Interpreting angiograms: Docs err toward higher severity

Twitter icon
Facebook icon
LinkedIn icon
e-mail icon
Google icon
 - doctor looking at scan

Physicians who visually interpreted the severity of coronary stenosis tended to estimate diameter stenosis higher than assessments via quantitative coronary angiography (QCA) in a study that compared the two methods. But the 8.2 percent difference actually heralds an improvement, wrote editorialists.

The study and editorial were published in the April 30 issue of Circulation.

Studies in the 1990s that compared visual assessment for stenosis severity in coronary angiograms with computer-assisted techniques in candidates for revascularization found intraobserver and interobserver variability. In the past two decades, imaging technology has improved, wrote Brahmajee K. Nallamothu, MD, MPH, of the University of Michigan in Ann Arbor, and colleagues. But how that has affected interpretation is unclear. “Indeed, interpretation may be even more challenging today because more decisions about revascularization are made during or just after the procedure is performed to maximize efficiency and to minimize costs (i.e., ad hoc PCI). “

Nallamothu et al randomly selected coronary angiograms from seven hospitals that participated in the CathPCI registry of patients who underwent PCI in 2011. They compared the clinical interpretations of stenosis severity for 175 patients (228 lesions) with measurements made with QCA, a tool used in clinical trials to ensure quality. None of the patients received PCI for urgent or emergency indications.

Twelve of lesions were described qualitatively and were not included in the clinical interpretation group. Of the remaining PCI-treated lesions, the mean percent stenosis was 84.2 percent, with 70 to less than 90 percent the most common range, followed by 90 to 100 percent. Only 1.4 percent fell in the less than 70 percent range, and none were less than 50 percent.

In the QCA group, the mean percent stenosis was 76.1 percent, with 70 to less than 90 percent the most common range, followed by 50 to less than 70 percent. None were less than 50 percent.

The mean difference in percent stenosis between the two methods was 8.2 percent. Of the lesions estimated to be 70 percent or greater by clinical interpretation, 26.3 percent were measured to be less than 70 percent via QCA and 4.7 percent were less than 60 percent.

Nallamothu et al also noted that variation across hospital sites ranged from 5.6 percent to 11.2 percent.

The gap identified in the study underscores that there is room for improvement in clinical interpretation of angiograms, which is still the standard for selecting patients for revascularization. They added QCA in isolation would lack important information such as the hemodynamic significance of the stenosis that also factors into revascularization decisions.

“Although differences between the clinical interpretation and QCA in an isolated patient should never be considered an automatic flag for inappropriate PCI, identifying where inconsistencies exist may provide opportunities for clinicians to understand ways to improve,” they suggested. They recommended as examples feedback and further testing using fractional flow reserve—which was used only 16 times in the study—before PCI, when warranted.

They acknowledged that their study relied on high-volume hospitals that initially joined the study as a pilot improvement initiative, making the results possibly “a best-case scenario,” and the sample size was limited.

In an accompanying editorial, R. David Anderson, MD, MS, and Carl J. Pepine, MD, of the University of Florida in Gainesville, emphasized that compared to previous studies, the finding by Nallamothu et al showed an improvement in clinical interpretation.

“In another study, also in patients undergoing percutaneous coronary intervention, the mean percent diameter stenosis was 87.9 ± 9.9 percent when assessed visually and 64.6 ± 9.2 percent when measured by QCA. Thus, the mean difference of 8.2 percent from the present study seems small in comparison, and perhaps a goal of complete concordance is not realistic,” they wrote.

They added that the variation by institution might serve as a quality metric, with hospitals that strayed too far from the QCA mean becoming possible targets for oversight.