Daniel Kopeinigg, Presenter: Nothing to Disclose

Dominik Fleischmann MD, Abstract Co-Author: Nothing to Disclose

Roland Bammer, Abstract Co-Author: Nothing to Disclose

Accurate prediction and controlling of individual arterial enhancement is highly desirable for continuously evolving cardiac CT. We report the accuracy of a Linear System Theory (LST) based 'forward approach' which mathematically extracts the patient-specific arterial enhancement response from a test bolus injection, to (i) predict subsequent coronary CT enhancement and to (ii) compute individual injection profiles to achieve a desired target enhancement profile.

Sixteen asymptomatic individuals (15m/1f, mean age 56y, range:45-65y) with at least one risk factor for CAD underwent coronary CTA (Siemens Somatom Definition). Individual time-attenuation responses (25 images, 0.5Hz) were extracted after a test bolus injection of 20mL contrast medium (CM) injected at 5mL/s, followed by a 40mL saline flush. Subsequent uniphasic or biphasic CM injections (82-126mL, at 3.3-6.0mL/s) were weight based (n=3) or computed using the LST-based method (n=13), with a target enhancement of 240-500HU above baseline (±10% error tolerance). Predicted individual enhancement profiles were compared to the actual attenuation profiles during coronary CTA (scan-times 7-19sec). The absolute (ΔHU) point-wise difference between the predicted and the measured arterial enhancement was calculated as the error metric. Average (ΔHUmean) and maximum (ΔHUmax) errors were derived for each individual.

LST-based prediction of enhancement and calculation of individual injection profiles took less than 30s of computation time per patient. Average arterial opacification achieved in the target vasculature was 386 HU (range: 273-520HU). The average ΔHUmean between predicted target enhancement and measured enhancement was 50HU (range: 18-119 HU), the average ΔHUmax was 75±34HU (range: 33-153HU) with a relative error range to the target enhancement of 5-28%.

LST-based prediction of individual CM enhancement is in good agreement with measured enhancement profiles and within the expected target opacification for coronary CTA.

Improved predictability and individualization of CM administration may reduce excessive opacification currently tolerated as a saftey cushion and thus reduce overall CM volumes in coronary CTA.

Kopeinigg, D, Fleischmann, D, Bammer, R, Linear Systems Theory (LST)-based Prediction and Individual Tailoring of Contrast Medium-enhancement for Coronary CTA.  Radiological Society of North America 2009 Scientific Assembly and Annual Meeting, November 29 - December 4, 2009 ,Chicago IL. http://archive.rsna.org/2009/8013343.html