Poor inter-observer agreement in anatomical classifications of infrapopliteal arterial disease due to mandatory selection of only one target artery

BACKGROUND: Established anatomical classifications of infrapopliteal arterial lesion severity are based on assessment of only one target artery, not including all infrapopliteal arteries although multivessel revascularization is common. PURPOSE: To investigate the reproducibility of one of these classifications and a new aggregated score. MATERIAL AND METHODS: A total of 68 patients undergoing endovascular infrapopliteal revascularization at Sahlgrenska University Hospital during 2008–2016 were included. Preoperative magnetic resonance angiographies (MRA) and digital subtraction angiographies (DSA) were evaluated by three blinded observers in random order, using the infrapopliteal TransAtlantic Inter-Society Consensus (TASC) II classification. An aggregated score, the Infrapopliteal Total Atherosclerotic Burden (I-TAB) score, including all infrapopliteal arteries, was constructed and used for comparison. RESULTS: Inter-observer agreement on lesion severity for each evaluated artery was good; Krippendorff’s α for MRA 0.64–0.79 and DSA 0.66–0.84. Inter-observer agreement on TASC II grade, based on the selected target artery as stipulated, was poor; Krippendorff's α 0.14 (95% confidence interval [CI]=−0.05 to 0.30) for MRA and 0.48 (95% CI=0.33–0.61) for DSA. Inter-observer agreement for the new I-TAB score was good; Krippendorff's α 0.76 (95% CI=0.70–0.81) for MRA and 0.79 (95% CI=0.74–0.84) for DSA. CONCLUSION: Reproducible assessment of infrapopliteal lesion severity can be achieved for separate arteries with both MRA and DSA using the TASC II definitions. However, poor inter-observer agreement in selecting the target artery results in low reproducibility of the overall infrapopliteal TASC II grade. An aggregated score, such as I-TAB, results in less variability and may provide a more robust evaluation tool of atherosclerotic disease severity.