Cognition of apalutamide + ADT treatment responders versus non-responders.

Abstract

5085 Background: LIBERTAS (NCT05884398) evaluates intermittent androgen deprivation therapy (ADT) with apalutamide monotherapy in metastatic castration-sensitive prostate cancer (mCSPC). ADT is known to affect cognitive domains such as visuospatial ability and executive function. However, links between cognition and treatment responses remain unclear. LIBERTAS offers an opportunity to examine this relationship in participants with mCSPC. Methods: All 410 participants received apalutamide and continuous ADT during the initial treatment phase, 144 without prior ADT exposure (ADT naive). Participants with PSA <0.2 ng/mL at 6 months were considered responders, otherwise they were non-responders. Cognition was assessed by two tasks from Cambridge Cognition’s CANTAB battery: Paired Associates Learning (PAL; visual memory) and Spatial Working Memory (SWM; executive function) at baseline (C1), 2 months (C3), and 4 months (C5). Outcomes included PAL First Attempt Memory Score (PALFAMS28, higher=better), PAL Total Errors Adjusted (PALTEA28, lower=better), and SWM Between Errors (SWMBE468, lower=better). At each visit, group differences were evaluated using the Mann-Whitney U test. Longitudinal change within and between groups were assessed using linear mixed effect models adjusted for age and testosterone. Results: At all visits, responders had higher PALFAMS28, lower PALTEA28, and lower SWMBE468 than non-responders in the entire sample (Table 1). In ADT-naïve participants, responders (N=100) had higher PALFAMS28 than non-responders (N=44) at C3 (10.48 ± 4.74 vs. 8.59 ± 5.00, p=0.01). SWMBE468 was lower in ADT-naïve responders than ADT-naïve non-responders at C1 (14.25 ± 9.12 vs. 19.30 ± 8.37, p=0.001) and C5 (15.42 ± 8.58 vs. 19.77 ± 11.06, p=0.01). Among all participants and within the ADT-naïve group, only the responders' performance changed over time: PALFAMS28 increased and PALTEA28 decreased from C1 to C5 and SWMBE468 increased from C1 to C3 (p <0.01). No significant time-by-group interaction emerged. Conclusions: Analyzed in retrospect, responders outperformed non-responders in visual memory and executive function at baseline and throughout initial treatment, maintaining a cognitive advantage. Though not tested, we hypothesize early cognitive performance could be considered among a constellation of clinical factors to prognosticate treatment response, support patient stratification and personalize treatment strategies. Clinical trial information: NCT05884398 . Significant differences between responders and non-responders. Measure Visit Responders (N=266): Mean ± SD Non-Responders (N=144): Mean ± SD P value PALFAMS28 C1 9.43 ± 4.53 8.18 ± 4.08 0.005 C3 9.78 ± 4.65 8.07 ± 4.30 <0.001 C5 10.16 ± 4.85 8.47 ± 4.80 0.003 PALTEA28 C1 28.52 ± 17.94 32.92 ± 17.83 0.02 C3 26.74 ± 17.90 31.90 ± 18.83 0.007 C5 25.74 ± 18.40 31.80 ± 20.00 0.004 SWMBE468 C1 15.26 ± 9.00 18.03 ± 7.95 0.008 C3 16.44 ± 9.08 18.73 ± 9.07 0.03 C5 14.79 ± 9.30 17.89 ± 9.83 0.004