Interoperator and intraoperator (in)accuracy of stent selection based on visual estimation

• Published in: Catheterization and cardiovascular interventions 2015-12, Vol.86 (7), p.1177-1183
• Main Authors: Campbell, Paul T., Mahmud, Ehtisham, Marshall, J. Jeffrey
• Format: Article
• Language: English
• Subjects: Adult; Aged; Clinical Competence; Coronary Angiography; Coronary Artery Disease - diagnostic imaging; Coronary Artery Disease - therapy; Coronary Vessels - diagnostic imaging; Humans; Judgment; lesion length; Middle Aged; Observer Variation; percutaneous coronary intervention; Percutaneous Coronary Intervention - instrumentation; Predictive Value of Tests; Prospective Studies; Prosthesis Design; quantitative coronary angiography; Reproducibility of Results; Stents; Visual Perception
• ISSN: 1522-1946; 1522-726X
• DOI: 10.1002/ccd.25780

Objectives The objectives of this study were to evaluate the ability of interventional cardiologists to accurately measure lesion length and select appropriate stents. Background Inaccurate measurement of lesion length during percutaneous coronary intervention (PCI) increases the risk of restenosis. Methods Interventional cardiologists (n = 40) evaluated 25 matched orthogonal angiographic images that were prescored using quantitative coronary angiography (QCA) by a core laboratory. Visual estimates of lesion length and stent length selection were compared to the maximum QCA value. A 2–4 mm stent overlap of both the proximal and distal lesion edges was considered to be optimal. Based on optimal stent overlap, accurate lesion lengths were those measured from −1 to +4 mm from the QCA. Likewise, appropriate stent lengths were those that measured between +4 mm to +8 mm from the QCA value. Five images were repeated to assess intrarater variability. Results Lesion length measurements were short and long for 51.1% (95% CI 47.6–54.6%) and 19.0% (95% CI 16.3–21.9%) of the images, respectively. Stent length selections that were short and long were recorded for 55.0% (95% CI 51.5–58.5%) and 22.8% (95% CI 19.9–25.8%) of the images, respectively. Intrarater variability evaluation indicated that 38.5% (95% CI 31.7–45.6%) of lesion length measurements and 37.5% (95% CI 30.8–44.6%) of stent length selections were >3 mm different between the first and second evaluation of repeated images. Conclusions Visual estimation of coronary lesion length has a high degree of variability, which may lead to inappropriate stent selection. Improving the accuracy of lesion length measurement may improve patient outcomes. © 2015 Wiley Periodicals, Inc.