Breast cancer neoadjuvant therapy prognosis multi-mode multi-task probability prediction method and system
By integrating various data from breast cancer patients using a multimodal, multi-task probabilistic prediction method, the potential outcomes of neoadjuvant therapy can be predicted. This addresses the issue of incomplete multimodal data integration in existing systems, improves the accuracy and interpretability of prognostic assessment, and provides objective support for individualized treatment plans.
Patent Information
- Application Number
- CN202610134701.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-01-30
- Publication Date
- 2026-05-29
AI Technical Summary
Existing prognostic systems for neoadjuvant therapy in breast cancer fail to effectively integrate multimodal data, ignore imaging and histopathological heterogeneity, lack model interpretability, and lack counterfactual predictive capabilities, making it difficult to select individualized treatment options.
A multimodal, multi-task probabilistic prediction method was adopted to integrate breast DCE-MRI images, radiological imaging reports, pathological reports, and clinical information. Feature extraction and fusion were performed using a deep learning model to predict the potential outcomes of different neoadjuvant therapy regimens, explicitly describe prediction uncertainty, and reduce selection bias.
It improves the accuracy and interpretability of prognostic assessment for neoadjuvant therapy, provides objective support for treatment decision-making, reduces selection bias, and enables the recommendation of individualized treatment plans.