Evaluation of a CT-like Multi-Echo 3D Gradient-Echo MR Sequence (FRACTURE) for Structural Assessment of the Sacroiliac Joints in Spondyloarthritis: Comparison with CT and T1-Weighted MR Imaging

CT remains the imaging reference standard for evaluating structural sacroiliac joint (SIJ) lesions in spondyloarthropathy (SpA), but its use is limited by ionizing radiation. This retrospective single-center study evaluated a CT-like multi-echo 3D gradient-echo MRI sequence (FRACTURE) for structural SIJ assessment by comparing it with CT and conventional T1-weighted MR imaging (T1-WI). Consecutive adults who underwent SIJ MRI between November 2023 and October 2024 were included. Four radiologists assessed inflammatory sclerosis, erosions, and joint space changes using a structural severity score (SSS) derived from the SIMACT study and assigned a three-point Likert score for SpA diagnosis. Forty-three patients (median age, 49 years), including 22 with SpA, were analyzed. The area under the ROC curve for diagnostic Likert scores was 0.842 for T1-WI, 0.792 for CT, and 0.767 for FRACTURE (all pairwise comparisons, p > 0.05). Considering only definite diagnoses (Likert = 2), diagnostic accuracy reached 81.4% for CT, 76.7% for T1-WI, and 69.8% for FRACTURE, with 100% specificity for all three techniques. FRACTURE detected additional lumbosacral neojoints and accessory articular facets compared with T1-WI and showed a stronger correlation with CT for structural severity assessment (ρ = 0.814 vs. ρ = 0.774). Intra- and inter-observer reproducibility were good (all ICCs > 0.700). In conclusion, FRACTURE provides CT-like structural information with good reproducibility and showed a numerically higher correlation with CT for overall structural severity than T1-WI, while also improving the depiction of certain osseous anatomical variants relevant to differential diagnosis. However, it did not improve diagnostic performance for SpA diagnosis and showed lower performance for erosion assessment. FRACTURE may therefore provide complementary added value within MRI-based SIJ assessment, although further prospective validation is warranted.