Prediction of complete regression in fertility-sparing patients with endometrial cancer and apical hyperplasia: the GLOBAL model in a large Chinese cohort

Autor: Xingchen Li, Yiqin Wang, Jiaqi Wang, Yuan Fan, Jianliu Wang
Jazyk: angličtina
Rok vydání: 2024
Předmět:
Zdroj: Journal of Translational Medicine, Vol 22, Iss 1, Pp 1-12 (2024)
Druh dokumentu: article
ISSN: 1479-5876
DOI: 10.1186/s12967-023-04671-w
Popis: Abstract Background Fertility preservation treatment is increasingly essential for patients with apical endometrial hyperplasia (AEH) and early endometrial cancer (EEC) worldwide. Complete regression (CR) is the main endpoint of this treatment. Accurately predicting CR and implementing appropriate interventions during treatment are crucial for these patients. Methods We conducted a retrospective study involving 193 patients diagnosed with atypical AEH or EEC, enrolled from January 2012 to March 2022 at our center. We evaluated 24 clinical parameters as candidate predictors and employed LASSO regression to develop a prediction model for CR. Subsequently, a nomogram was constructed to predict CR after the treatment. We evaluated the performance of the nomogram using receiver operator characteristic (ROC) curve and decision curve analysis (DCA) to assess its predictive accuracy. Additionally, we employed cumulative curves to determine the CR rate among patients. Results Out of the 193 patients, 173 achieved CR after undergoing fertility preservation treatment. We categorized features with similar properties and provided a list of formulas based on their coefficients. The final model, named GLOBAL (including basic information, characteristics, blood pressure, glucose metabolism, lipid metabolism, immunohistochemistry, histological type, and medication), comprised eight variables identified using LASSO regression. A nomogram incorporating these eight risk factors was developed to predict CR. The GLOBAL model exhibited an AUC of 0.907 (95% CI 0.828–0.969). Calibration plots demonstrated a favorable agreement between the predicted probability by the GLOBAL model and actual observations in the cohort. The cumulative curve analysis revealed varying cumulative CR rates among patients in the eight subgroups. Categorized analysis demonstrated significant diversity in the effects of the GLOBAL model on CR among patients with different total points (p
Databáze: Directory of Open Access Journals
Externí odkaz:
Nepřihlášeným uživatelům se plný text nezobrazuje