Fig. 4: Example of a patient undergoing single-photon emission computed tomography/computed tomography (SPECT/CT) myocardial perfusion imaging with an extracardiac structure increasing the highest risk of all-cause mortality.
From: Holistic AI analysis of hybrid cardiac perfusion images for mortality prediction
Waterfall plot shows top 10 structures influencing mortality risk in the computed tomography attenuation correction (CTAC) model, highlighting Shapley Additive Explanations values (X-axis) and key structures. A 66-year-old male patient was classified to be at higher risk of death with the lungs (right lower lobe) contributing the most to the elevated risk (red arrow on the waterfall plot). a CTAC, axial view, with a corresponding deep learning structures segmentation (b) revealed a 39×39 mm solid mass with irregular margins in the right lower lobe. c CTAC with an overlayed SPECT scan showed no uptake of the radiotracer. d, e 3D reconstruction of all segmented and ranked structures. The patient had abnormal myocardial perfusion (total perfusion deficit of 7.65) and died 48 days after the exam.
