Reliability of Contrast-Enhanced MR Angiography in Evaluation of Nitinol Stent Restenosis in Iliac Arteries

In this observational study, contrast-enhanced MR angiography (MRA) was evaluated against digital subtraction angiography (DSA) in the stented segments of the iliac arteries to determine MRA's accuracy in diagnosing restenosis following stent placement. Six radiologists independently evaluated the degree of restenosis in the MRA and DSA images of 87 nitinol stents implanted in 53 patients with peripheral arterial disease. The differences in restenosis assessment between these two modalities, along with inter-observer agreement, were analysed using DSA as a reference. Discrepancy between methods was quantified by the MRA-DSA difference in % restenosis. Correlation, regression, and agreement analysis tools were used for evaluation. A significant mean MRA-DSA difference of 14.0% (P < 0.001) in the assessment of restenosis severity between the MRA and DSA was found. In isolated cases, the discrepancy reached 50–60%. The discrepancy negatively correlated with the degree of restenosis, showing a systematic proportional bias. The agreement analysis revealed a good-to-excellent inter-observer reliability measured by the intraclass correlation coefficient ICC. In general, the DSA observers were more similar and consistent in their ratings compared to the MRA observers (0.876 and 0.651, respectively). Diagnostic performance of the MRA in detecting high-grade (≥ 50%) restenosis yielded a sensitivity of 78.57% and a specificity of 67.12%. MRA showed an acceptable-to-good sensitivity and specificity for diagnosing high-grade restenosis after placement of nitinol stent in iliac arteries, though it significantly overestimated lumen diameter narrowing at low-grade restenosis.