Work

Magnetic resonance imaging-based radiomic model to predict the risk of intraoperative massive hemorrhage in patients with cesarean scar pregnancy

Year: 2025

Type: article

Abstract: OBJECTIVES: Development of magnetic resonance imaging (MRI)-based radiomic models to predict the risk of intraoperative massive hemorrhage in patients with cesarean scar pregnancy (CSP). MATERIAL AND METHODS: CSP patients (n = 126) from Center 1 were randomly assigned in a 7:3 ratio into a training set (n = 88) and an internal validation set (n = 38), and patients (n = 32) from Center 2 into an external validation set. Afterward, the clinical and radiomic features related to intraoperative massive hemorrhage were fed into the k-nearest Neighbor (KNN), support vector machine (SVM), Light Gradient Boosting Machine (Light GBM), and Multi- Layer Perception (MLP) to construct predictive clinical, radiomic, and combinatorial models. The performance of these models was assessed using area under curve (AUC), Delong's test, Decision Curve Analysis (DCA), and calibration curves. Youden's index was used to determine the optimal threshold. RESULTS: Eleven radiomic characteristics were found to be substantially linked to intraoperative massive hemorrhage. The combined in the gestational sac and peripheral to the gestational sac (IP) model (AUC = 0.959), constructed by MLP, had the best performance, with an optimal risk threshold of 0.180, as compared to the clinical model (AUC = 0.500) and the nomogram (AUC = 0.283). DCA and calibration curves demonstrated the IP model's good clinical predictive performance. CONCLUSIONS: The IP model for CSP was superior to the other models in this study in predicting the risk of intraoperative massive hemorrhage, which was significantly increased when the risk threshold exceeded 0.180. The model may help clinicians make individualized treatment decisions.

Source: Ginekologia Polska

DOI: 10.5603/gpl.102853

Authors: Bingjie Ye, Chenyu Chen, Ke Su, Rujia Fan, Bo Yuan

Institutions: Second Affiliated Hospital of Zhengzhou University, First Affiliated Hospital of Zhengzhou University, Henan Provincial People's Hospital, Zhengzhou University

Language: English

Topic: Radiomics and Machine Learning in Medical Imaging

Subfield: Radiology, Nuclear Medicine and Imaging

Field: Medicine

Domain: Health Sciences

SDG: Good health and well-being

Open Access status: gold

Estimated APC paid: $431