Shilpy Chowdhury MD, MPH, Presenter: Nothing to Disclose

Ananth Madhuranthakam PhD, Abstract Co-Author: Employee, General Electric Company

Ramesh Venkatesan PhD, Abstract Co-Author: Employee, General Electric Company

Arvind Sonik MD, Abstract Co-Author: Nothing to Disclose

Peng Lai PhD, Abstract Co-Author: Employee, General Electric Company

Anja C. S. Brau PhD, Abstract Co-Author: Employee, General Electric Company

Shreyas Shreenivas Vasanawala MD, PhD, Abstract Co-Author: Research support, General Electric Company Consultant, Zimmer Holdings, Inc Consultant, ArthroCare Corporation

MRI for pediatric bone tumors and osteomyelitis requires joint-to-joint anatomic coverage, which can be time-consuming for long bones. This study assesses the potential of a fast single sequence MRI protocol based on a two-echo modified 2-point Dixon technique that yields volumetric (3D) fat, water, in-phase, and opposed-phase T2-weighted images.  

17 consecutive patients were identified retrospectively with lesions of a long bone that underwent conventional (longitudinal T1, short-axis T1 and fat-suppressed T2 sequences) and 2-point 3D MRI. Two radiologists independently assessed quality of fat suppression, lesion detection and extent, value of in/opposed phase images, joint/neurovascular involvement.

No case had significant fat suppression failure on 3D images, whereas minor fat suppression failure was seen in 18% (95% C.I. 0-38%). No case had major fat sat failure. Overall fat suppression was superior in 3D exams (binomial test, p <0.001), but no statistically significant difference in fat suppression near the lesion was demonstrated (p = .33). In/opposed phase imaging was useful in 47% (95% C.I. 21-73%). No bone lesion was seen by conventional imaging alone, and no significant difference in the longitudinal measurement of intraosseous bone lesion extent between two methods was demonstrated (p = .23, paired t-test). With conventional MRI as the gold standard for extraosseous extension of bone lesions, 3D had a sensitivity of 100%, specificity of 90% and high agreement with 2D (Cohen’s kappa = 0.88). No cases had joint or neurovascular involvement, so assessment of these features is limited. Mean total scan times for 2D sequences and 3D were 17.6 and 6.5 minutes, respectively

This preliminary study indicates a single-sequence technique may permit fast MRI evaluation of pediatric bone tumors and osteomyelitis in terms of lesion detection, characterization, and extent. Further study should be aimed at assessing accuracy of assessment of neurovascular and joint involvement.    

This preliminary study indicates a single-sequence technique may permit fast MRI evaluation of pediatric bone tumors and osteomyelitis in terms of lesion detection, characterization, and extent.

Chowdhury, S, Madhuranthakam, A, Venkatesan, R, Sonik, A, Lai, P, Brau, A, Vasanawala, S, Volumetric Fat-Water Separated MRI of Long Bones.  Radiological Society of North America 2011 Scientific Assembly and Annual Meeting, November 26 - December 2, 2011 ,Chicago IL. http://archive.rsna.org/2011/11012917.html