LAUSR

Objective This study aimed to calculate scanner-, kilovoltage peak (kVp)–, and patient size–specific computed tomography (CT) number thresholds for determining Agatston score (AgSc). Methods The proposed method was validated using calcium measurements in an anthropomorphic phantom for 4 CT scanners made by 4 vendors. The derived mass concentration (γ) thresholds were used to calculate kVp- and patient size–specific CT number thresholds. Two models were applied to reduce intrascanner and interscanner AgSc variation, respectively. Results The mean error of the modeled CT numbers is 1.8% (0.1%–4.4%). Model 1 has comparable results to the published phantom calibration method for an average-size patient (error, 1.5%; 0.1%–5.1%). The size- and the kVp-dependent fitting of modeled results have R2 greater than 0.965. Conclusions Our results show a potential to enable accurate determination of AgSc under diverse conditions (eg, reduced tube potential) and are more easily applicable to different patient sizes than the phantom calibration method.