CN-122000079-A - Child tumor management system and method based on multi-mode large model

CN122000079ACN 122000079 ACN122000079 ACN 122000079ACN-122000079-A

Abstract

The invention discloses a child tumor management system and method based on a multi-mode large model, and relates to the technical field of child tumor management. The multimodal data analysis module processes clinical text, medical images, pathology reports, gene sequencing, nutrition assessment and psychological scales to generate structured data. The retrieval enhancement generation module retrieves the first information fragment from the child tumor field knowledge base according to the structured data, generates first traceable response content through a semantic rearrangement technology, and further generates visual information and text information related to the health condition of the patient to be provided for the patient or family members. The invention integrates multi-mode data to generate structured information, suppresses artificial intelligent illusion through traceable response, outputs health information in popular form, solves the problems of information overload and fragmentation of patient management, and improves diagnosis and treatment efficiency and accuracy.

Inventors

GONG MENGCHUN
Lv Maoxin
CHEN YICHAO
FAN JUNKANG
FU TAOTAO

Assignees

神州医疗科技股份有限公司

Dates

Publication Date: 20260508
Application Date: 20251229

Claims (10)

1. The child tumor management system based on the multi-mode large model is characterized by comprising a multi-mode data analysis module, a retrieval enhancement generation module and an information generation module; The multi-modal data analysis module is used for processing multi-modal data of the child tumor patient to generate structured data, wherein the multi-modal data comprises clinical texts, medical images, pathology reports, gene sequencing, nutrition evaluation and psychological scales; The retrieval enhancement generation module is used for retrieving related first information fragments from a knowledge base in the field of children tumor according to the structured data, sequencing the first information fragments by utilizing a semantic rearrangement technology, and generating first traceable response content; the information generation module is used for generating visual information and/or text information related to the health condition of the pediatric tumor patient based on the first traceable response content and providing the visual information and/or text information to the pediatric tumor patient or family members of the pediatric tumor patient.
2. The system for managing the children tumor based on the multi-mode big model of claim 1, further comprising a patient end module, wherein the patient end module is used for sending the interaction content with a target user to the retrieval enhancement generation module, and the target user is the children tumor patient or the family member of the children tumor patient; The retrieval enhancement generation module is also used for retrieving relevant second information fragments from the child tumor field knowledge base according to the structured data and the interaction content with the target user, and sequencing the second information fragments by utilizing a semantic rearrangement technology to generate second traceable response content; The patient-side module is further configured to generate intelligent question-answer content, psychological screening assessment information, peer care information, follow-up management advice, and nutrition guidance advice based on the second traceable response content, and provide the intelligent question-answer content, psychological screening assessment information, peer care information, follow-up management advice, and nutrition guidance advice to the target user.
3. The child tumor management system based on the multi-mode large model of claim 2, further comprising a doctor-side module for sending the interaction content with the doctor to the retrieval enhancement generation module; The retrieval enhancement generation module is further specifically configured to retrieve related third information segments from the knowledge base of the pediatric tumor field according to the structured data, the interaction content with the target user, and the interaction content with the doctor, and sort the third information segments by using a semantic rearrangement technology to generate third traceable response content; the doctor-side module is also used for generating intelligent question-answer content, a patient management overview and a follow-up record overview based on the third traceable response and providing the intelligent question-answer content, the patient management overview and the follow-up record overview to a doctor.
4. A pediatric tumor management system based on a multimodal big model according to any of the claims 1-3, wherein the pediatric tumor field knowledge base stores authoritative diagnosis and treatment guidelines, expert clinical experience and desensitization case data for the pediatric tumor field.
5. A method for managing children's tumor based on a multi-modal large model, comprising: Processing multi-modal data of a child tumor patient to generate structured data, wherein the multi-modal data comprises clinical texts, medical images, pathology reports, gene sequencing, nutrition evaluation and psychological scales; According to the structured data, retrieving related first information fragments from a knowledge base in the field of children tumor, and sequencing the first information fragments by utilizing a semantic rearrangement technology to generate first traceable response content; Based on the first traceable response content, visual information and/or text information related to the health condition of the pediatric tumor patient is generated and provided to the pediatric tumor patient or the family members of the pediatric tumor patient.
6. The method for managing children's tumors based on the multimodal big model of claim 5, further comprising: According to the structured data and the interaction content with a target user, retrieving relevant second information fragments from the child tumor field knowledge base, and sequencing the second information fragments by utilizing a semantic rearrangement technology to generate second traceable response content, wherein the target user is the child tumor patient or the family of the child tumor patient; Intelligent question-answer content, psychological screening assessment information, peer care information, follow-up management advice, and nutrition guidance advice are generated based on the second traceable response content and provided to the target user.
7. The method for managing children's tumors based on the multimodal big model of claim 6, further comprising: According to the structured data, the interaction content with the target user and the interaction content with doctors, retrieving relevant third information fragments from the knowledge base of the children tumor field, and sequencing the third information fragments by utilizing a semantic rearrangement technology to generate third traceable response content; Intelligent questioning content, patient management overview, and follow-up record overview are generated based on the third traceable response and provided to the physician.
8. The method of any one of claims 5 to 7, wherein the pediatric tumor domain knowledge base stores authoritative diagnosis and treatment guidelines, expert clinical experience, and desensitized case data for the pediatric tumor domain.
9. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing a method of childhood tumor management based on a multimodal big model as claimed in any of claims 5 to 8 when the computer program is executed.
10. A computer readable storage medium, characterized in that the computer readable storage medium has stored thereon a computer program which, when executed by a processor, implements a method for managing children's tumors based on a multimodal big model as claimed in any of the claims 5 to 8.

Description

Child tumor management system and method based on multi-mode large model Technical Field The invention relates to the technical field of childhood tumor management, in particular to a childhood tumor management system and method based on a multi-mode big model. Background The pathological parting of the malignant tumor of children is complex, the progress is rapid, the treatment window is narrow, the child is in the growth and development stage, and the requirements on diagnosis and treatment precision and long-term life quality are extremely high. The current diagnosis and treatment mode faces the following pain points that professional resources are scarce and unevenly distributed, the center pediatric tumor expert is concentrated in a large city, the diagnosis and treatment capability of basic medical institutions is insufficient, the diagnosis and treatment delay and diagnosis of the child are caused, information overload and decision making are complex, the child tumor relates to multi-dimensional heterogeneous data such as clinic, image, pathology, genes and the like, doctors are difficult to integrate rapidly and form an optimal diagnosis and treatment scheme, and the patient management fragmentation is carried out, and all links are disjointed from pre-diagnosis, treatment to long-term follow-up, nutrition and psychological support and lack of an integrated full-period management tool. In the prior art, generic large models or simple medical question-answering systems have been attempted to be applied to assist in medical decision making. These systems are typically based on natural language processing techniques that enable preliminary analysis and answers to medical data in textual form. Some systems also attempt to integrate limited medical image data to provide auxiliary diagnostic advice via image recognition techniques. These prior art solutions provide knowledge questions and answers and simple decision support in specific medical scenarios, mainly by means of model fine tuning or rule engines. However, the prior art has significant drawbacks. The general large model or the simple medical question-answering system has the artificial intelligent illusion problem, inaccurate or fictional information is output, the specialty is insufficient, and the multi-mode medical data in the field of children tumor cannot be safely and effectively processed. The systems lack deep optimization aiming at the specificity of the children tumor, and are difficult to realize the integrated analysis and fusion treatment of multi-mode data such as clinical texts, medical images, pathology reports, gene sequencing, nutrition evaluation, psychological scales and the like. Meanwhile, the existing system cannot ensure the traceability of output, privacy leakage risks exist when sensitive medical data are processed, and medical-grade accuracy and safety compliance requirements are difficult to meet. Aiming at the defects in the prior art, an intelligent system capable of effectively processing multi-mode data, ensuring output accuracy and providing full-period management support is urgently needed so as to improve the efficiency and quality of children tumor diagnosis and treatment. Disclosure of Invention The invention aims to solve the technical problems of the prior art, and particularly provides a multi-mode large model-based child tumor management system and method, which are as follows: 1) In a first aspect, the invention provides a child tumor management system based on a multi-mode large model, which comprises the following specific technical scheme: the system comprises a multi-mode data analysis module, a retrieval enhancement generation module and an information generation module; The multi-modal data analysis module is used for processing multi-modal data of the child tumor patient to generate structured data, wherein the multi-modal data comprises clinical texts, medical images, pathology reports, gene sequencing, nutrition evaluation and psychological scales; The retrieval enhancement generation module is used for retrieving related first information fragments from the knowledge base of the children tumor field according to the structured data, and sequencing the first information fragments by utilizing a semantic rearrangement technology to generate first traceable response content; the information generation module is used for generating visual information and/or text information related to the health condition of the pediatric tumor patient based on the first traceable response content and providing the visual information and/or text information to the pediatric tumor patient or family members of the pediatric tumor patient. The child tumor management system based on the multi-mode large model has the following beneficial effects: The multi-mode data analysis module is used for carrying out integrated processing on multi-mode data such as clinical texts, medical images, pathology reports, gene sequencin