Skip to main content
. 2016 Jun 13;5(6):e003226. doi: 10.1161/JAHA.116.003226

Figure 5.

Figure 5

Reclassification of patient risk. The addition of CTA‐derived data to the Cox proportional hazards model improved the classification of primary outcome events over that from the UKPDS and CAC scores alone. Left: classification into low, intermediate, and high risk by UKPDS and CAC scores alone. Right: Classification with the Cox prognostic model based on the clinical risk scores and the CTA findings. Above are subjects with no CHD event (noncases) and below are subjects with a CHD event (cases). Each mannequin represents 1 subject. Among the noncases, 102 subjects (43% of risk groups 2 and 3) were correctly reclassified to a lower risk at the cost of incorrect reclassification of 45 (9% of risk groups 1 and 2) to a higher‐risk group. Among the cases, 11 (44% of risk groups 1 and 2) were correctly reclassified to a higher‐risk at the cost of incorrect reclassification of 5 (16% of risk groups 2 and 3) to a lower‐risk group. The net reclassification improvement was 0.24. CAC indicates coronary artery calcium; CHD, coronary heart disease; CTA, computed tomography angiography; UKPDS, United Kingdom Prospective Diabetes Study.