Diabetes mellitus and responsiveness of endometrial hyperplasia and early endometrial cancer to conservative treatment

AT-rich interaction domain 1A (ARID1A) is a tumor suppressor protein involved in endometrioid carcinogenesis. The expression of ARID1A may be lost in the premalignant phase. Our aim was to assess ARID1A as: (i) diagnostic marker to differentiate premalignant endometrial hyperplasia (EH) form benign EH; (ii) prognostic marker for the risk of occult cancer in premalignant EH. A systematic review and meta-analysis were performed by searching electronic databases from their inception to October 2018 for all studies assessing ARID1A in EH by immunohistochemistry. Sensitivity, specificity, positive and negative likelihood ratios (LR+ and LR-), and diagnostic odds ratio (DOR) were calculated for both diagnostic and prognostic accuracy. LR+ > 5, LR- < 0.2, DOR > 25 defined good accuracy; LR+ > 10, LR- < 0.1, DOR > 100 defined excellent accuracy. Seven studies with 467 EH were included. As diagnostic marker, ARID1A showed sensitivity = 0.12, specificity = 0.99, LR+ = 4.34, LR- = 0.85, DOR = 5.12. As prognostic marker for occult cancer, ARID1A showed sensitivity = 0.33, specificity = 0.99, LR+ = 20.70, LR- = 0.49, DOR = 49.59. In conclusion, ARID1A loss is highly specific, but little accurate as diagnostic marker of premalignant EH. Instead, ARID1A loss in premalignant EH is an accurate and almost perfectly specific prognostic marker for coexistent cancer.