Abstract

5612 Background: Pembrolizumab and lenvatinib in combination (pembro-lenv) has resulted in improved outcomes compared to standard chemotherapy for second-line treatment of endometrial cancer. Lenvatinib currently is only indicated for microsatellite stable (MSS) tumors as it is associated with significant toxicities, and microsatellite instability high (MSI-H) tumors respond to pembrolizumab alone. We seek to further specify which patients benefit from the addition of lenvatinib through analysis of outcomes by genomic subtypes and immune microenvironments. Methods: 2,532 endometrial cancer patients who received pembro or pembro-lenv were analyzed using NGS (NextSeq, 592 genes or NovaSeq, WES) and RNA (NovaSeq, WTS) (Caris Life Sciences, Phoenix, AZ). Overall survival (OS) was obtained from insurance claims data and calculated from first treatment to last contact. Hazard ratio (HR) was calculated by Cox proportional hazards, with p-value calculated by log-rank test. Responders (R) and non-responders (NR) were determined as those with above and below median survival, respectively. Immune-cell infiltrates calculated by Quantiseq. Statistical significance calculated by Mann-Whitney U test. Results: There were 37 POLE-mt, 724 MSI-H (POLE-wt), 1,222 TP53-mt (POLE-wt and MSS), and 549 TP53-wt (POLE-wt and MSS) patients treated with pembro alone or with lenvatinib. When comparing pembro-lenv vs pembro in the POLE-mt and MSI-H cohorts, there was no significant difference in OS (p=0.70 and p=0.60). There was also no significant difference in the TP53-mt cohort (17.2 vs 15.2 mo; HR 0.91 (95%CI 0.77-1.08), p=0.28). In the TP53-wt cohort, the addition of lenvatinib was associated with longer median OS (31.6 vs 26 mo; HR 0.73 (95%CI 0.54-0.99), p=0.039). Among the TP53 mutated (POLE-wt/MSS) cohort we sought to identify potential molecular differences associated with response in patients who were treated with pembro-lenv compared to pembro alone: ARID1A-mt (17.3% vs 6%, p<0.05) and ERBB3-mt (4.31% vs 0%, p<0.05) were associated with R vs NR in patients treated with pembro-lenv but not pembro alone (ARID1A-mt: 7.7% vs 9.8%, ERBB3-mt: 1.7% vs 2.4%) (p>0.05). ARID1A-mt trended towards improved post-pembro-lenv survival (28.5 vs 15.6 mo; HR 0.91 (95% CI 0.37-1.15), p=0.14) whereas ARID1A-mt was associated with worse post-pembro survival (9.57 vs 15.4 mo; HR 1.44 (95% CI 1.0 -2.06), p=0.048). Conclusions: Among POLE-wt/MSS patients, TP53 wild type patients have longer OS after pembrolizumab with lenvatinib compared to pembrolizumab alone, but in the TP53 mutated cohort, there was no difference. Among TP53-mt patients, ARID1A-mt is associated with improved pembro-lenv survival but not in pembro alone. Our findings suggest a need to readdress when to use lenvatinib in TP53-mt patients and further explore genomic alterations that may promote treatment response to optimize use of this agent in endometrial cancer.