Distinct clinical outcomes according to molecular subgroups in relapsed endometrial carcinoma: A cohort study.

[OBJECTIVE] This study evaluated time to progression and post-recurrence disease-specific survival in molecularly classified endometrial carcinoma to improve understanding of disease biology and factors influencing tumor aggressiveness.

[METHODS] In this retrospective cohort study, immunohistochemistry and polymerase-ϵ (POLE) sequencing were used for molecular classification and determination of estrogen receptor and programmed death-ligand 1 (PD-L1) expression.

[RESULTS] We identified 1146 patients with molecularly classified endometrial carcinoma, of whom 220 (19.2 %) experienced relapse (median follow-up: 65 months). The no specific molecular profile (NSMP) subgroup (n = 60) showed longer time to progression than the mismatch repair deficient (MMRd, n = 95; P = 0.047) and p53-abnormal (p53abn, n = 63; P = 0.009) subgroups. Only 2 POLE ultramutated tumors were present, precluding meaningful comparisons. In the NSMP subgroup, chemotherapy ± radiotherapy was associated with a longer time to progression compared to no adjuvant therapy (hazard ratio 0.13, 95 % confidence interval 0.044-0.37; P < 0.001). Patterns of relapse suggested a tendency toward local relapses in NSMP, regional in MMRd, and distant in p53abn (P = 0.002). Pairwise comparisons of post-recurrence disease-specific survival indicated longer survival for NSMP than for MMRd (P = 0.014) or p53abn (P < 0.001). Within NSMP, post-recurrence disease-specific survival was poorer for high-grade tumors, irrespective of estrogen receptor status, than low-grade tumors, all estrogen receptor-positive. PD-L1 expression was not associated with progression-free or disease-specific survival in molecular subgroup-specific analyses; however, PD-L1 positivity correlated with poorer post-recurrence disease-specific survival in MMRd tumors (P < 0.001).

[CONCLUSION] Our findings support the clinical utility of molecular classification in relapsed endometrial carcinoma and underscore the need to consider both molecular subtype and disease phase when assessing immune-related biomarkers.