based on GS had higher risk of metastasis and death from prostate cancer.
The identification of subgroups of high-risk patients who have a poor response to conventional treatment with EBRT and ADT is critical. The NCCN stratification scheme yields a broad continuum of disease. The presence of one NCCN high-risk feature increased the absolute rate of PCSM at 8 years by 10% (from 2% to 12%), and the presence of three NCCN high-risk features further increased this risk by another 27% (from 12% to 39%). These results support the reports by Nguyen et al. (17) and Zelefsky et al. (22) incorporating the number of risk factors, thus demonstrating that even in the context of dose-escalated RT, the number of risk factors still predicts for worse outcome.
After high-dose EBRT, our results indicate that adjuvant ADT only yielded improvements in BF for patients with elevated GS. By contrast, significant reductions in the risk of both metastasis and PCSM where identified for patients meeting NCCN high-risk criteria or those with either GS ≥8 or PSA >20 ng/mL, with hazard ratios ranging from 0.2 to 0.4 and from 0.4 to 0.5, respectively. The discrepancy between outcomes with ADT use based on BF and both metastasis and PCSM raises caution about the strength of PSA as a surrogate for clinical outcome. For NCCN high-risk patients treated with dose-escalated EBRT and ADT, the 8-year rates of metastasis (20%) and PCSM (13%) seem to be at least comparable to the outcomes previously reported for combined EBRT and ADT (7), (9), and (23). For instance, Zelefsky et al. (23) reported slightly higher rates of both metastasis (24%) and PCSM (19%) when measured at 7 years in their NCCN-defined high-risk cohort treated with dose-escalated EBRT to the prostate only (without pelvic RT) and less frequent ADT (delivered at most for 1 year).
Furthermore, our results may underestimate the benefit of ADT when combined with EBRT, for although 79% (186/234) of NCCN high-risk patients were treated with ADT, only 55% (103/186) received ADT for >1 year (median, 26 months), whereas 45% (83/186) received ADT for <1 year (median, 6 months). Both the RTOG and the European Organization for the Research and Treatment of Cancer have previously demonstrated that long-term ADT provides greater benefit than short-term ADT when combined with conventional-dose EBRT (9) and (10). A recent meta-analysis of 4,128 patients treated with conventional-dose EBRT on four randomized RTOG trials (24) demonstrated that short-term ATD yielded significant reductions in both metastasis (HR 0.8, p < 0.007) and PCSM (HR 0.8, p < 0.05), and a greater benefit was observed with long-term ADT (HR for metastasis 0.5, p < 0.0001; HR for PCSM 0.5, p < 0.0001). Furthermore, long-term ADT was associated with a reduction in all-cause mortality (HR 0.7, p < 0.0001), but STAD was not (HR 0.9, p > 0.2).
Although the benefit of ADT in the setting of dose-escalated RT has not been examined prospectively, our results confirm those of randomized trials using conventional-dose RT. For instance, in patients with GS 8 to 10 disease (including 26% with two and 17% with three NCCN high-risk features), we observed a reduction in the 8-year rate of metastasis from 51% (SEM 39–63) in those treated with RT alone to 26% (SEM 18–34) in those treated with STAD and 16% (SEM 11–21) in those treated with long-term ADT (p < 0.002). A similar pattern was seen for PCSM, with 8-year rates of 37% (SEM 25–49) without ADT, 25% (SEM 17–33) with short-term ADT, and 15% (SEM 8–22) with long-term ADT (p < 0.06).
The optimal local therapy for high-risk prostate cancer patients is subject to debate (25) and (26). Recent retrospective surgical series have claimed comparable or superior outcomes in patients treated with radical prostatectomy compared with EBRT with or without ADT (21), (26), and (27). These comparisons, however, are complicated by the heterogeneous distribution of patients within the high-risk classification, patient selection for different primary treatments, varying use and duration of ADT in RT cohorts, differing salvage regimens, and the inherent bias in comparing treatment results across retrospective series. Not surprisingly, pretreatment risk factors are typically greater in the EBRT series. Arcangeli et al. (25) demonstrated that when the pretreatment prognostic factors are properly balanced in high-risk patients and uniform ADT is used with adequate RT doses, the BF rate is significantly better for EBRT than for radical prostatectomy, even with the use of adjuvant RT after radical prostatectomy. Furthermore, the often-cited advantage of radical prostatectomy is that it is a single treatment modality and spares men ADT. However, in reports of radical prostatectomy for high-risk prostate cancer, the use of salvage ADT after radical prostatectomy may be as high as 50–80% with EBRT (either adjuvant or salvage) used in 25–40% of patients (21), (27), and (28). These are higher than the rates of salvage ADT we observed here for patients treated with EBRT and adjuvant ADT, particularly for patients without multiple high-risk features, who may be most similar to those undergoing radical prostatectomy. Therefore, multimodality management is common in prostate cancer patients with high-risk disease, and the role of each component of treatment must be critically appraised.
At present, Level I evidence from Phase III trials for locally advanced and high-risk prostate cancer support an overall survival advantage when LTAD is added to conventional-dose RT delivered to the prostate and pelvic lymph nodes (6), (7), (9), and (10), with a similar overall survival benefit observed with the addition of RT to long-term ADT (29) . Studies also support adjuvant EBRT after RP with high-risk pathologic features in decreasing recurrence and metastasis (30), (31), and (32) and in improving both PCSM and overall survival (30) . To date, however, no prospective studies have assessed planned surgical resection for high-risk patients with subsequent EBRT and ADT as compared to EBRT and long-term ADT up front. Indeed, many patients taking part in the trials of adjuvant EBRT after RP would not have been classified as having either high-risk or locally advanced disease before surgery (30), (31), and (32). Nevertheless, well-performed retrospective studies of EBRT using treatments as supported by the randomized trials would suggest that the added benefit of surgical resection followed by both EBRT and ADT is at best small and is associated with both increased morbidity and financial cost.
There are several limitations to this report. First, it is retrospective in nature and subject to the biases of treatment selection instead of random allocation. Second, there was a close correlation between the use of ADT and pelvic RT, which partially confounds this analysis. However, the stronger association with ADT in improving clinical outcome from randomized trials than with pelvic RT would support ADT use as the predominant factor; however, one cannot rule out the potential benefit of pelvic RT in patients with high-risk prostate cancer (here used in 91% of such patients). Third, although distant metastasis and PCSM are clinically meaningful endpoints, a lack of uniformity in the administration and timing of adjuvant and salvage ADT may have rendered our estimates of metastasis-free survival and cause-specific survival inaccurate. Newer understanding about the timing of PSA failure (33) and about PSA kinetics at the time of failure (34) may help define the appropriate timing of salvage therapy.
These results demonstrate a significant improvement in metastasis-free survival and cause-specific survival with the addition of adjuvant ADT to dose-escalated EBRT across multiple definitions of high-risk disease. Although the majority of high-risk patients remain free of salvage hormonal therapy, metastasis, and death from prostate cancer at 8 years after combined EBRT and ADT, there are subsets of patients who continue to be at significant risk for metastasis and PCSM. These men, characterized by the presence of multiple high-risk features or the presence of Gleason Pattern 5 disease, should be considered for clinical trials investigating novel local and/or systemic therapies.
- 1 A.V. D’Amico, R. Whittington, S.B. Malkowicz, et al. Biochemical outcome after radical prostatectomy, external beam radiation therapy, or interstitial radiation therapy for clinically localized prostate cancer. JAMA. 1998;280:969-974
- 2 A.W. Partin, M.W. Kattan, E.N. Subong, et al. Combination of prostate-specific antigen, clinical stage, and Gleason score to predict pathological stage of localized prostate cancer: A multi-institutional update. JAMA. 1997;277:1445-1451
- 3 M. Kattan, M. Zelefsky, P. Kupelian, et al. Pretreatment nomogram for predicting the outcome of three-dimensional conformal radiotherapy in prostate cancer. J Clin Oncol. 2000;18:3352-3359
- 4 J. Mohler, R.R. Bahnson, B. Boston, et al. NCCN clinical practice guidelines in oncology: Prostate cancer. J Natl Compr Canc Netw. 2010;8:162-200
- 5 S.J. Freedland, E.B. Humphreys, L.A. Mangold, et al. Risk of prostate cancer-specific mortality following biochemical recurrence after radical prostatectomy. JAMA. 2005;294:433-439
- 6 M.V. Pilepich, K. Winter, C.A. Lawton, et al. Androgen suppression adjuvant to definitive radiotherapy in prostate carcinoma: Long-term results of phase III RTOG 85-31. Int J Radiat Oncol Biol Phys. 2005;61:1285-1290
- 7 M. Bolla, L. Collette, L. Blank, et al. Long-term results with immediate androgen suppression and external irradiation in patients with locally advanced prostate cancer (an EORTC study): A phase III randomised trial. Lancet. 2002;360:103-106
- 8 M.V. Pilepich, K. Winter, M.J. John, et al. Phase III radiation therapy oncology group (RTOG) trial 86-10 of androgen deprivation adjuvant to definitive radiotherapy in locally advanced carcinoma of the prostate. Int J Radiat Oncol Biol Phys. 2001;50:1243-1252
- 9 E.M. Horwitz, K. Bae, G.E. Hanks, et al. Ten-year follow-up of radiation therapy oncology group protocol 92-02: A phase III trial of the duration of elective androgen deprivation in locally advanced prostate cancer. J Clin Oncol. 2008;26:2497-2504
- 10 M. Bolla, T.M. de Reijke, G. Van Tienhoven, et al. Duration of androgen suppression in the treatment of prostate cancer. N Engl J Med. 2009;360:2516-2527
- 11 A.L. Zietman, M.L. DeSilvio, J.D. Slater, et al. Comparison of conventional-dose vs high-dose conformal radiation therapy in clinically localized adenocarcinoma of the prostate: A randomized controlled trial. JAMA. 2005;294:1233-1239
- 12 D.A. Kuban, S.L. Tucker, L. Dong, et al. Long-term results of the M. D. Anderson randomized dose-escalation trial for prostate cancer. Int J Radiat Oncol Biol Phys. 2008;70:67-74
- 13 S.T.H. Peeters, W.D. Heemsbergen, P.C.M. Koper, et al. Dose-response in radiotherapy for localized prostate cancer: Results of the Dutch multicenter randomized phase III trial comparing 68 Gy of radiotherapy with 78 Gy. J Clin Oncol. 2006;24:1990-1996
- 14 D.P. Dearnaley, M.R. Sydes, J.D. Graham, et al. Escalated-dose versus standard-dose conformal radiotherapy in prostate cancer: First results from the MRC RT01 randomised controlled trial. Lancet Oncol. 2007;8:475-487
- 15 M. Gleave, W.K. Kelly. High-risk localized prostate cancer: A case for early chemotherapy. J Clin Oncol. 2005;23:8186-8191
- 16 A. Nanda, M.H. Chen, A.A. Renshaw, et al. Gleason Pattern 5 prostate cancer: Further stratification of patients with high-risk disease and implications for future randomized trials. Int J Radiat Oncol Biol Phys. 2009;74:1419-1423
- 17 P.L. Nguyen, M.H. Chen, W.J. Catalona, et al. Predicting prostate cancer mortality among men with intermediate to high-risk disease and multiple unfavorable risk factors. Int J Radiat Oncol Biol Phys. 2009;73:659-664
- 18 M.W. Kattan, M.J. Zelefsky, P.A. Kupelian, et al. Pretreatment nomogram for predicting the outcome of three-dimensional conformal radiotherapy in prostate cancer. J Clin Oncol. 2000;18:3352-3359
- 19 Z. Symon, K.A. Griffith, P.W. McLaughlin, et al. Dose escalation for localized prostate cancer: Substantial benefit observed with 3D conformal therapy. Int J Radiat Oncol Biol Phys. 2003;57:384-390
- 20 M. Roach 3rd, G. Hanks, H. Thames Jr., et al. Defining biochemical failure following radiotherapy with or without hormonal therapy in men with clinically localized prostate cancer: Recommendations of the RTOG-ASTRO Phoenix Consensus Conference. Int J Radiat Oncol Biol Phys. 2006;65:965-974
- 21 O. Yossepowitch, S.E. Eggener, A.M. Serio, et al. Secondary therapy, metastatic progression, and cancer-specific mortality in men with clinically high-risk prostate cancer treated with radical prostatectomy. Eur Urol. 2008;53:950-959
- 22 M.J. Zelefsky, S.A. Leibel, P.B. Gaudin, et al. Dose escalation with three-dimensional conformal radiation therapy affects the outcome in prostate cancer. Int J Radiat Oncol Biol Phys. 1998;41:491-500
- 23 M.J. Zelefsky, Y. Yamada, Z. Fuks, et al. Long-term results of conformal radiotherapy for prostate cancer: Impact of dose escalation on biochemical tumor control and distant metastases-free survival outcomes. Int J Radiat Oncol Biol Phys. 2008;71:1028-1033
- 24 D.A. Hamstra, K. Bae, M.V. Pilepich, et al. Older age predicts decreased metastasis and prostate cancer-specific death for men treated with radiation therapy: Meta-analysis of Radiation Therapy Oncology Group trials. Int J Radiat Oncol Biol Phys. 2011; In press
- 25 G. Arcangeli, L. Strigari, S. Arcangeli, et al. Retrospective comparison of external beam radiotherapy and radical prostatectomy in high-risk, clinically localized prostate cancer. Int J Radiat Oncol Biol Phys. 2009;75:975-982
- 26 M.J. Zelefsky, J.A. Eastham, A.M. Cronin, et al. Metastasis after radical prostatectomy or external beam radiotherapy for patients with clinically localized prostate cancer: A comparison of clinical cohorts adjusted for case mix. J Clin Oncol. 2010;28:1508-1513
- 27 M. Spahn, C. Weiss, P. Bader, et al. Long-term outcome of patients with high-risk prostate cancer following radical prostatectomy and stage-dependent adjuvant androgen deprivation. Urol Int. 2010;84:164-173
- 28 M. Spahn, S. Joniau, P. Gontero, et al. Outcome predictors of radical prostatectomy in patients with prostate-specific antigen greater than 20 ng/ml: A European multi-institutional study of 712 patients. Eur Urol. 2010;58:1-7 discussion 10–11
- 29 A. Widmark, O. Klepp, A. Solberg, et al. Endocrine treatment, with or without radiotherapy, in locally advanced prostate cancer (SPCG-7/SFUO-3): An open randomised phase III trial. Lancet. 2009;373:301-308
- 30 I.M. Thompson, C.M. Tangen, J. Paradelo, et al. Adjuvant radiotherapy for pathological T3N0M0 prostate cancer significantly reduces risk of metastases and improves survival: Long-term followup of a randomized clinical trial. J Urol. 2009;181:956-962
- 31 T. Wiegel, D. Bottke, U. Steiner, et al. Phase III postoperative adjuvant radiotherapy after radical prostatectomy compared with radical prostatectomy alone in pT3 prostate cancer with postoperative undetectable prostate-specific antigen: ARO 96-02/AUO AP 09/95. J Clin Oncol. 2009;27:2924-2930
- 32 M. Bolla, H. van Poppel, L. Collette, et al. Postoperative radiotherapy after radical prostatectomy: A randomised controlled trial (EORTC trial 22911). Lancet. 2005;366:572-578
- 33 M.K. Buyyounouski, A.L. Hanlon, E.M. Horwitz, et al. Interval to biochemical failure highly prognostic for distant metastasis and prostate cancer-specific mortality after radiotherapy. Int J Radiat Oncol Biol Phys. 2008;70:59-66
- 34 A.K. Lee, L.B. Levy, R. Cheung, et al. Prostate-specific antigen doubling time predicts clinical outcome and survival in prostate cancer patients treated with combined radiation and hormone therapy. Int J Radiat Oncol Biol Phys. 2005;63:456-462
∗ Department of Radiation Oncology, University of Michigan, Ann Arbor, MI
† The Department of Radiation Oncology, Cedars Sinai Medical Center, Los Angeles, CA
‡ Department of Radiology, The Ann Arbor VA Medical Center, Ann Arbor, MI
∗ Reprint requests to: Daniel A. Hamstra, M.D., Ph.D., University of Michigan Health System, 1500 East Medical Center Drive, UH B2 C490, Ann Arbor, MI 48109-5010. Tel: (734) 936-4302; Fax: (734) 763-7370
Conflict of interest: none.
© 2011 Elsevier Inc., All rights reserved.