Cardiac CT progresses in assessing myocardium

Cardiac CT is moving closer to competing with MR in imaging damage to the myocardium caused by infarction.

Dr. Andreas H. Mahnken, a radiologist at Aachen University of Technology in Germany, detailed the incremental studies needed to validate 16-slice CT for this indication in three papers presented at the ECR on Sunday.

The first report showed that 16-slice CT can reliably assess left ventricular volumes and regional wall motion at rest. But the researchers found that 16-slice CT cannot assess functional parameters that directly depend on temporal resolution. Mahnken and colleagues reached this conclusion by comparing multislice CT with MRI in pigs.

In the second report, the Aachen researchers found that delayed contrast-enhanced CT is as reliable as delayed contrast-enhanced MR to assess infarct size. They performed reperfused myocardial infarction contrast-enhanced cardiac MSCT (16 slice, 50 kVp, 500 mAs) and late-enhancement 1.5T MRI in 28 patients.

There was good agreement between MRI and late-enhancement CT. The results were poorer but still within a good range when MRI and early-phase CT were compared, Mahnken said. On early- and late-phase CT, density values of infarcted myocardium were significantly different from viable myocardium.

Mahnken said he routinely uses coronary CT angiography, and this study was an attempt to determine what other information could be seen with a lower CT dose.

The third study determined that perfusion-weighted color maps from cardiac CT data improve detection of acute myocardial infarction.

The researchers developed a software tool for semiautomated detection of the long axis of the left ventricle and assignment of left-ventricular segments. The software color-codes changes in the myocardial contrast enhancement pattern. Normal myocardium is encoded in green and infarction in blue.

The group used cardiac CT (120 kVp, 550 mAs) and late-enhancement MR to examine 15 patients. MRI showed acute MI in 78 of 255 myocardial segments. Routine CT correctly detected MI in 58 segments (74% sensitivity, 96% specificity). Using the color-coded maps, observers increased sensitivity to 83%, identifying MI in 65 segments.

The agreement between MRI and routine MSCT images was κ = 0.76. Using postprocessed images, the agreement improved to κ = 0.81.