Diagnostic Utility of 3D Gradient-Echo MR Imaging Sequences through the Filum Compared with Spin-Echo T1 in Children with Concern for Tethered Cord

BACKGROUND AND PURPOSE: Fatty intrathecal lesions are a cause of tethered cord, and detection of these on spinal MR imaging is paramount.

Conventional T1 FSE sequences are the mainstay of detecting fatty elements; however, 3D gradient-echo MR images, volumetric interpolated breath-hold examination/liver acquisition with volume acceleration (VIBE/LAVA), are popular, given the increased motion resistance.

We sought to evaluate the diagnostic accuracy of VIBE/LAVA compared with T1 FSE for detection of fatty intrathecal lesions.

MATERIALS AND METHODS: In this retrospective, institutional review board–approved study, 479 consecutive pediatric spine MRIs obtained to evaluate cord tethering between January 2016 and April 2022 were reviewed.

Inclusion criteria were patients who were 20 years of age or younger who underwent spine MRIs containing both axial T1 FSE and VIBE/LAVA sequences of the lumbar spine.

The presence or absence of fatty intrathecal lesions was recorded for each sequence.

If fatty intrathecal lesions were present, anterior-posterior and transverse dimensions were recorded.

VIBE/LAVA and T1 FSE sequences were evaluated on 2 separate occasions (VIBE/LAVAs first followed by T1 FSE several weeks later) to minimize bias.

Basic descriptive statistics compared fatty intrathecal lesion sizes on T1 FSEs and VIBE/LAVAs.

Receiver operating characteristic curves were used to determine minimal fatty intrathecal lesion size detectable by VIBE/LAVA.

RESULTS: Sixty-six patients were included, with 22 having fatty intrathecal lesions (mean age, 7.2 years).

T1 FSE sequences revealed fatty intrathecal lesions in 21/22 cases (95%); however, fatty intrathecal lesions on VIBE/LAVA were detected in 12/22 patients (55%).

Mean anterior-posterior and transverse dimensions of fatty intrathecal lesions measured larger on T1 FSE compared with VIBE/LAVA sequences (5.4 × 5.0 mm versus 1.5 × 1.6 mm, respectively; P values = .039 anterior-posterior; .027 transverse).

CONCLUSIONS: While T1 3D gradient-echo MR images may have decreased the acquisition time and are more motion-resistant than conventional T1 FSE sequences, they are less sensitive and may miss small fatty intrathecal lesions.