Nomogram using human epididymis protein 4 predicted concurrent endometrial cancer from endometrial atypical hyperplasia before surgery.

ABSTRACT

Objective: To establish a nomogram based on presurgical predictors of concurrent endometrial cancer (EC) for patients diagnosed with endometrial atypical hyperplasia before definitive surgery (preoperative-EAH) to improve the risk stratification and clinical application. Methods: Preoperative-EAH patients who underwent hysterectomy in a tertiary hospital from January 2020 to December 2022 were retrospectively analyzed. Independent predictors from the multivariate logistic regression model were used to establish a nomogram, and bootstrap resampling was used for internal validation. Results: Of 370 preoperative-EAH patients, 23.4% were diagnosed with EC after definitive surgery (final-EC). Multivariate analyses found three independent predictors of final EC human epididymis protein 4 (HE4) ≥43.50 pmol/L [odds ratio (OR) = 3.70; 95% confidence intervals (CI) = 2.06-6.67], body mass index (BMI) ≥ 28 kg/m2 (OR = 2.05; 95% CI = 1.14-3.69), and postmenopausal status, particularly at postmenopausal time ≥5 years (OR = 5.84, 95% CI = 2.51-13.55), which were used to establish a nomogram model. The bootstrap-corrected C-index of the nomogram was 0.733 (95% CI = 0.68-0.79), which was significantly higher than that of each individual factor. The calibration curve and decision curve showed good consistency and clinical net benefit of the model. At the maximum Youden index, 49.4% (43/87) of women in the high-risk group defined by nomogram had concurrent EC, versus 16.6% in the low-risk group (P< 0.001). Conclusion: The nomogram based on HE4, menopausal status, and BMI was found with an improved predictive value to stratify preoperative-EAH patients at high risk of concurrent EC for better clinical management.