Osteoporotic compression fracture screening method based on multi-modal large language model

By employing a multimodal large language model screening method, the safety and accuracy issues of early screening for osteoporotic vertebral compression fractures have been addressed. This method enables efficient screening without imaging equipment, reduces costs, and improves the safety and standardization of home screening.

CN121983318BActive Publication Date: 2026-06-09THE AFFILIATED HOSPITAL OF SOUTHWEST MEDICAL UNIV
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
THE AFFILIATED HOSPITAL OF SOUTHWEST MEDICAL UNIV
Filing Date
2026-04-07
Publication Date
2026-06-09

AI Technical Summary

Technical Problem

Existing technologies struggle to achieve safe, standardized, and clinically significant early screening for osteoporotic vertebral compression fractures without imaging equipment and full physician involvement. This is particularly problematic in home and primary healthcare settings, where there is a high risk of concealed misdiagnosis, inaccurate assessment, and insufficient safety.

Method used

An osteoporotic compression fracture screening method based on a multimodal large language model is adopted. Through data collection, posture sequence construction, multimodal semantic feature extraction, feature aggregation and machine learning risk modeling, a structured quantitative scoring result is generated, and the probability prediction value of osteoporotic vertebral compression fracture is output.

Benefits of technology

It enables OVCF screening without the need for specialized imaging equipment, reducing medical costs, improving the safety and reliability of screening, enhancing the standardization and repeatability of home and community screening, supporting rapid screening and tiered medical decision-making, and adapting to deployment in primary care and multiple scenarios.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN121983318B_ABST
    Figure CN121983318B_ABST
Patent Text Reader

Abstract

The present application relates to the technical field of early screening of osteoporotic vertebral compression fractures, and proposes an osteoporotic compression fracture screening method based on a multi-modal large language model, which takes patient posture images and action videos and other unstructured visual data as input, introduces a multi-modal large language model to quantitatively evaluate key functions such as patient posture alignment, motion coordination and pain-related responses from images and videos, and outputs in the form of standardized scores, automatically extracting structured features with clear clinical significance. Based on the above structured features, a machine learning model is constructed to assess the risk of OVCF, and combined with SHAP feature contribution analysis and decision tree visualization methods, the key discriminant factors and their action directions are clarified, realizing the interpretable expression of the prediction process. A safe, low-cost, interpretable and easy-to-promote technical solution is provided for the early screening of OVCF, which is suitable for various application scenarios such as community and home screening.
Need to check novelty before this filing date? Find Prior Art