Related Experiment Video
Updated: Jul 10, 2026
05:41
Synchronous Triplanar Reconstruction Integrated with Color Doppler Mapping for Precise and Rapid Localization of Thyroid Lesions
Published on: February 9, 2024
LLM-Enhanced Multimodal Fusion of SPECT Radiomics and Clinical Data for Predicting 131I Therapeutic Response in
Hangyu Wang1,2,3, Erhao Chen4, Liqin Pan4
1School of Biomedical Engineering, Southern Medical University, 1023 Shatai Road, Guangzhou, 510515, China.
Molecular Imaging and Biology
|July 8, 2026
Summary
This study developed a new model combining imaging features and clinical data to predict radioactive iodine treatment success in differentiated thyroid cancer patients. The advanced model accurately predicts treatment response, aiding personalized patient management.
Area of Science:
- Nuclear medicine imaging
- Radiomics and artificial intelligence
- Oncology
Background:
- Differentiated thyroid cancer (DTC) patients treated with radioactive iodine (RAI) face risks of poor outcomes.
- Accurate early prediction of RAI efficacy is crucial for effective patient management.
Purpose of the Study:
- To create a predictive model integrating Single Photon Emission Computed Tomography (SPECT) radiomics and clinical data for DTC patients.
- To assess the model's ability to predict the therapeutic efficacy of 131I treatment.
Main Methods:
- Extracted 1,688 radiomic features from SPECT images of 311 DTC patients.
- Developed multimodal models using early-fusion and late-fusion strategies, combining radiomics with clinical data.
- Utilized a Large Language Model (LLM) fine-tuned with low-rank adaptation for enhanced prediction.
Main Results:
- Radiomic models showed improved prediction over clinical data alone.
- The LLM-enhanced multimodal model achieved a high Area Under the Curve (AUC) of 0.85.
- Significant improvements in sensitivity and F1-score were observed with the final model.
Conclusions:
- The multimodal framework offers accurate early prediction of RAI response in DTC.
- This approach shows potential for optimizing individualized treatment planning and follow-up strategies.
