Patients’ clinical characteristics
Patient characteristics and demographics are given in Table 1. Among the 135 study participants with CRPC who had received enzalutamide, 89 were chemo-naive, and 46 were post-chemo patients. The median age at CRPC progression of all patients was 76 years (IQR, 70–79), and the median age of post-chemo patients was 73 years (IQR, 67-78), which was significantly younger than those of chemo-naive patients at 77 years (IQR, 71–81) (p = 0.0002). The median follow-up period for all patients, chemo-naive patients only, and post-chemo patients only were 63 months (IQR, 33–108), 75.5 months (IQR, 38.3–111.8), and 55.0 months (IQR, 30.3–96.8), respectively. All patients underwent ADT combined with first-generation anti-androgen agent bicalutamide administration before CRPC. No patients received up-front androgen-receptor signaling axis-targeting agent (ARAT) or chemotherapy in the hormone-sensitive setting.
The median time to CRPC of all, chemo-naive, and post-chemo patients were 19 (IQR, 10–50), 26.0 (IQR, 11–67), and 12.0 months (IQR, 6.0–28.8), respectively. After the CRPC setting, 17.0% (23/135), 24.7% (22/89), and 2.2% (1/46) of patients were diagnosed with non-metastatic CRPC. The rate of localized therapy, prior-abiraterone acetate use before enzalutamide administration, and bone modifying agents (BMA) treatment with enzalutamide were not significantly different between chemo-naive and post-chemo CRPC groups. The median serum PSA value of post-chemo patients at initiation of enzalutamide (36.7 ng/mL) was significantly higher than that of chemo-naive patients (10.2 ng/mL) (p = 0.0005).
PSA response
The maximum PSA changes during enzalutamide treatment are depicted in a waterfall plot chart (Fig. 1A-C). PSA responses were observed in 80.0% (108/135) of all, 84.3% (75/89) of chemo-naive, and 71.7% (33/46) of post-chemo CRPC patients. There was no statistically significant difference in PSA response rate between the chemo-naive and post-chemo groups (p = 0.085; Fisher’s exact test). On the other hand, a PSA response rate of >= 50% decline from baseline was observed in 55.6% (75/135) of all, 62.9% (56/89) of chemo-naive, and 41.3% (19/46) of post-chemo CRPC patients, respectively. A PSA response rate of > 50% was significantly higher in chemo-naive CRPC patients than post-chemo CRPC patients (p = 0.017).
Causes of treatment discontinuation and adverse events
Among all 135 patients, 75.6% (102/135) discontinued enzalutamide therapy. The causes of discontinuation included PSA progression 82.5% (84/102), radiographic progression 2.0% (2/102), adverse event 2.9% (3/102), progression of other diseases 2.9% (3/102), and deterioration of general condition 9.8% (10/102) (Table 2). Notably, three patients who experienced adverse events during enzalutamide use experienced symptoms of fatigue (grade 2), itchy skin (grade 2), and rash (grade 2). These events improved after stopping enzalutamide treatment.
Progression-free survivals
The median PFS of all CRPC patients was 301 days (95% CI, 239–363) (Fig. 2A). The median PFS of chemo-naive and post-chemo CRPC patients were 362 days (95% CI, 279–444) and 245 days (95% CI, 174–316), respectively, which was statistically significantly different between the two groups (HR 0.48, 95% CI, 0.32–0.74, p < 0.001) (Fig. 2B). There was no significant difference in the median PFS between non-metastatic patients at 392 days (95% CI, 287–497) and metastatic patients at 266 days (95% CI, 204–328) CRPC patient groups (HR 0.61, 95% CI, 0.35-1.06, p = 0.079) (Fig. 2C). The median PFS for patients with prior abiraterone acetate use was 161 days (95% CI, 60.5–261), shorter than the others at 350 days (95% CI, 290–410) (HR 0.56, 95% CI, 0.36–0.83, p = 0.010) (Fig. 2D).
Overall survivals from initiation of enzalutamide
The median OS from initiation of enzalutamide was 790 days (95% CI, 667–913 days) for all CRPC patients (Fig. 3A). The median OS of chemo-naive and post-chemo CRPC patients were 1142 days (95% CI, 579–1704) and 540 days (95% CI, 467–613), respectively, which was statistically significantly different between the two groups (HR 0.47, 95% CI, 0.28–0.79, p = 0.004) (Fig. 3B). There was a significant difference in the median OS between non-metastatic (1422 days (95% CI, 1246 – 1597)) and metastatic (730 days (95% CI, 667–913)) CRPC patient groups (HR 0.36, 95% CI, 0.16–0.86, p = 0.020) (Fig. 3C). Additionally, there was a significant difference in the median OS between patients with prior-abiraterone acetate use at 478 days (95% CI, 375–580) and those without at 869 days (95% CI, 696–1042) patient (HR 0.85, 95% CI, 0.49–1.48, p = 0.56) (Fig. 3D).
Overall survivals after CRPC setting
The OS from the time of CRPC diagnosis was also analyzed. The overall median OS from CRPC diagnosis was 1514 days (95% CI, 1339–1689 days) (Fig. 4A). The median OS of chemo-naive and post-chemo CRPC patients were 1968 days (95% CI, 1191–2754) and 1273 days (95% CI, 984–1562) from CRPC diagnosis, respectively, which was not a statistically significant difference between the two groups (HR 0.70, 95% CI, 0.43–1.13, p = 0.15) (Fig. 4B). The median OS in patients with non-metastatic (not reached) tended to be longer than that of metastatic CRPC patients at 1426 days (95% CI, 1126-1725 days) but did not reach statistical significance (HR 0.44, 95% CI, 0.19–1.01, p = 0.052) (Fig. 4C). The median OS of patients was 1426 days (95% CI, 1160 –1692 days) with prior abiraterone acetate use and 1514 days (95% CI, 1100–1927 days) without prior abiraterone acetate use before enzalutamide treatment, which was also not a statistically significant difference (HR 1.00, 95% CI, 0.60–1.67, p = 0.99) (Fig. 4D).
Risk factors of enzalutamide therapy related to PFS and OS
To search for the variables predicting shorter PFS and OS after initiation of enzalutamide for CRPC patients, we performed Cox proportional hazards analyses (Table 3, 4 and 5). Regarding PFS, time from ADT therapy initiation to CRPC diagnosis < median value (20 months), prior-docetaxel use, and prior-abiraterone acetate use were risk factors identified in univariate analysis. These three were also risk factors identified by multivariate analysis (Table 3). On the other hand, regarding OS from initiation of enzalutamide, time to CRPC < 20 mo., non-metastatic CRPC status, prior-docetaxel use, and prior-abiraterone acetate use were risk factors in univariate analysis, and only time to CRPC < 20 mo. and prior-docetaxel use were risk factors in multivariate analysis (Table 4). Interestingly, in multivariate analysis, only time to CRPC < 20 mo. was associated with unfavorable OS from the time of CRPC diagnosis for the patients treated with enzalutamide (Table 5).