A dual-energy computed tomography-based radiomics nomogram for predicting time since stroke onset: a multicenter...

Area of Science:

• Radiology
• Medical Imaging
• Stroke Medicine

Background:

• Accurate determination of time since stroke (TSS) is critical for treatment decisions.
• Dual-energy computed tomography (DECT) offers advanced imaging capabilities for stroke assessment.

Purpose of the Study:

• To develop and validate a radiomics nomogram using DECT images and clinical data to classify TSS.
• To facilitate stroke decision-making through improved TSS classification.

Main Methods:

• Retrospective study of 488 stroke patients undergoing DECT.
• Extraction of 535 radiomics features from various DECT image types.
• Development of a radiomics nomogram combining radiomics and clinical factors using multivariate logistic regression.
• Evaluation of predictive performance using ROC and DCA.

Main Results:

• A radiomics nomogram incorporating clinical and radiomics features demonstrated superior predictive value for TSS.
• The nomogram achieved a higher AUC (0.936) in the validation cohort compared to radiomics-only (0.905) and clinical-only (0.824) models.
• Decision curve analysis confirmed the clinical utility of the radiomics nomogram.

Conclusions:

• DECT-based radiomics nomograms show promise for predicting TSS.
• The developed nomogram supports potential clinical use for predicting TSS and facilitating individualized stroke management.