Objectives Atypical endometrial hyperplasia (AEH) is considered precursor of endometrioid endometrial carcinoma. The 2014 WHO classification divides endometrial hyperplasia into two categories: hyperplasia without atypia and atypical hyperplasia. However, this classification ignores the degree of nuclear atypia. The objective of this study was to show the importance of grading nuclear atypia (low vs high-grade) and find out the risk of developing endometrial carcinoma following a diagnosis of AEH. In addition, we investigated the potential role of genes known to be involved in endometrial carcinogenesis such as ARID1A, PIK3CA, PTEN, KRAS, CTNNB1 and mismatch repair genes.

Methods We reviewed 91 biopsies of AEH from 91 patients who subsequently underwent hysterectomy within 1 year interval. The association between the grade of nuclear atypia at biopsy and findings at hysterectomy was assessed via a Fisher’s exact test. Targeted sequencing was performed in 30 cases.

Results The grade of nuclear atypia at biopsy was highly predictive of the findings at hysterectomy (P=5.0x10–25), with none of the low-grade AEH having a diagnosis of high-grade AEH/carcinoma at hysterectomy, whereas 9 (29%) of the high-grade AEH had high-grade AEH and 22 (71%) FIGO grade-1 carcinoma. None of the genes tested showed a mutational load significantly associated with the degree of nuclear atypia.

Conclusions In AEH is crucial to assess the degree (low or high) of nuclear atypia. Our data strongly support that low-grade AEH is inconsequential, questioning the need of hysterectomy for such patients.