Thin-cut coronary calcium quantification: Advantages compared with standard 3 mm slices

Research comparing the accuracy of three MDCT slice thicknesses has found that 3-mm slices underestimated coronary artery calcium (CAC) scores at every level of calcification. The inaccuracies were caused by partial volume averaging errors.

"Our analysis proved this concept and showed that CAC can be more accurately measured with 0.5 or 1 mm using isotropic data acquisition obtained by a volume scanner at identical radiation dose ," said Farhood Saremi, MD, University of Southern California. "Coronary artery calcium can be more accurately measured with 0.5- or 1-mm slices using isotropic data acquisition obtained by a volume scanner at the identical radiation dose.

The study is featured in an electronic exhibit at the ARRS 2015 Annual Meeting in Toronto.

Abstract

Cardiac Imaging

E4911. Thin-Cut Coronary Calcium Quantification: Advantages Compared With Standard 3-mm

Saremi A, Ng B, Lin L, Yong C, Saremi F. University of California, Los Angeles, Los Angeles, CA

Objective: Traditional 3-mm slice reconstructions for coronary artery calcium score (CAC) quantification may underestimate scores and lead to inaccurate measurements due to partial volume averaging errors. A 320-MDCT scanner is capable of quantifying CAC at different slice thicknesses from 0.5-mm source images without increased radiation. We took advantage of this capability to compare CAC at 0.5-, 1-, and 3-mm reconstructions.

Materials and Methods: We analyzed the CAC in 33 major vessel distributions (mean age, 65 ± 12.5 years). A 320-slice MDCT was used, and prospectively triggered unenhanced 0.5-, 1-, and 3-mm slices were reconstructed at mid diastole. Agatston score and volume of calcified plaques were calculated using standard semiautomated software at a 130-HU threshold. Wilcoxon signed rank tests were used to compare the pairwise difference among the three reconstructions. Spearman correlation and intraclass correlation were used to assess the interobserver variability. The Bonferroni adjustment was used for multiple comparisons.

Results: The median Agatston scores of 0.5-mm [45.5; interquartile range (IQR), 14-165] and 1-mm (67.5; IQR, 24-213) slices were significantly higher than those of 3-mm (71; IQR, 23-225) slices. The difference between 3 mm and 1 mm was 11.5 (IQR, 1.5-27.5; p < 0.01), between 3 mm and 0.5 mm was 12.5 (IQR, 3-30.5; p < 0.01), between 1 mm and 0.5 mm was 2 (IQR, 0-6; p < 0.01). At low CAC, 0.5-mm slices showed small calcified lesions that were not measurable by 3-mm slices. There was excellent correlation between the three techniques (R2 = 0.97-"0.99, p < 0.0001). Interobserver variability was 0.988. Dose-length product was 133 mGy · cm, and volume CT dose index was 8.3 mGy.

Conclusion: Three-millimeter slices underestimated CAC at every level of coronary calcifications. CAC can be more accurately measured with 0.5- or 1-mm slices using isotropic data acquisition obtained by a volume scanner at identical radiation dose.