Specific Pathology Features Enrich Selection of Endometrial Carcinomas for POLE Testing

Abstract

Identification of ultramutated/POLE-mutated endometrial carcinomas (POLE M ECs) has important implications given its association with better prognosis. However, POLE mutation testing is not widely available. Our objective was to evaluate POLE M ECs versus POLE wild-type (POLE WT) ECs, within a cohort of consultation cases with features suggestive of an ultramutated phenotype. Consultation cases of EC that had undergone POLE hotspot mutation testing over a 3.5-year period were included. Tumor morphology and immunohistochemistry were reviewed for both groups. Chi-square test and t test were used for statistical analysis. Of 25 consultation cases, 12 harbored a POLE mutation (48%) and 13 were wild-type (52%). Patients with POLE M ECs were younger (59 vs. 71.3 y; P=0.01). Ambiguous histomorphology (5/12 vs. 1/13; P=0.04) and the presence of more than rare bizarre nuclei (8/12 vs. 2/12; P=0.01) differed significantly between POLE M and POLE WT ECs, respectively. In the POLE M group, one case (1/12) demonstrated PMS2 loss, and one (1/12) showed subclonal MLH1/PMS2 loss. Among POLE WT ECs, 3/13 (23%) showed MLH1/PMS2 loss. p53 was subclonally overexpressed in 4/10 POLE M and 1/13 POLE WT cases (P=0.06). Mutant p53 patterns were seen in 1/10 POLE M versus 6/13 of POLE WT ECs, respectively (P=0.06). Within our cohort, the specificity of ambiguous histomorphology, bizarre nuclei, subclonal biomarker expression, and marked tumor-infiltrating lymphocytes for POLE M EC was 83%, 80%, 80%, and 71%, respectively. Where universal POLE testing is not available, these data suggest that morphologic screening (particularly ambiguous histomorphology and the presence of more than rare bizarre nuclei) can be useful for selective enrichment of ECs for POLE testing.