Real-world ICI adoption and molecular landscape characterization in pMMR/MSS recurrent/metastatic endometrial cancer

SITC 2026

Nov 03, 2026
Oncology
Presentation

Malvika Pillai, Michelle Ting-Lin, Victoria L Chiou, Chithra Sangli, Halla Nimeiri, Jay Carlson

Background
Despite demonstrated benefits of immune checkpoint inhibitors (ICIs) across four pivotal phase III trials, patients with mismatch repair proficient/microsatellite stable (pMMR/MSS) endometrial cancer exhibit attenuated and heterogenous benefit. Although ICI-containing regimens are now standard of care regardless of MMR status, no validated biomarker exists to identify which pMMR/MSS patients derive benefit. Furthermore, pMMR/MSS endometrial cancer harbors a heterogeneous molecular landscape ranging from immune-cold (e.g., serous) to more immunogenic (e.g., endometrioid) histologies. Here, we characterize real-world ICI adoption patterns, biomarker testing rates, and the molecular profiles of a large, sequenced cohort.

 

Methods
We conducted a retrospective observational study using the Tempus multimodal de-identified real-world database. All pMMR/MSS-confirmed, metastatic or recurrent endometrial carcinoma with ≥1 curated line of therapy were eligible for this descriptive study. A subset underwent DNA (Tempus xT) or RNA (Tempus xR, whole transcriptome) next-generation sequencing after metastatic diagnosis (sequenced sub-cohort). Treatment patterns, ICI adoption, and clinical biomarker testing rates (HER2, ER, PR) were characterized in the treated cohort. Molecular biomarkers (e.g., TMB, PDL1, somatic variants) were examined in the sequenced sub-cohort.

 

Results
Among 1,272 treated pMMR/MSS patients (53.9% recurrent, 46.1% de novo metastatic; median age 65), histologies included endometrioid (40.1%), serous (33.0%), carcinosarcoma (9.5%), and clear cell carcinoma (5.6%).

 

At first-line (1L, n = 1250), carboplatin/paclitaxel (37.8%) and carboplatin/paclitaxel + pembrolizumab (14.8%) were the most common regimens. Of 1,051 HER2-tested patients (82.6% of cohort, 9.8% positive), 6.4% received a trastuzumab-containing 1L regimen. 690 patients (54.2%) received pembrolizumab, and 130 patients (10.2%) received dostarlimab-containing regimens. Notably, 5.8% received 1L Lenvatinib + pembrolizumab, a regimen NCCN recommends for second-line or later (2L+) in pMMR. 49.8% of 1L & 2L ICI records pre-dated formal approval, predominantly off-label pembrolizumab, suggesting clinical utilization preceded guideline changes. Approximately 45% of patients progressed to ≥2L, where Lenvatinib + pembrolizumab predominated.

 

In the sub-cohort (n = 935), molecular profiles differed by histology (table 1). In patients sequenced pre-treatment (n = 292), PDL1 positivity was 5.1% and TMB-high was 3.4%. Biomarker co-occurrence differed across HER2 categories: HER2-positive patients (n = 74) had higher TP53 mutation rates (87.8%) and lower PTEN (1.4%) and CTNNB1 (0.0%) rates than HER2-low patients (PTEN 25.5%, CTNNB1 12.2%) or HER2-negative patients (PTEN 29.2%, CTNNB1 12.3%).

 

Conclusions
Real-world adoption of ICIs in pMMR/MSS endometrial cancer has rapidly increased and preceded formal regulatory and guideline changes. Biomarker heterogeneity across histologies underscores the lack of validated biomarkers for treatment selection. These real-world insights complement trial data and highlight where biomarker-driven ICI selection is most needed.