Carla M Thompson MS, Presenter: Nothing to Disclose

Jens Wiegert PhD, Abstract Co-Author: Employee, Koninklijke Philips Electronics NV

Jeffrey Harold Yanof PhD, Abstract Co-Author: Employee, Koninklijke Philips Electronics NV

Kavitha Yaddanapudi, Abstract Co-Author: Nothing to Disclose

Sandra Simon Halliburton PhD, Abstract Co-Author: Research Grant, Siemens AG Research Grant, Koninklijke Philips Electronics NV

The DLP conversion ("k") factors ubiquitously used for CT effective dose estimation are based on Monte Carlo (MC) simulations with standard reference phantoms. The aim of this study was to compare effective doses (ED) estimated from patient-specific CT dose distribution maps with those estimated using DLP conversions factors.

Ten cardiac CT data sets (mean patient age 56.6 yrs; 5 F/5 M; BMI 22-42 kg/m2) were obtained with a 256-slice scanner (iCT, Philips Healthcare). These contrast-enhanced scans were performed with non-gated helical, retrospectively ECG-gated helical or prospectively ECG triggered axial techniques. A full field of view (500 mm) image reconstruction was used for the coronary arteries, pulmonary veins, or thoracic aorta with scan lengths ranging from 120-600 mm. The data sets were processed offline with a Monte Carlo simulation tool that 1) enabled the adjustment of tube potential and tube current-time product, 2) utilized original scan parameters, and 3) modeled scanner geometry including the tube start angle. Each image data set was ‘voxelized’ to create a patient-specific "virtual phantom": CT numbers were assigned a mass density and a molecular formula per ICRU 44. Absorbed dose values were computed for each voxel by simulated irradiation of the virtual phantoms to create individual dose maps. Dose maps were segmented by two consulting observers to determine patient-specific organ doses. Average EDs based on 1) weighted sum of organ doses (ICRP 60) and 2) ED= kchest x DLP were compared with a two-tailed paired Student's t-test. The mean of absolute percent differences between the two ED estimates was computed. EDdose map/EDk-factor ratios and BMI were plotted in Cartesian coordinates.

Average ED was not significantly different for the dose-map and k-factor methods (4.7 mSv vs. 4.3 mSv, p=0.76). However, EDs differed by an absolute mean percentage of 33.5%. ED ratios vs BMI showed a trend that ED was underestimated (overestimated) by the k-factor method with BMI < 28 kg/m2 (BMI > 28 kg/m2) relative to the dose map method.

A standard effective dose estimate based on a k-factor can be significantly different from an estimate based on a patient-specific dose map.

Insight gained from patient-specific dose maps may lead to more accurate dose metrics and reporting.

Thompson, C, Wiegert, J, Yanof, J, Yaddanapudi, K, Halliburton, S, Estimating Effective Dose for CT with Patient-specific Dose Maps in Comparison with DLP Conversion ("K Factor") Methods.  Radiological Society of North America 2012 Scientific Assembly and Annual Meeting, November 25 - November 30, 2012 ,Chicago IL. http://archive.rsna.org/2012/12043854.html