A Radiotherapy Nutritional Risk Assessment System and Method Based on Multimodal Data

By constructing a radiotherapy nutrition risk assessment system based on multimodal data, integrating multimodal data and introducing an attention mechanism, the system addresses the issues of lag and one-sidedness in existing nutritional risk assessment technologies. This enables accurate prediction of nutritional risks for radiotherapy patients and personalized intervention recommendations, thereby improving the timeliness and precision of nutritional support.

CN121862316BActive Publication Date: 2026-05-26NANTONG TUMOR HOSPITAL
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
NANTONG TUMOR HOSPITAL
Filing Date
2026-03-17
Publication Date
2026-05-26

AI Technical Summary

Technical Problem

In existing technologies, nutritional risk assessment for radiotherapy patients mainly relies on single or static clinical indicators, lacking integrated analysis of multimodal time-series data. This results in delayed and one-sided risk assessment, making it impossible to accurately predict the evolution trajectory of patients' nutritional risks, and intervention recommendations lack individualization and dynamic adjustment capabilities.

Method used

The radiotherapy nutrition risk assessment system based on multimodal data integrates structured clinical data, unstructured text records, and time-series radiomics data of radiotherapy patients to construct a multimodal high-dimensional feature tensor time-series sequence. It introduces an attention mechanism to fusion network adaptively learn the dynamic weights of each modality feature, generates a weighted fusion multimodal feature vector, establishes a radiotherapy nutrition risk time-series prediction model, and generates individualized nutrition intervention suggestions by associating with the hospital's nutrition intervention knowledge base.

Benefits of technology

It enables a detailed portrayal of the patient's condition throughout the entire life cycle, improves the accuracy of nutritional risk level prediction and the identification of risk evolution trajectory, generates highly individualized and dynamically adjustable nutritional intervention recommendations, and realizes a shift in decision-making mode from passive response to proactive early warning.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN121862316B_ABST
    Figure CN121862316B_ABST
Patent Text Reader

Abstract

This invention discloses a radiotherapy nutrition risk assessment system and method based on multimodal data, belonging to the field of medical information technology. It includes: a high-dimensional data feature module for each modality, integrating structured clinical data, unstructured text records, and time-series radiomics data of radiotherapy patients to form a multimodal data stream and establishing a time-series sequence of high-dimensional feature tensors for the patient; a radiotherapy nutrition risk time-series prediction module, adaptively learning the dynamic weights of each modality feature based on the patient's current state, generating a weighted fusion multimodal feature vector, establishing a time-series prediction model for the patient's radiotherapy nutrition risk, and predicting the nutritional risk level of the patient after N treatment fractions, identifying its risk evolution trajectory; and a radiotherapy nutrition risk assessment module, associating with a paired hospital nutrition intervention knowledge base to generate individualized nutritional intervention suggestions for the patient, realizing radiotherapy nutrition risk assessment based on multimodal data. This invention achieves a refined characterization of the patient's state throughout the entire treatment cycle.
Need to check novelty before this filing date? Find Prior Art