Frequency and outcomes of co-mutations according to ProMisE classifiers in endometrial cancer
Recommended Citation
Ali-Fehmi R, Toboni M, Ketch P, Krause H, Wu S, Zaiem F, Wallbillich J, Morris R, Gogoi R, Wong T, Kheil M, Oberley MJ, Winer I, Chapel D, Hirst J, Jones N, Thaker PH, Powell M, Herzog T. Frequency and outcomes of co-mutations according to ProMisE classifiers in endometrial cancer. Gynecol Oncol 2024; 190(S1):S170.
Document Type
Conference Proceeding
Publication Date
11-1-2024
Publication Title
Gynecol Oncol
Abstract
Objectives: Proactive Molecular Risk Classifier for Endometrial Cancer (ProMisE) classifies endometrial cancer (EC) into four molecular subtypes: DNA polymerase epsilon (POLE)-mutated, mismatch repair deficient (MMRd), wild-type and mutant p53 (TP53mut). There is a limited understanding of prognosis when tumors have alterations in multiple classifiers. Here in this study, we reported the frequency and outcomes of multi-classifier tumors in addition to high-grade biomarkers (loss of heterozygosity [LOH] and cyclin E1 amplification [CCNE1-amp]). Methods: A total of 5158 EC underwent whole-exome sequencing. MMRd was defined as a complete loss of ≥1 IHC stain (MLH1, MSH2/6, or PMS2). MSI-High (determined from 7000 targeted microsatellite loci) was used as a surrogate for MMRd. TP53mut was defined as any pathogenic or likely pathogenic (PLP) single nucleotide variation (SNV) or an insertion or deletion (indel). POLEmut was defined as PLP mutations in the exonuclease domain. Autosomal chromosomes were split into 552 segments, and the LOH within each segment was calculated (LOH-High [H] ≥ 16 %). CCNE1-amp was defined as ≥6 gene copies. Real-world overall survival (OS) was obtained from insurance claims and calculated from tissue collection to last contact; Kaplan-Meier estimates were calculated for molecularly defined patients. Results: Concurrence between MMRd, TP53mut, and POLEmut was calculated with overlapping subtypes occurring in 4.1 % of cases (MMRd and TP53mut, n = 172 [3.3 %]; MMRd and POLEmut, n = 8 [0.2 %]; TP53mut and POLEmut, n = 29 [0.6 %]) (Table 1). Tumors that were exclusively TP53mut had significantly lower median OS as compared to TP53mut/POLEmut and POLEmut alone (median OS: 30 vs 56 months vs median not reached [MNR], respectively, P = 0.009) (Fig. 1a). The median OS was not significantly different between TP53mut/MMRd, TP53mut, and MMRd (MNR vs 30 vs 40 months, P = 0.02) (Fig. 1b). Nor was there a significant difference in OS for TP53mut versus TP53mut/CCNE1-amp tumors (median OS: 30 [n = 662] vs 65 months [n = 61], P = 0.29) (Fig. 1c). No significant difference was observed between TP53mut, TP53mut/LOH-H, and LOH-H tumors (median OS: 30 vs 29.1 vs 45.5 months; P = 0.27) (Fig. 1d). Conclusions: We report on the co-occurrence of MMR, LOH, TP53, POLE, and CCNE1 alterations in a large cohort of EC. We note that co-occurrence POLEmut/TP53mut favors POLEmut alone in contrast to TP53mut alone and follows the ProMisE algorithm. The co-occurrence of MMRd/TP53mut tumors is not significantly different than either alone. Future work should investigate treatment options for these distinct subtypes. [Formula presented]
Volume
190
Issue
S1
First Page
S170
