Endometrioid ovarian carcinoma landscape: pathological and molecular characterization

Abstract

Endometrioid ovarian cancers (EOvC) are usually managed as serous tumors. In this study, we conducted a comprehensive molecular investigation to uncover the distinct biological characteristics of EOvC. This retrospective multicenter study involved patients from three European centers. We collected clinical data and formalin-fixed paraffin-embedded (FFPE) samples for analysis at the DNA level using panel-based next-generation sequencing and array-comparative genomic hybridization. Additionally, we examined mRNA expression using NanoString nCounter® and protein expression through tissue microarray. We compared EOvC with other ovarian subtypes and uterine endometrioid tumors. Furthermore, we assessed the impact of molecular alterations on patient outcomes, including progression-free survival (PFS) and overall survival (OS). Preliminary analysis of clinical data from 668 patients, including 86 (12.9%) EOvC, revealed more favorable prognosis for EOvC compared with serous ovarian carcinoma (5-year OS of 60% versus 45%; P = 0.001) driven by diagnosis at an earlier stage. Immunohistochemistry and copy number alteration (CNA) profiles of 43 cases with clinical data and FFPE samples available indicated that EOvC protein expression and CNA profiles were more similar to endometrioid endometrial tumors than to serous ovarian carcinomas. EOvC exhibited specific alterations, such as lower rates of PTEN loss, mutations in DNA repair genes, and P53 abnormalities. Survival analysis showed that patients with tumors harboring loss of PTEN expression had worse outcomes (median PFS 19.6 months vs. not reached; P = 0.034). Gene expression profile analysis confirmed that EOvC differed from serous tumors. However, comparison to other rare subtypes of ovarian cancer suggested that the EOvC transcriptomic profile was close to that of ovarian clear cell carcinoma. Downregulation of genes involved in the PI3K pathway and DNA methylation was observed in EOvC. In conclusion, EOvC represents a distinct biological entity and should be regarded as such in the development of specific clinical approaches.

Keywords: endometrioid carcinoma; gene expression profiling; genomic; ovarian cancer; tissue microarray.

Fig. 1

Study flow‐chart. aCGH, comparative genomic hybridization analysis; FFPE, formalin‐fixed paraffin‐embedded; IHC, immunohistochemistry.

Fig. 2

Protein expression and ProMisE classification in endometrioid ovarian cancer (EovC). (A) Protein expression levels in EOvC, high‐grade serous ovarian cancer (SerOvC), and endometrioid uterine carcinoma. (B) Examples of PTEN positive (A) and negative (B) cases identified by tissue microarray analysis (zoom ×5). (C) Kaplan–Meier curves for progression‐free survival according to PTEN expression. (D) ProMisE classification based on immunohistochemistry data. (E) Kaplan–Meier curves for progression‐free survival according to ProMisE classification. MMRd, mismatch repair deficient; NSMP, non‐specific molecular subtype. P‐values are evaluated using the log‐rank test.

Fig. 3

Copy number alterations in endometrioid ovarian cancer (N = 33). (A) Frequency plot of recurrent copy number alterations identified in endometrioid ovarian cancer (EOvC) tumors using the GISTIC algorithm. Frequencies of gains (red) and losses (blue) are plotted as a result of chromosome location. X‐axis: top = log–scale ratio; bottom = q‐values. Green lines represent the threshold for significance. (B) Supervised analysis comparing EOvC samples to high‐grade serous ovarian carcinoma and endometrial tumors from TCGA datasets. Dotted line of the bottom charts comparing EOvC to high‐grade serous ovarian cancer or endometrioid endometrial cancer: threshold of significance associated with a False Discovery Rate < 0.25.

Fig. 4

Comparison of transcriptomic profiles of endometrioid ovarian cancer (EOvC), clear cell ovarian cancer, and serous ovarian cancer (SerOvC). (A) Principal component analysis of ovarian cancers on 747 genes showing that the centroid of the EOvC set is closer to ovarian clear cell carcinoma than to serous ovarian cancer. (B) Principal component analysis including ovarian and endometrial cancer subtypes on 747 genes showing that centroids of ovarian cancer subtypes do not overlap with endometrial tumors and that EOvC are distinct from endometrioid endometrial cancer. (C) Volcano plot of differential mRNA expression between EOvC vs. SerOvC and endometrioid endometrial cancer identified 60 genes (highlighted in yellow) specifically expressed in EOvC (moderated t‐test: P < 5%, q < 10% & |FC| > 1.5×). (D) Ontology analysis of these 60 genes using the Reactome database [46, 47, 48, 50, 51].