Developing a calibration method to utilize low-dose chest CT for assessment of coronary artery calcification score.

PURPOSE: This study aimed to develop a calibration approach for low-dose CT (LDCT) used in lung cancer screening to enable coronary artery calcium (CAC) scores that reflect those obtained from conventional electrocardiogram (ECG)-gated calcium scoring CT (CSCT).

METHODS

An electron density phantom-based calibration curve was developed to align Hounsfield unit (HU) values measured on LDCT with those from ECG-gated CSCT. To evaluate the efficacy of the proposed method, 139 patients undergoing physical examination who received both LDCT and CSCT on the same day were retrospectively enrolled. Agatston and volume scores were quantified for the right coronary artery, left anterior descending artery, left circumflex artery, and left main artery. The Bland-Altman analysis and two-sample t-test were performed to compare CAC scores measured on CSCT and LDCT before and after HU-based calibration.

RESULTS

Compared with the mean bias observed in Bland-Altman analysis between CSCT and LDCT, the mean bias between CSCT and calibration LDCT (cLDCT) was substantially reduced. Based on two-sample t-test, statistically significant differences in total Agatston scores and total volume scores calculated were observed between CSCT and LDCT for the four coronary arteries. In contrast, no significant differences were found between CSCT and cLDCT. Among patients undergoing diastolic reconstruction in CSCT, the p-values comparing CSCT and cLDCT were 0.94 for the Agatston score and 0.43 for the volume score, whereas corresponding p-values of 0.95 and 0.63 were observed in patients undergoing systolic reconstruction in CSCT.

CONCLUSION

The proposed calibration approach enables consistent HU mapping across protocols and supports simultaneous assessment of coronary calcium burden and pulmonary pathology on a single LDCT scan.