METHODS: Using a 17-segment model, coronary computed tomography angiography images were analyzed for subendocardial and transmural attenuation and the corresponding blood pool. The segment with the lowest subendocardial attenuation and transmural attenuation were normalized to the segment with the highest subendocardial and transmural attenuation, respectively (SUBnormalized, and TRANSnormalized, respectively). We evaluated the independent and incremental value of myocardial attenuation to predict the composite of cardiovascular death or nonfatal myocardial infarction.

RESULTS: Of a total of 995 coronary CTA VISION (Coronary Computed Tomographic Angiography and Vascular Events in Noncardiac Surgery Patients Cohort Evaluation Study) patients, 735 had available images and complete data for these analyses. Among these patients, 60 had MACE. Based on Revised Cardiovascular Risk Index, 257, 302, 138, and 38 patients had scores of 0, 1, 2, and ≥3, respectively. On coronary computed tomography angiography, 75 patients had normal coronary arteries, 297 patients had nonobstructive coronary artery disease, 264 patients had obstructive disease, and 99 patients had extensive obstructive coronary artery disease. SUBnormalized was an independent and incremental predictor of events in the model that included Revised Cardiovascular Risk Index and coronary artery disease severity. Compared with patients in the highest tertile of SUBnormalized, patients in the second and first tertiles had an increased hazards ratio for events (2.23 [95% CI, 1.091-4.551] and 2.36 [95% CI, 1.16-4.81], respectively). TRANSnormalized, as a continuous variable, was also found to be a predictor of MACE (P=0.027).