Fig. 5
From: Machine learning models using dual-phase CT radiomics for early detection of PRISm
Decision curve analysis (DCA) results of the three models based on expiratory-phase CT images across the three cohorts. Panels (a), (b), and (c) show the DCA curves for the training cohort, internal test cohort, and external test cohort, respectively. The blue, orange, and green lines represent the net benefits of the clinical model, radiomics model, and combined model (integrating clinical and radiomics features), respectively. The black dashed line indicates the strategy of treating no patients, while the black solid line indicates the strategy of treating all patients. The combined model demonstrated higher net benefit across a range of decision thresholds, suggesting its superior clinical utility in identifying PRISm. Figures were automatically generated using the Onekey AI platform, and font sizes and label styles follow system defaults.
