A feasibility study of quantitative molecular characterization of musculoskeletal lesions by proton MR spectroscopy at 3 T

Abstract

Objective: The purpose of this study is to establish the feasibility and potential value of measuring the concentration of choline-containing compounds by proton MR spectroscopy (MRS) in musculoskeletal lesions at 3 T.

Subjects and methods: Thirty-three subjects with 34 musculoskeletal lesions (four histologically proven malignant, 13 histologically proven benign or proven benign by follow-up analysis, and 17 posttreatment fibrosis with documented stability for 6-36 months) underwent single-voxel 3-T MRS studies. In each case, both water-suppressed and water-unsuppressed scans were obtained. The quality of the scans was recorded as excellent, adequate, or nondiagnostic, and the choline concentration was measured using water as the internal reference. The choline concentrations of benign and malignant lesions were compared using the Mann-Whitney test.

Results: Spectral quality was excellent in 26 cases, adequate in four cases, and nondiagnostic in four cases. For malignant lesions (three sarcomas), the choline concentrations were 1.5, 2.9, and 3.8 mmol/kg, respectively. For five benign lesions (two neurofibromas, two schwannomas, and one enchondroma), the choline concentrations were 0.11, 0.28, 0.13, 0.8, and 1.2 mmol/kg, respectively. For seven benign lesions (two hematomas, two bone cysts, one lipoma, one giant cell tumor, and one pigmented villonodular synovitis), the spectra showed negligible choline content. For three posttreatment fibrosis cases, the choline concentration range was 0.2-0.4 mmol/kg. For the remaining 12 posttreatment fibrosis cases, the spectra showed negligible choline content. Average choline concentrations were different for malignant and benign lesions (2.7 vs 0.5 mmol/kg; p = 0.01).

Conclusion: The measurement of choline concentration within musculoskeletal lesions by MRS is feasible using an internal water-referencing method at 3 T and has potential for characterizing lesions for malignancy.

Fig. 1

25-year-old woman with Ewing sarcoma of sacrum. A, Single-voxel proton MR spectroscopy performed within sacral mass, with weak water suppression, shows discrete choline peak (choline concentration, 2.9 mmol/kg). B, Axial T1-weighted MR image (spin echo; TR/TE, 690/15) shows left sacral mass as low-signal-intensity lesion. C, Axial fat-suppressed T2-weighted MR image (fast spin echo; 2,886/100) shows mass involving neural foramina and crossing sacroiliac joint. Anteriorly, mass is associated with soft-tissue component. Although mass has malignant features and was categorized as probably malignant by conventional MRI, diagnosis of giant cell tumor was also entertained given some characteristic features, such as crossing of sacroiliac joint. However, elevated choline concentration (2.9 mmol/kg) detected by MR spectroscopy points toward malignancy.

Fig. 2

64-year-old man with left thigh hematoma. Biopsy showed no evidence of malignant cells, and lesion subsequently decreased in size over time. A, Single-voxel proton MR spectroscopy obtained within mass, with weak water suppression, shows no evidence of discrete choline peak. B, Axial fat-suppressed T2-weighted image (fast spin echo; TR/TE, 2,886/100) shows indeterminate mass of slightly heterogeneous T2 signal with hemosiderin ring. C, Axial T1-weighted image (spin echo; 690/15) shows mass of increased T1 signal suggesting possibility of blood products, but with underlying tumor difficult to exclude. D, Axial contrast-enhanced T1-weighted image (gradient-refocused echo; 436/4.9) shows questionable contrast enhancement within mass, because it contains increased T1 signal on unenhanced imaging. Differential diagnosis includes hematoma as well as underlying malignant neoplasm. Negligible choline content may help guide diagnosis toward benign cause.