194 Background: In TALAPRO-2, 1L TALA + ENZA showed significantly prolonged OS and rPFS vs placebo (PBO) + ENZA in pts unselected (cohort 1) and selected for homologous recombination repair gene alterations (HRRm cohort 2). Long-term safety and patient-reported outcomes (PRO) were reported previously (data cutoff: Sep 3, 2024). This post hoc analysis evaluated the longitudinal Q-TWiST benefit of TALA + ENZA vs PBO + ENZA in both cohorts by contextualizing the clinical benefit of extended life and delay in disease progression, accounting for impact on QoL due to toxicity and disease progression, and integrating them into a single value using TALAPRO-2 data. Methods: Using restricted mean survival time, mean OS was partitioned into time experiencing toxicity prior to progression (TOX), time without toxicity or symptoms of disease progression (TWiST), and time in state of disease progression assessed by blinded independent central review (REL). Utility value was calculated from individual EQ-5D-5L responses. A threshold analysis was conducted using a hypothetical range of values typically reported in similar analyses and a range of scores estimated from TALAPRO-2 data. Time in each health state was weighted by utilities then summed to estimate Q-TWiST. Mean between-treatment differences for each health state were calculated, and bootstrap methods used to estimate CIs for means and mean differences. Results: In both cohorts, time in TOX was slightly longer with TALA + ENZA vs PBO + ENZA while time spent in TWiST was much longer with TALA + ENZA (Table). In both cohorts, time in REL was shorter with TALA + ENZA (more time spent without progression) vs PBO + ENZA (Table). OS was numerically longer with TALA + ENZA vs PBO + ENZA for both cohorts (Table). Regardless of the range of utility values explored for each health state, Q-TWiST estimates for both cohorts were numerically longer for TALA + ENZA vs PBO + ENZA (mean differences ranged from 3.1 to 11.7 months). Conclusions: In TALAPRO-2, TALA + ENZA showed longer Q-TWiST vs PBO + ENZA, indicating a net benefit in unselected and HRRm mCRPC pts. Those receiving TALA + ENZA had a greater quality-adjusted time (shorter REL) because pts spent more time without progression in TALA + ENZA than PBO + ENZA. The resulting Q-TWiST benefit further supports use of 1L TALA + ENZA in unselected and HRRm selected mCRPC pts. Clinical trial information: NCT03395197 . Cohort 1 (Unselected) Cohort 2 (HRRm only) Mean (95% CI), months TALA + ENZA (n=402) PBO + ENZA (n=403) Mean difference TALA + ENZA (n=200) PBO + ENZA (n=199) Mean difference TOX 2.1 (1.6, 2.6) 1.4 (0.9, 1.9) 0.6 (-0.1, 1.3) 1.8 (1.3, 2.3) 0.6 (0.3, 0.9) 1.2 (0.6, 1.8) TWiST 34.0 (31.4, 36.6) 27.5 (24.8, 30.2) 6.5 (2.8, 10.3) 30.3 (27.2, 33.4) 19.8 (16.8, 22.7) 10.5 (6.2, 14.8) REL 7.0 (3.6, 10.4) 10.5 (7.0, 13.9) -3.4 (-8.3, 1.4) 8.2 (4.2, 12.3) 13.2 (9.3, 17.1) -5.0 (-10.6, 0.6) OS 43.1 (40.8, 45.3) 39.4 (37.1, 41.6) 3.7 (0.6, 6.9) 40.3 (37.7, 43.0) 33.6 (31.0, 36.2) 6.7 (3.0, 10.4)
Agarwal et al. (Sun,) studied this question.