Predicting Response to Radiotherapy in Locally Advanced Cervical Squamous Cell Carcinoma Based on Multisequence MRI
Yanhong Zhuo1, Youjia Wang1, Yimin Li1
1Department of Radiation Oncology, Zhangzhou Hospital Affiliated to Fujian Medical University, Zhangzhou, Fujian 350600, China (Y.Z., Y.W., Y.L., L.Q., D.K., Y.X.).
Academic Radiology
|July 31, 2026
Summary
A machine learning model integrating radiomics and clinical data accurately predicts tumor shrinkage rate (TSR) in locally advanced cervical cancer (LACC) patients undergoing external beam radiotherapy (EBRT). This aids in early identification of high-risk individuals for personalized treatment.
Area of Science:
- Oncology
- Radiology
- Artificial Intelligence
Background:
- Locally advanced cervical cancer (LACC) is a significant cause of cancer mortality in women.
- Accurate prediction of treatment response is crucial for effective management.
Purpose of the Study:
- To develop and validate a machine learning model for predicting tumor shrinkage rate (TSR) after external beam radiotherapy (EBRT) in LACC patients.
- To integrate radiomic features from MRI with clinical data for enhanced predictive performance.
Main Methods:
- Retrospective analysis of 248 LACC patients.
- Deep learning extraction of radiomic features from multi-sequence MRI (T1-weighted, T2-weighted, T2-SPAIR, DWI).
- Fusion of radiomic and clinical features; training of multiple machine learning models (CWGBS, XGBoost, GBST, RSF).
Main Results:
- Machine learning fusion models demonstrated robust TSR prediction performance.
- Area Under the Curve (AUC) for fusion models ranged from 0.824 to 0.863.
- The Random Survival Forest (RSF) model achieved the highest performance: accuracy 0.920, sensitivity 0.937, specificity 0.833, and AUC 0.863.
Conclusions:
- A deep-learning radiomics-clinical model effectively predicts TSR after EBRT in LACC.
- Early identification of patients likely to have poor response facilitates individualized treatment planning.
