Prediction of Neoadjuvant Therapy Sensitivity in Rectal Cancer Patients based on Deep Imaging Omics Models
10.11783/j.issn.1002-3674.2025.05.004
- VernacularTitle:基于深度影像组学模型的直肠癌患者新辅助治疗敏感性预测研究
- Author:
Guohong GAO
1
;
Zhipeng DING
;
Yan LI
Author Information
1. 哈尔滨医科大学附属肿瘤医院(150081);哈尔滨医科大学公共卫生学院卫生统计学教研室
- Publication Type:Journal Article
- Keywords:
Rectal cancer;
Neoadjuvant therapy;
Ensemble learning;
Classification model
- From:
Chinese Journal of Health Statistics
2025;42(5):661-665,671
- CountryChina
- Language:Chinese
-
Abstract:
Objective Exploring the performance of a sensitivity recognition model for neoadjuvant therapy based on ensemble learning algorithms for predicting neoadjuvant therapy sensitive patients.Methods In this study,255 rectal cancer patients who underwent standard neoadjuvant therapy and surgery at the Affiliated Cancer Hospital of Harbin Medical University were collected,of which 139 patients were sensitive to neoadjuvant therapy and 116 patients were not.The recursive feature elimination method was used to screen deep learning features and imaging histology features to construct an integrated learning model.The generalization performance of the model was verified using the leave-out method,and the sensitivity,specificity,positive predictive value,negative predictive value,accuracy,G-mean,F-measure,Mathews correlation coefficient(MCC),and area under the receiver operating charactpristic curve(AUC)were used to evaluate and compare the model performance.Results The performance of the integrated model on the external validation set was:an AUC value of 0.916(95％CI:0.899 to 0.935),a sensitivity of 1.000,a specificity of 0.833,a positive predictive value of 0.875,a negative predictive value of 1.000,an accuracy of 0.923,a G-mean of 0.913,an F-measure of 0.933,a MCC was 0.854.Conclusion The integration model of depth features and imaging omics features is superior to the individual depth feature model and imaging omics feature model,and has good practicality and reliability in identifying sensitive patients with neoadjuvant therapy,which can be used as a reference for clinical practice.
