To evaluate the predictive performance of a multiparameter MRI-based radiomics model for progressive neurologic deterioration in patients with subacute ischemic perforator artery cerebral infarction.
Key Findings:
The combined model achieved an AUC of 0.844 in the training cohort and 0.824 in the validation cohort.
Sensitivity, specificity, and accuracy for the combined model were approximately 76%, 75%, and 76% in the training cohort, and 65%, 73%, and 70% in the validation cohort.
Single-sequence models had lower AUC values ranging from 0.769 to 0.820 in the training cohort.
Interpretation:
The multiparameter MRI-based radiomics model outperforms single-sequence models in predicting progressive cerebral infarction, indicating its potential utility in clinical settings.
Limitations:
Modest sample size and restriction to single perforating artery infarctions.
Heterogeneity in infarct location and lack of external validation.
Retrospective design and potential partial volume effects.
Conclusion:
A radiomics-based model using multiple MRI sequences can effectively predict clinical symptom progression in patients with subacute ischemic perforator artery cerebral infarction.
Radiologists assigned to receive step-by-step explanations from a large language model achieved higher diagnostic accuracy in a randomized vignette study, while differential-diagnosis outputs may have increased inappropriate reliance on incorrect model suggestions.
