Improving preoperative diagnosis in endometrial cancer using systematic morphological assessment and a small...

Preoperative histopathological classification determines the primary surgical approach in endometrial carcinoma (EC) patients but has only moderate agreement between preoperative and postoperative diagnosis. The aim of the PIpelle Prospective ENDOmetrial carcinoma (PIPENDO) study is to determine whether histopathological assessment and a small panel of diagnostic biomarkers decreases discrepancies between preoperative and postoperative diagnosis in EC. Preoperative endometrial tissue of 378 included patients with EC was stained with 15 different antibodies. Clinically relevant discrepancies in grade or histological subtype between original preoperative and reviewed postoperative diagnosis were observed in 75 (20%) patients. Highest clinically relevant discrepancy was found in grade 2 ECs (20%), compared to 5% and 14% in respectively grade 1 and 3 endometrioid endometrial carcinomas (EECs). A practical two-biomarker panel with PR and p53 improved diagnostic accuracy (AUC = 0.92; 95%CI = 0.88-0.95) compared to solely morphological evaluation (AUC = 0.86). In preoperative high-grade EC, the diagnostic accuracy of histological subtype was improved by a three-immunohistochemical biomarker panel (PR, IMP3, and L1CAM) (AUC = 0.93; 95%CI = 0.88-0.98) compared to solely morphological evaluation (AUC = 0.81). In conclusion to improve correct preoperative diagnosis in EC, we recommend use of a panel of at least two easily accessible immunohistochemical biomarkers (PR and p53), only in grade 2 ECs. Overall, this will reduce clinically relevant discrepancies in tumor grade and subtype with postoperative diagnosis with 6% (from 20% to 14%). Addition of PR, IMP3, and L1CAM for histological subtyping in high-grade EECs resulted in a further decrease in discrepancies with 8% (from 20% to 12%).