# An Evaluation of Magnetic Resonance Imaging Dixon Sequence Fat–Water Separation Techniques (T2w Dixon FSTs) to Detect Dorso-Lumbar Structural Lesions in Patients with Axial Spondyloarthritis

## TL;DR

This study compares MRI techniques for detecting bone lesions in patients with axial spondyloarthritis and finds that Dixon sequence fat–water separation performs as well as traditional T1-weighted imaging.

## Contribution

Demonstrates that T2w Dixon FSTs can replace T1w images for evaluating structural lesions in axial spondyloarthritis.

## Key findings

- T2w Dixon FSTs showed high sensitivity (≥75%), specificity (100%), and accuracy (96%) for detecting structural lesions.
- Diagnostic performance of T2w Dixon FSTs was comparable to T1w images for evaluating fatty lesions, erosions, sclerosis, and ankylosis.
- Findings suggest T1-weighted images may be unnecessary when using multi-parameter Dixon sequences.

## Abstract

Objective: The aim of this study was to assess and compare the diagnostic accuracy of two MRI techniques for identifying structural bone lesions (fatty lesions [BMFs], subchondral erosions [BMEs], sclerosis [BMS], and ankylosis [A]) in the dorso-lumbar joints. This assessment specifically focused on the application of MRI Dixon sequence fat–water separation techniques (T2w Dixon FSTs) when acquiring T1-weighted (T1w) images as the reference standard, among patients diagnosed with axial spondyloarthritis (SpA). Methods: Conducted at a single center, this study involved the recruitment of patients who underwent both spinal and sacroiliac (SI) joint MRI between 2019 and 2022, with follow-up continuing until 2023. In 2023, three independent readers reassessed the initial MRI datasets to evaluate specific radiological features of SpA. They recorded confidence estimates regarding the use of T2w Dixon FSTs when acquiring T1w images. The centralized MRI interpretations were then compared to established rheumatologic diagnoses. Results: A total of 73 patients (42 men and 31 women) were included in the study. The mean sensitivity, specificity, and accuracy of T2w Dixon FSTs (fat-only images) were at least 75%, 100%, and 96%, respectively, based on the 2023 assessment for all considered items. The diagnostic performance of T2w Dixon FSTs was comparable to that of T1w images. Conclusions: The diagnostic performance of T2w Dixon FSTs (fat-only images) matched that of T1w images not only in assessing structural and fatty lesions, but also in the evaluation of subchondral erosions, sclerosis, and ankylosis in the dorso-lumbar joints of patients with a rheumatologic diagnosis of SpA. These findings suggest the potential avoidance of T1-weighted images when employing multi-parameter, multi-sequence imaging, such as the Dixon sequence.