Purpose or Learning Objective: Femoroacetabular Impingement (FAI) is a known cause of hip pain in young patients.
Femoral version (FV) is associated with extra-articular hip impingement. Measurement
of FV with magnetic resonance imaging (MRI) is controversial. Knee bolstering is commonly
used for MRI. We compared MRI- and computed tomography (CT)-based measurements of
FV and asked this question: What is the reliability in terms of mean difference and
correlation with and without bolstering, and also between two readers?
Methods or Background: A retrospective comparative radiologic study approved by the institutional review
board involving a total of 100 hips was performed. Of them, 52 hips (46 patients)
underwent standard MRI with knee bolstering; 48 hips had no knee bolstering and an
additional MRI sequence. All patients (100 hips) had symptomatic FAI (mean age: 28 ± 10
years). All patients underwent pelvic CT and MRI of the same hip joint (2016–2019).
Mean interval between CT and MRI was 20 days. In addition to the routine unilateral
multiplanar protocol for chondrolabral lesions, bilateral fast T1 volumetric interpolated
breath-hold examination (VIBE) Dixon of the pelvis and of the knee was acquired for
48 hips without bolstering to measure FV. Two readers independently measured FV on
both CT and MRI in two separate sessions (Murphy method).
Results or Findings: (1) Mean absolute difference between MRI with bolstering and CT-based measurements
of FV of 52 hips was 6.0 degrees ± 3 (− 10 to 7) for reader 1 and 8.4 degrees ± 5
(− 21 to 12) for reader 2. Mean absolute difference between MRI without bolstering
and CT-based measurements of FV of 48 hips decreased to 1.3 degrees ± 0.8 (0.2–3.0)
for reader 1 and 2.3 ± 1.9 degrees (0–10) for reader 2. Correlation of FV between
MRI without bolstering and CT-based measurements was r = 0.993 (p < 0.001) for reader 1 and r = 0.975 (p < 0.001) for reader 2. (2) Mean absolute difference of CT-based measurements of FV
between two readers was 2.9 degrees ± 2 (− 0 to 8), and correlation was r = 0.969
(p < 0.001). Mean difference of MRI-based measurements without bolstering of FV between
two readers was 3.3 degrees ± 3 (0–10), and correlation was r = 0.943 (p < 0.001).
Conclusion: MRI-based measurement of FV is as accurate and reliable as CT-based measurements
when using T1 VIBE Dixon and in patients with FAI. No bolstering during MRI decreased
measurement errors in FV, leading to fewer misdiagnoses of FV.
