ABSTRACT
OBJECTIVE: Coronary artery calcification (CAC) is frequently detected on low-dose ungated MDCT performed for lung cancer screening. We aimed to determine the concordance of CAC scores on low-dose ungated and regular-dose ECG-gated MDCT. SUBJECTS AND METHODS: The subjects were 513 patients consecutively registered for health screening and undergoing both low-dose ungated (120 kVp, 20 mAs) and regular-dose ECG-gated MDCT (120 kVp, 150 mAs, retrospective ECG gating). The first 30 cases were used for protocol optimization and a training session. Agatston score on regular-dose ECG-gated and low-dose ungated MDCT in the other 483 cases (320 men; mean age, 62.2 +/- 13.2 [SD] years) was calculated by two observers in a blinded manner. Interobserver and intertechnique scoring variability and concordance were calculated. RESULTS: The mean of interobserver scoring variability for regular-dose ECG-gated MDCT was 3.6% and for low-dose ungated MDCT was 9.6%. Regular-dose ECG-gated MDCT depicted CAC in 221 (46%) of the subjects. With low-dose ungated MDCT, observers 1 and 2, respectively, had five and seven false-positive and five and four false-negative predictions. All the miscategorized scores were 12 or less. The negative predictive values of CAC on low-dose ungated MDCT were 98% and 99% for observers 1 and 2, respectively. For patients with CAC, the mean intertechnique scoring variability was 40-43%. For all 483 subjects, the intertechnique concordance of the four major score ranks (0, 1-100, 101-400, > 400) was high (kappa = 0.89 for the two observers). CONCLUSION: Low-dose ungated MDCT with an optimized protocol is reliable for prediction of the presence of CAC and categorization of the four major Agatston score ranks. This technique may be useful for coronary artery disease risk stratification of persons undergoing low-dose ungated MDCT for lung cancer screening.
