Comparison of Magnetic Resonance Angiography and Digital Subtraction Angiography for the Assessment of Infrapopliteal Arterial Occlusive Lesions, Based on the TASC II Classification Criteria

Abstract

This paper aimed to study the agreement and repeatability, both intra- and interobserver, of infrapopliteal lesion assessment with magnetic resonance angiography (MRA), using the TransAtlantic Inter-Society Consensus (TASC) II criteria, with perioperative digital subtraction angiography (DSA) as a reference. Sixty-eight patients with an MRA preceding an endovascular infrapopliteal revascularization were included. Preoperative MRAs and perioperative DSAs were evaluated in random order by three independent observers using the TASC II classification. The results were analyzed using visual grading characteristics (VGC) analysis and Krippendorff’s α. No systematic difference was found between modalities: area under the VGC curve (AUCVGC) = 0.48 (p = 0.58) or intraobserver; AUCVGC for Observer 1 and 2 respectively, 0.49 (p = 0.85) and 0.53 (p = 0.52) for MRA compared with 0.54 (p = 0.30) and 0.49 (p = 0.81) for DSA. Interobserver differences were seen: AUCVGC of 0.63 (p < 0.01) for DSA and 0.80 (p < 0.01) for MRA. These results were confirmed using Krippendorff’s α for the three observers showing 0.13 (95% confidence interval (CI) −0.07–0.31) for MRA and 0.39 (95% CI 0.23–0.53) for DSA. Poor interobserver agreement was also found in the choice of a target vessel on preoperative MRA: Krippendorff’s α = 0.19 (95% CI 0.01‒0.36). In conclusion, infrapopliteal lesions can be reliably determined on preoperative MRA, but interobserver variability regarding the choice of a target vessel is a major concern that appears to affect the overall TASC II grade.

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

“…In a previous study, we investigated the agreement on infrapopliteal lesions between two imaging modalities, magnetic resonance angiography (MRA) versus digital subtraction angiography (DSA). We found good agreement between modalities but observed a large inter-observer variability regarding the choice of a target artery (6). If the choice of target artery has low reproducibility, then the scientific value of comparative studies based on the established classifications will also be low.…”

Comparison of image quality assessments between interventional radiographers and interventional radiologists using digital subtraction angiography

“…Studies have previously used VGC analysis for extremity computed tomography (CT), [16][17][18] chest, appendicular, and spine radiography, [19][20][21][22][23][24] abdominal CT, [25][26][27] head CT, 28,29 pelvic CT, 30 propagation-based phase-contrast CT, 31 pediatric radiograph, 32 and magnetic resonance imaging (MRI) olfactory bulb volumetry sequences. 33 Fridh et al 10 used angiographic images from lower limb DSA runs to detect infrapopliteal lesions, where the observers were interventional radiologists or vascular surgeons. This differentiates from our study as we compared two groups of observers who are partnered clinicians in the Interventional Catheter Laboratory when performing these procedures in a large private hospital setting.…”

“…Studies have previously used VGC analysis for extremity computed tomography (CT), 16 – 18 chest, appendicular, and spine radiography, 19 – 24 abdominal CT, 25 – 27 head CT, 28 , 29 pelvic CT, 30 propagation-based phase-contrast CT, 31 pediatric radiograph, 32 and magnetic resonance imaging (MRI) olfactory bulb volumetry sequences 33 . Fridh et al 10 . used angiographic images from lower limb DSA runs to detect infrapopliteal lesions, where the observers were interventional radiologists or vascular surgeons.…”

“…Interobserver differences were seen, where AUC VGC ¼ 0.63 (P < 0.01) for DSA and 0.80 (P < 0.001) for MRA. 10 However, this study was using VGC analysis to assess four independent observers who were interventional radiologists or vascular surgeons and did not include interventional radiographers as an observer group.…”

“…An AUCVGC of 0.5 indicates that the assessment for the two groups is similar and does not determine if the quality of the two groups is high or low. Fridh et al 10 . compared magnetic resonance angiography (MRA) and DSA for the assessment of infrapopliteal arterial occlusive lesions, using VGC analysis to determine the agreement of observers in evaluating the MRAs and DSAs using the TransAtlantic Inter-Society Consensus II classification.…”