Hybrid model of CT-fractional flow reserve, pericoronary fat attenuation index and radiomics for predicting the progression of WMH: a dual-center pilot study

Figures

• 

  Figure 1. Patients disposition and research design. CCTA, coronary computed tomography angiography; MRI, magnetic resonance imaging; ACS, acute coronary syndrome.

• 

  Figure 2. The specific calculation process of CT-FFR and pFAI and radiomics extraction. (A) Original CCTA images. (B,C) CCTA reconstruction images and lesion segments display. (D) PCAT ROI. (E) The calculation process and formula of CT-FFR. (F) pFAI acquisition and radiomics features analysis. CT-FFR, CT fractional flow reserve; pFAI, percoronary fat attenuation index; CCTA, coronary computed tomography angiography; PCAT, pericoronary adipose tissue.

• 

  Figure 3. Procedures used for radiomics analysis. CCTA, coronary computed tomography angiography; CT-FFR, CT fractional flow reserve; pFAI, pericoronary fat attenuation index; WMH, white matter hyperintensity.

• 

  Figure 4. Selection of myocardium and PCAT radiomics signatures for the prediction models. (A,C) Myocardium (top) and PCAT (bottom) signatures selection by Elastic Net Regression with the optimal penalization coefficient lambda (λ) using 10-fold cross-validation and the minimal criteria process. The x-axis shows Lambda (λ), and the y-axis shows the model AUC. (B,D) Elastic Net Regression coefficient profiles of the myocardium signatures (top) and PCAT signatures (bottom). (E,F) Importance ranking of the myocardium signatures (left) and PCAT signatures (right) that had the greatest correlations with WMH progression. PCAT, pericoronary adipose tissue, WMH, white matter hyperintensity.

• 

  Figure 5. Scatter box plot of radiomics signatures scores. (A–C) Myocardium signatures radiomics scores in the “WMH progression” group and the “no WMH progression” group in the training set, internal validation set and external validation set. (D–F) PCAT signatures radiomics scores in the “WMH progression” group and the “no WMH progression” group in training set, internal validation set and external validation set. The horizontal line in the box represents the median of the signatures score, the upper and lower borders of the box represent the upper and lower quartiles of signatures score, and the scattered points around the box represent the distribution position of each radiomics signatures score. PCAT, pericoronary adipose tissue; WMH, white matter hyperintensity. *Represents p < 0.05, **represents p < 0.01, *** represents p < 0.001, and ***represents p < 0.0001.

• 

  Figure 6. The variables included in the hybrid model are presented as a nomogram.

• 

  Figure 7. (A–C) calibration curves of hybrid model for WMH progression in the training set, internal validation set and external validation set. The x-axis represents the predicted probability of WMH progression estimated by the nomogram, and the y-axis represents the actual WMH progress probability. The red solid line represents logistic regression calibration, and the green dashed line represents nonparametric testing. The closer the red solid fit to the green dashed, the higher the accuracy of nomogram for predicting WMH progression. The triangle represents the grouped observations. (D–F) The predictive probability of the hybrid model in the “WMH progression” group and the “no WMH progression” group in the training set, internal validation set and external validation set. The horizontal line in the box represents the median of the predictive probability, the upper and lower borders of the box represent the upper and lower quartiles of predictive probability, and the scattered points around the box represent the distribution position of predictive probability. CT-FFR, CT fractional flow reserve; pFAI, pericoronary fat attenuation index; PCAT, pericoronary adipose tissue; radscore,…

• 

  Figure 8. Diagnostic performance of different models in predicting WMH progression. (A–C) ROC curves for different models for prediction of WMH progression in the training set, internal validation set and external validation set. (D–F) DCA of different models from the training set, internal validation set and external validation set. The horizontal black line indicates the net benefit assuming that no patients have WMH progression and the smooth gray line indicates the net benefit assuming that all patients have WMH progression. PCAT, pericoronary adipose tissue; pFAI, pericoronary fat attenuation index; CT-FFR, CT fractional flow reserve; ROC, receiver operating characteristic; DCA, decision curve analysis.