Objective: To determine its ability of in-phase (IP) and out-of-phase (OOP) chemical shift imaging (CSI) to distinguish non-neoplastic marrow lesions, benign bone tumours and malignant bone tumours.
Methods: CSI was introduced into our musculoskeletal tumour protocol in May 2018 to aid in characterisation of suspected bone tumours. The % signal intensity (SI) drop between IP and OOP sequences was calculated and compared to the final lesion diagnosis, which was classified as non-neoplastic (NN), benign neoplastic (BN) or malignant neoplastic (MN).
Results: The study included 174 patients (84 males; 90 females: mean age 44.2 years, range 2-87 years). Based on either imaging features (n = 105) or histology (n = 69), 44 lesions (25.3%) were classified as NN, 66 (37.9%) as BN and 64 (36.8%) as MN. Mean % SI drop on OOP for NN lesions was 36.6%, for BN 3.19% and for MN 3.24% (p < 0.001). The sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and diagnostic accuracy of CSI for differentiating NN from neoplastic lesions were 65.9%, 94.6%, 80.6%, 89.1%% and 87.4% respectively, and for differentiating BN from MN were 9.1%, 98.4%, 85.7%, 51.2 and 53.1% respectively.
Conclusion: CSI is accurate for differentiating non-neoplastic and neoplastic marrow lesions, but is of no value in differentiating malignant bone tumours from non-fat containing benign bone tumours.
Advances in knowledge: CSI is of value for differentiating non-neoplastic marrow lesions from neoplastic lesions, but not for differentiating benign bone tumours from malignant bone tumours as has been previously reported.