Radiology: Automation cuts plaque measurement time to 20 seconds

Researchers used an automated software program to accurately measure noncalcified coronary plaque volumes with CT in less than 20 seconds, a fraction of the time of an intravascular ultrasound measurement and five to 15 minutes less than a standard CT coronary plaque measurement, according to a study published in the October edition of Radiology.

Intravascular ultrasound (IVUS) is considered the most accurate method for coronary plaque quantification, but the procedure is invasive and includes additional risk factors such as allergic reaction and blood vessel damage.

Coronary CT angiography (CCTA) is an increasingly common alternative. However, CCTA lacks a validated approach for measuring coronary plaque and may be prone to intraobserver variability, explained the researchers.

Damini Dey, PhD, of Cedars-Sinai Medical Center and University of California Los Angeles and colleagues utilized their own interactive automated computer software (AUTOPLAQ, or APQ) to calculate the volumes of 22 noncalcified proximal to middle coronary plaques in CT studies of 20 patients.

All patients underwent IVUS, and within 48 hours of the IVUS study, patients were injected with intravenous contrast and underwent CCTA imaging.

Researchers quantified all plaque volumes using three methods: IVUS, manual CT measurement and the automated CT APQ algorithm. For automated quantifications, two reviewers marked distal and proximal limits of coronary plaque, circled a broader region of interest and marked lumen midlines. With these reference points, the APQ software automatically calculated plaque volumes.

The total “time for the automated plaque segmentation and quantification was less than 20 seconds,” they reported. IVUS plaque quantification took 15 to 35 minutes, and manual CT tracing took 5 to 15 minutes.

The mean plaque volumes calculated by reviewers using IVUS, manual CT and automatic CT were 105.9 mm³, 100.8 mm³ and 116.6 mm³, respectively. None of the variations in quantifications was statistically significant, and both automatic and manual CT measurements showed excellent correlation with the clinically accepted IVUS values.

Several factors may account for wide variations in IVUS versus manual and automate CT measurements, according to the authors. They include:

• Limited temporal and spatial resolution of CT scanners, making accurate delineation of the coronary vessel wall from surrounding epicardial fat challenging;
• IVUS excludes the adventitia, whereas CT does not;
• Misregistration of the landmarks, although landmarks "were carefully identified"; and
• Segmentation of the coronary arteries and quantification of plaque with APQ was achieved using a fully 3D approach, whereas for manual quantification with intravascular US and CT summed the plaque area.

The authors acknowledged that the study’s relatively small sample size (22 identified lesions) and exclusion of calcified coronary plaque presented limitations. Moreover, the authors only used good and excellent CT images, as they were unsure how accurately lower quality CT images could be measured using the automated program. 

The results indicate that automated CT measurement of noncalcified coronary plaque may provide an accurate, quick and noninvasive alternative to IVUS and manual CT quantification, concluded Dey.