Image prediction tumor treatment reactivity classification method, device, equipment and medium

By introducing deep supervision strategies and multi-layer perceptron structure into the tumor treatment responsiveness prediction model, combining imaging and clinical data, the problems of poor prediction results and poor interpretability in the prior art are solved, and a higher-precision tumor treatment responsiveness prediction is achieved.

CN119693724AActive Publication Date: 2025-03-25WEST CHINA HOSPITAL SICHUAN UNIV
4 Cites 0 Cited by

Patent Information

Application Number
CN202510207158.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-02-25
Publication Date
2025-03-25
Estimated Expiration
2045-02-25

AI Technical Summary

Technical Problem

When predicting tumor treatment responsiveness, the prior art relies on pre-treatment imaging data and fails to effectively combine clinical data and biochemical indicators, resulting in poor prediction effect and poor interpretability.

Method used

By obtaining the basic network model and introducing deep supervision strategies and multi-layer perceptron structure to it, the initial network model is built. Then, based on the image data set, including pre-treatment and post-treatment CT images, the teacher-student network model is trained to obtain the target model. This model is able to predict based on imaging and clinical features.

Benefits of technology

It improves the accuracy of tumor treatment response prediction, can more accurately predict the complete pathological remission rate, and enhances the interpretability of the model.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention provides an image prediction tumor treatment reactivity classification method and device, equipment and a medium, and relates to the technical field of image recognition, and the method comprises the steps: obtaining a basic network model, introducing a depth supervision strategy and a multi-layer perceptron structure into the basic network model, and obtaining an initial network model; determining a teacher-student network model according to the initial network model; an image data set is obtained, the image data set comprises a first data subset and a second data subset, the first data subset comprises a plurality of pieces of first data, the second data subset comprises a plurality of pieces of second data, and each piece of second data corresponds to one piece of first data; training a teacher-student network model based on the first data subset and the second data subset to obtain a target model; and obtaining a to-be-classified image, and inputting the to-be-classified image into the target model to obtain a corresponding classification result. According to the invention, the prediction precision of the pathological complete remission condition is improved.
Need to check novelty before this filing date? Find Prior Art