Compressed Sensing SEMAC MRI of Hip, Knee, and Ankle Arthroplasty Implants: A 1.5-T and 3-T Intrapatient Performance Comparison for Diagnosing Periprosthetic Abnormalities.

Abstract

Background. The utility of 3-T MRI for diagnosing joint disorders is established, but its performance for diagnosing abnormalities around arthroplasty implants is unclear. Objective. This study compared 1.5-T and 3-T compressed sensing slice encoding for metal artifact correction (CS-SEMAC) MRI for diagnosing periprosthetic abnormalities around hip, knee, and ankle arthroplasty implants. Methods. Forty-five participants (mean age ± SD, 71 ± 14 years; 26 women, 19 men) with symptomatic lower extremity arthroplasty (hip, knee, and ankle, 15 each) prospectively underwent consecutive 1.5- and 3-T MRI with intermediate-weighted (IW) and STIR CS-SEMAC sequences. Using a Likert scale, three radiologists evaluated presence or absence of periprosthetic abnormalities, including bone marrow edema-like signal, osteolysis, stress reaction/fracture, synovitis, tendon abnormalities and collections, image quality, and visibility of anatomic structures. Statistical analysis included nonparametric comparison and interchangeability testing. Results. For diagnosing periprosthetic abnormalities, 1.5-T and 3-T CS-SEMAC MRI were interchangeable. Across all three joints, 3 T produced lower noise than 1.5 T (median IW/STIR score at 3 T vs 1.5 T, 4/4 [range, 2-5/3-5] vs 3/3 [2-5/2-4]; p < .01 for both), sharper edges (4/4 [both 2-5] vs 3/3 [2-4/2-5]; p < .02/p < .05), and more effective metal artifact reduction (4/4 [3-5/2-5] vs 4/4 [both 3-5]; p < .02/p = .72). Agreement was moderate to substantial for image contrast (IW/STIR, 0.66/0.54 [95% CI, 0.41-0.91/0.29-0.80]; p = .58/p = .16) and joint capsule visualization (0.57/0.70 [0.32-0.81/0.51-0.89]; p = .16/p = .19). The bone-implant interface was more visible at 1.5 T (median IW/STIR score, 4/4 [range, both 2-5] at 1.5 T vs 3/3 [both 2-5] at 3 T; p = .08/p = .58), but periprosthetic tissues had superior visibility at 3 T (4/4 [both 3-5] at 3 T vs 4/4 [2-5/3-5] at 1.5 T; p = .07/p = .19). Conclusion. Optimized 1.5-T and 3-T CS-SEMAC MRI are interchangeable for diagnosing periprosthetic abnormalities, although metallic artifacts are larger at 3 T. Clinical Impact. With CS-SEMAC MRI, lower extremity arthroplasty implants can be scanned at 3 T rather than 1.5 T.