Clinical and ultrasound predictors of clinically significant endometrial pathology in women with histologically confirmed lesions: a retrospective case-control study.

Objective: To identify clinical and ultrasound predictors of clinically significant endometrial pathology and to develop an internally validated clinicosonographic model in women with histologically confirmed endometrial or intracavitary lesions. Methods: This unmatched retrospective case-control study included women with histologically confirmed lesions who had complete clinicosonographic data. All eligible women with clinically significant pathology during the study period were included (n=500), and a random 1:1 sample of eligible benign-pathology controls was selected. The primary outcome was clinically significant pathology, defined as endometrial intraepithelial neoplasia (EIN)/atypical hyperplasia or endometrial carcinoma. Univariable and multivariable logistic regression analyses were performed to evaluate clinical and sonographic predictors. Three prespecified models were developed: clinical, sonographic, and combined clinicosonographic models. Model discrimination was assessed using the area under the receiver operating characteristic curve (AUC), and internal validation was performed with bootstrap resampling. Results: Compared with benign pathology, clinically significant pathology was associated with older age, lower parity, greater endometrial thickness, larger lesion size, mixed/heterogeneous echogenicity, abnormal endomyometrial junction, higher Color Doppler flow imaging [CDFI] score, and more frequent polypoid mass-like morphology, while the bright-edge sign was more common in benign pathology. In the final combined model, independent predictors of clinically significant pathology were age (adjusted odds ratio [aOR] per 5-year increase 11.49, 95% confidence interval [CI] 8.10-16.28), parity (aOR per additional birth 0.63, 95% CI 0.53-0.75), endometrial thickness (aOR per 5-mm increase 1.98, 95% CI 1.55-2.53), lesion length (aOR per 1-mm increase 1.52, 95% CI 1.29-1.80), mixed/heterogeneous echogenicity (aOR 1.87, 95% CI 1.13-3.09), abnormal endomyometrial junction (aOR 1.80, 95% CI 1.14-2.83), minimal/moderate/abundant CDFI score versus none, bright-edge sign (aOR 0.37, 95% CI 0.20-0.66), and polypoid mass-like lesion (aOR 3.85, 95% CI 2.42-6.12). The combined model showed the best discrimination, with an AUC of 0.959 and optimism-corrected AUC of 0.957. Conclusions: Clinically significant endometrial pathology is associated with a distinct clinicosonographic profile that includes older age, lower parity, thicker endometrium, larger lesions, heterogeneous echogenicity, junctional abnormality, increased vascularity, and polypoid mass-like morphology, whereas the bright-edge sign appears reassuring. A combined clinical and ultrasound model showed excellent internal discrimination and may help structure risk assessment in women with pathology-confirmed endometrial or intracavitary lesions.