[Comparative study of next generation sequencing and immunohistochemistry on molecular classification of endometrial...

Objective: To investigate the differences in molecular classification of endometrial carcinoma (EC) between various technical methods and to explore molecular classification schemes suitable for Chinese population. Methods: The study used a comprehensive scheme of next generation sequencing (NGS) and immunohistochemistry for molecular classification of 254 EC cases that were obtained at Department of Pathology, Obstetrics and Gynecology Hospital of Fudan University, Shanghai, China from April 2021 to March 2022. According to the recommended threshold of Sanger sequencing which was approximate-20% variant allele fraction (VAF), NGS data were extracted to simulate the results of Sanger sequencing. Results: The 254 EC patients had a mean age of 51 years (range, 24 to 89 years). Combination of POLE (9-14 exons), TP53 total exons and microsatellite instability (MSI) detection was a better single scheme than NGS alone, while combination of MSI fragment analysis and conventional immunohistochemistry was the best solution and seemed best aligned with TCGA data and recent studies. POLE ultramuted type, mismatch repair defect type, TP53 mutant type and non-specific molecular characteristic type accounted for 11.4% (29/254), 31.5% (80/254), 22.4% (57/254) and 34.6% (88/254) of the cases, respectively. If Sanger sequencing was adopted for POLE and TP53 detection, the frequencies of these EC types were 9.1% (23/254), 31.5% (80/254), 12.9% (33/254) and 46.6% (118/254), respectively, with greatly increasing non-specific molecular characteristics cases. If POLE was detected by Sanger sequencing and others by immunohistochemistry, they were 9.1% (23/254), 42.2% (92/218), 13.8% (35/254) and 40.9% (105/254), respectively, with increasing the false positive rates of the mismatch repair defect group. Conclusions: Small and medium-sized NGS panels with MSI detection is a better solution than NGS alone. Sanger sequencing is currently available for POLE mutation detection, which is not sensitive enough for TP53 mutation detection, and seems equivalent to the efficiency of TP53 by immunohistochemistry. Further optimization of small and medium-sized NGS panels covering MSI detection and POLE and TP53 full exons may be the best choice for the future to meet national conditions.