POLE mutations in endometrial carcinoma: Clinical and genomic landscape from a large prospective single-center cohort

Abstract

Background: To date, 11 DNA polymerase epsilon (POLE) pathogenic variants have been declared "hotspot" mutations. Patients with endometrial cancer (EC) characterized by POLE hotspot mutations (POLEmut) have exceptional survival outcomes. Whereas international guidelines encourage deescalation of adjuvant treatment in early-stage POLEmut EC, data regarding safety in POLEmut patients with unfavorable characteristics are still under investigation. On the other hand, the spread of comprehensive genome profiling programs has underscored the need to interpret POLE variants not considered to be hotspots.

Methods: This study provides a comprehensive analysis of 596 sequenced patients with EC. The genomic landscape of POLEmut EC was compared with cases harboring nonhotspot POLE mutations within the exonuclease domain. Additionally, the genomic characteristics of multiple classifiers, as well as those exhibiting unfavorable histopathological and clinical features, were examined.

Results: No significant genomic differences were observed among patients with POLEmut EC when comparing multiple classifiers to not-multiple classifiers or those with unfavorable clinical features. However, the tumor mutational burden differed in both comparisons, whereas the percentage of C>G mutations only differed in the comparison based on clinical features. Specific POLE mutations, even if not considered to be hotspots, have genomic features comparable to POLEmut.

Conclusions: The present findings confirm the absence of significant genomic differences among POLEmut patients regardless of multiple-classifier status or association with high-risk clinical features. Prognostic data will be essential to elucidate the clinical significance of POLE mutations not classified as hotspots that exhibit genomic characteristics similar to those in POLEmut patients.

Keywords: DNA polymerase epsilon (POLE); POLE hotspots; POLE multiclassifier; adjuvant therapy; comprehensive genomic profiling; endometrial cancer; molecular classification.

FIGURE 1

Integrative genomic and clinical characterization of patients in group A. (A) Lollipop plot representing mutation distribution in the POLE gene. Missense mutations and functional domains are reported. Most common mutations were P286R/S and V411L. (B) Oncoplot displaying variants in the most frequently mutated genes and clinical features. Bar chart on the left represents the mean VAF. (C–D) Single nucleotides substitution types. (C) Percentage of single nucleotide substitutions per sample and (D) Box plot illustrating single nucleotide substitutions frequencies for different substitution types. Points outside the whiskers represent outliers. (E) COSMIC mutational signature decomposition. Bar plot displaying the contribution of mutational signatures to each sample. Different colors represent distinct signatures. COSMIC, catalogue of somatic mutations in cancer; ER, estrogen receptor; FIGO, International Federation of Gynecology and Obstetrics; G, grade; LVSI, lymphovascular space invasion; MSI, microsatellite instability; POLE, DNA polymerase epsilon; PR, progesterone receptor; TMB, tumor mutational burden; VAF, variant allele frequency.

FIGURE 2

Comparison and principal component analysis of The Cancer Genome Atlas parameters across group A. Top row, from left to right: Distribution of TMB, Indels, C>A substitutions and T>G substitutions: between high‐risk (red triangles) and low‐risk (blue circles) samples within the POLE group. Bottom left: Principal Component Analysis (PCA) biplot showing the separation of high‐risk and low‐risk samples along the PC1 and PC2 axes. Vectors represent mutational features contributing to sample clustering. Right: Bar plot indicating the percentage of variance explained by the top five principal components (PC1–PC5) (Top row) and loading values of mutational features (C>G, TMB, T>G, C>A, Indels) for PC1 and PC2 (Bottom row). Indel, insertion and deletion variant; PC, principal component; POLE, DNA polymerase epsilon; TMB, tumor mutational burden.

FIGURE 3

Integrative genomic and clinical characterization of patients in group B. (A) Lollipop plot representing mutation distribution in the POLE gene. Missense mutations and functional domains are reported. (B) Oncoplot displaying variants in the most frequently mutated genes and clinical features. Bar chart on the left represents the mean VAF. (C) Chord plot representing co‐mutation patterns. Width of the edges is proportional to co‐occurrence instances. (D–E) Single nucleotides substitution types. (D) Percentage of single nucleotide substitutions per sample and (E) Box plot illustrating single nucleotide substitutions frequencies for different substitution types. Points outside the whiskers represent outliers. (F) COSMIC mutational signature decomposition. Bar plot displaying the contribution of mutational signatures to each sample. Different colors represent distinct signatures. COSMIC, catalogue of somatic mutations in cancer; ER, estrogen receptor; FIGO, International Federation of Gynecology and Obstetrics; G, grade; LVSI, lymphovascular space invasion; MSI, microsatellite instability; POLE, DNA polymerase epsilon; PR, progesterone receptor; TMB, tumor mutational burden; VAF, variant allele frequency.

FIGURE 4

Comparison and principal component analysis of The Cancer Genome Atlas parameters across the four molecular risk groups. (A) Comparison of TMB across molecular subtypes (MMRd, NSMP, p53abn, POLE). (B) Comparison of the proportion of indels and single nucleotide substitution (C>A, C>T, T>G, T>C) across molecular risk groups. Significant differences are highlighted with asterisks. (C) Principal Component Analysis plot showing clustering of samples based on principal components (PC1 and PC2). Samples are colored by molecular risk groups, and representative sample (P1–P8) are highlighted. (D) Bar plot displaying the variance explained by the top five principal components (PC1–PC5). (E) Loading values for mutational features in PCA to PC1 and PC2. Indel, insertion and deletion variant; MMRd, mismatch repair deficient; ns, not significant; NSMP, nonspecific molecular profile; P, patient; PC, principal component; POLE, DNA polymerase epsilon; p53abn, p53 abnormal; TMB, tumor mutational burden.