Potential effectiveness of local radiotherapy for extending survival and reducing symptomatic local events in patients with de novo metastatic prostate cancer

Abstract Objectives To evaluate the association between the use of local radiotherapy (RT) with the survival of patients with de novo metastatic prostate cancer and symptomatic local events (SLEs). Patients and methods Patients were initially diagnosed with metastatic prostate cancer between 2008 and 2017 at 30 institutes in Japan. Prostate‐specific antigen (PSA) progression‐free survival (PSA‐PFS) under initial androgen deprivation therapy and overall survival (OS) was compared between patients receiving local RT (RT group) and no RT (no‐RT group) by multivariate Cox proportional hazard analyses. The occurrence rate of grade ≥2 SLEs was compared by multivariate logistic regression analyses. Propensity score matching (PSM) analyses were performed to compare PSA‐PFS and OS of the groups in the high and low metastatic burden cohort. Results Two hundred and five (7%) of 2829 patients received RT before PSA progression. Median PSA‐PFS and OS were significantly longer in the RT group than in the no‐RT group and the difference was significant in multivariate analyses (HR = 0.44, 95% CI = 0.33‐0.57 and HR = 0.40, 95% CI = 0.27‐0.60, respectively). The occurrence rate of grade ≥2 SLEs was significantly lower in the RT group (2%) than the no‐RT group (9%) and the difference was significant in multivariate analyses (HR = 0.28, 95% CI = 0.10‐0.76). Using PSM analyses, PSA‐PFS and OS remained significantly different (HR = 0.64, 95% CI = 0.46‐0.89 and HR = 0.47, 95% CI = 0.30‐0.72, respectively), between the RT (n = 182) and the no‐RT (n = 182) groups. The difference in OS was significant in the high metastatic burden cohort (HR = 0.55, 95% CI = 0.37‐0.81). Conclusions Addition of local RT to standard treatment for de novo metastatic prostate cancer patients tends to have the potential to extend survival, even in patients with high metastatic burden, and to reduce SLEs.


| INTRODUC TI ON
Prostate cancer is the second leading cause of cancer-related death in men in western countries 1 and its incidence in Japan is increasing rapidly. 2 Between 10% and 15% of prostate cancer patients present with metastatic disease at first diagnosis and typically receive systemic treatment with androgen deprivation therapy (ADT) combined with abiraterone or docetaxel and other androgen receptor axis-targeted therapies, especially those with a high metastatic burden. Retrospective analyses have noted an association between RT of the primary tumor in patients with metastatic prostate cancer and improved overall survival (OS). [3][4][5] Therefore, local treatment of the primary tumor in patients with metastatic prostate cancer might be more useful than previously appreciated. The HORRAD trial randomized patients with metastatic prostate cancer to ADT, with or without local RT, and found no evidence of an OS benefit. However, the trial raised the possibility that survival may be improved in a subgroup of patients with fewer than five bone metastases. 6 The STAMPEDE trial showed no OS benefit of local RT in all metastatic prostate cancer patients; however, a subgroup analysis showed that RT improved the survival in men with a low metastatic burden. 7 It was reported that the prognosis of prostate cancer patients receiving primary ADT in Japan is significantly better than in western counties. 8 Therefore, the efficacy of local RT for Japanese patients with metastatic prostate cancer might differ from the previously reported results. Some advanced prostate cancer patients experience local symptoms such as hematuria, urinary retention, or ureteral obstruction during long-term treatment. 9 Local treatment might have the potential to prevent these symptomatic local events (SLEs). 10 However, SLEs such as hematuria or urinary frequency occur secondary to the local RT. 11 The objective of our retrospective study was to evaluate the efficacy of local RT for extending the survival time and reducing SLEs in de novo metastatic prostate cancer patients with high and low metastatic burdens at institutes participating in the Japan Urologic Oncology Group (JUOG).

| Study population
This was a retrospective study of patients who were initially diagnosed with metastatic prostate cancer between 2008 and 2017 at 30 university, public, and private hospitals participating in the JUOG.
This study was approved by the institutional review board of each institute. The approval number was O-0403 for Miyazaki University CI = 0.30-0.72, respectively), between the RT (n = 182) and the no-RT (n = 182) groups. The difference in OS was significant in the high metastatic burden cohort (HR = 0.55, 95% CI = 0.37-0.81).
Conclusions: Addition of local RT to standard treatment for de novo metastatic prostate cancer patients tends to have the potential to extend survival, even in patients with high metastatic burden, and to reduce SLEs.

K E Y W O R D S
metastasis, prostate cancer, radiation therapy Hospital. All patients had pathologically proven adenocarcinoma of the prostate, and extra-regional lymph node or distant metastasis was detected by computed tomography (CT) or bone scan at the time of diagnosis. The patient backgrounds and survival data were retrospectively obtained from medical records. In the patients who received local RT during treatment, the date and radiation dose were checked. The treatment start was defined as the date of initial ADT.
The state of prostate-specific antigen (PSA) progression under primary ADT was defined as PSA levels increased by 25% from nadir and higher than 2.0 ng/mL under testosterone levels of <50 ng/dL (1.73 nmol L −1 ). Patients who received radical prostatectomy were excluded. Patients who received RT before PSA progression at a radiation dose of 50 Gy or higher were included in the RT group, and others were included in the no-RT group. The patients were separated into those with a low metastatic burden (extra regional lymph node metastasis or <4 bone metastases without visceral metastasis) and those with a high metastatic burden (≥4 bone metastases or visceral metastasis). SLEs based on the Common Terminology Criteria for Adverse Events (CTCAE) version 5.0 were defined as cystitis, hematuria, urinary frequency, urinary retention, ureteral obstruction, and urinary tract pain. The type, grade, and occurrence date of the SLEs were evaluated based on medical records.

| Statistical analysis
The data are presented as medians and interquartile ranges (IQR).
PSA-progression free survival (PFS) and OS were estimated using the Kaplan-Meier method. Associations of clinicopathological variables including age ≥73 y, PSA ≥160 ng/mL, hemoglobin (Hb) ≤13 g/ dL, lactate dehydrogenase (LDH) ≥200 IU/L, alkaline phosphatase (ALP) ≥300 IU/L, Gleason score (GS) ≥9, prostate volume (PV) ≥50 mL, clinical stage ≥T3b, high metastatic burden, and receiving RT, with PSA-PFS and OS, were assessed using univariate and multivariate Cox proportional hazards models. The association with the occurrence of grade ≥2 SLEs was assessed by univariate and multivariate logistic regression analyses. Baseline characteristics were compared across groups using Fisher's exact and Mann-Whitney tests. To match the patients' background between the groups, propensity score matching (PSM) was performed using a caliper for neighborhood-based estimation using the propensity scores calculated by the logistic regression models using the parameters of age, levels of PSA, Hb, LDH and ALP, GS ≥ 9, and high metastatic burden.
They were significantly different between the groups and correlated with PSA-PFS or OS. After PSM, the baseline characteristics data were presented as mean and standard deviation (SD) and compared using the paired t test. Thereafter, PSA-PFS and OS were compared in the low and high metastatic burden cohorts, and interactions for variables of metastatic burden and the effect of local RT on survival were analyzed using Cox proportional hazards models adjusted for stratification factors with emphasis on metastatic burden. 12 All statistical analyses were performed with EZR (Saitama Medical Center, Jichi Medical University, Saitama, Japan), which is a graphical user interface for R (R Foundation for Statistical Computing, Vienna, Austria). P < .05 was considered statistically significant.

| RE SULTS
The median follow-up duration from the start of initial ADT was 36 months (IQR, 19-61). Among 2823 patients included in this study, 384 (14%) received local RT during the follow-up duration, of whom 205 (7%) received RT at a dose of 50 Gy or more before PSA progression and were included in the RT group. The median radiation dose in the RT group was 70 Gy (IQR, 70-74). The primary treatment was ADT alone in 380 (13%), ADT+bicalutamide in 2365 (84%), ADT+docetaxel in 62 (2%), and ADT+abiraterone in one (0.04%).
In total, 1893 patients (57%) had PSA progression under primary treatment and the median PSA-PFS was 23 months. Between the no-RT and RT groups, age, levels of PSA, LDH, ALP, Hb, the number of patients with a total GS of 9 or higher, and the number with a high metastatic burden were significantly different (Table 1). PSA-PFS was significantly longer in the RT group than in the no-RT group (HR = 0.35, 95% CI = 0.28-0.44, P < .001) ( Figure 1A). In univariate Cox proportional hazard analyses, PSA <160 ng/mL, Hb >13 g/ dL, LDH <200 IU/L, ALP <300 IU/L, GS <9, clinical stage <T3b, low metastatic burden, and RT+ were significantly associated with longer PSA-PFS. In multivariate analyses of these parameters, RT+ was still significantly associated with longer PSA-PFS (HR = 0.44, 95% CI = 0.33-0.57, P < .001) ( Table 2).
To reduce the selection bias between the no-RT and RT groups, PSM was performed. The background of the patients was not significantly different between the no-RT group (n = 182) and the RT group (n = 182) after PSM (Table 4) than in the no-RT group (224/2618, 9%) (P < .001), although the median follow-up was significantly longer in the RT group than in the no-RT group (56 and 34 months, respectively; P < .001). In univariate logistic regression analyses, GS ≥9, PV ≥50 mL, and clinical stage ≥T3b were positively associated with the occurrence of grade ≥2 SLEs, while ALP ≥300 IU/L and RT+ were negatively associated. In multivariate analyses including these parameters, RT+ was still significantly associated with a lower rate of grade ≥2 SLEs (HR = 0.28, 95% CI = 0.10-0.76, P = .013) ( Table 5). After PSM, the occurrence rate of grade ≥2 SLEs was still significantly lower in the RT group (4/182, 2%) than in the no-RT group (26/182, 14%) (P < .001).      18 We previously reported that major SLEs that required therapeutic intervention were observed in 28% of patients who died of prostate cancer. 9 Local RT was effective for the palliation of these SLEs. 19 Therefore, prophylactic local RT might have the potential to reduce the occurrence of these SLEs to prevent local progression. However, local RT might increase some treatment-related SLEs. 11 The STAMPEDE study demonstrated that the number of patients with one or more symptomatic local event did not differ between the RT and the control groups. 7 However, in our study, the occurrence rate of grade ≥2 SLEs was significantly lower in the RT group (4/205, 2%) than in the no-RT group (224/2618, 9%) (P < .001). In a previous retrospective study, local treatment by radical prostatectomy or RT significantly reduced the incidence of SLEs compared with no primary treatment (33% vs 55%, P = .001). 10 The occurrence rate of SLEs in our study was lower than that in the previous studies, probably because the data were retrospectively collected from medical records and the median fol-  patients receiving abiraterone is being tested in the PEACE1 trial. 22 Further RCTs are needed to evaluate the efficacy of local RT under the various therapies for patients, especially those with a high metastatic burden.

TA B L E 4 Background of patients in RT and no-RT groups after prpoensity score matching
In conclusion, our retrospective multi-institutional study indicated that addition of local RT to the standard treatment for metastatic prostate cancer has the potential to reduce SLEs and prolong patient survival. Local RT might be a treatment option for selected patients with metastatic prostate cancer.

ACK N OWLED G M ENTS
We are grateful to the members of the Japan Clinical Oncology Group  .013 * *P < .05.