Risk factors and predictive modeling for occult endometrial cancer in women with atypical hyperplasia: a retrospective study

Abstract Purpose Atypical endometrial hyperplasia (AEH) is a known precursor to endometrioid endometrial carcinoma. However, occult carcinoma may already be present at diagnosis, complicating surgical planning. Accurate preoperative risk stratification is crucial, especially for guiding the selective use of sentinel lymph node biopsy. This study aimed to identify predictors of occult carcinoma and develop a model to estimate the risk of malignancy. Methods We conducted a retrospective case–control study of 101 women diagnosed with AEH who underwent hysterectomy between 2010 and 2024 at Galilee Medical Center. Clinical, metabolic, and imaging data were extracted. Patients were stratified based on the final pathology into two groups: those with occult carcinoma and those with AEH only. Multivariable logistic regression was employed to identify independent predictors and construct a predictive model. Results Occult endometrial carcinoma was identified in 37 women (36.6%). Women with occult endometrial carcinoma were older and more likely to present with postmenopausal bleeding. Occult carcinoma was more frequently detected after Pipelle biopsy than after hysteroscopy or dilation and curettage (43.2% vs. 17.2%). In multivariable analysis, Pipelle biopsy (OR 4.68), hyperlipidemia (OR 5.86), obesity (OR 2.97), and increasing age (OR 1.07 per year) were independently associated with occult carcinoma. A predictive model estimated individual risk ranging from 5.6% to 95.0% according to accumulation of risk factors. Conclusion Older age, biopsy method, obesity, hyperlipidemia, and bleeding presentation are independently associated with an occult endometrial carcinoma in women with atypical endometrial hyperplasia. The proposed model may support preoperative risk stratification and counseling, but it requires external validation before clinical use, including decisions regarding sentinel lymph node biopsy.