Differentiation of Schmorl Nodes From Bone Metastases of the Spine: Use of Apparent Diffusion Coefficient Derived From DWI and Fat Fraction Derived From a Dixon Sequence

Abstract

OBJECTIVE. The objective of our study was to evaluate the diagnostic performance of quantitative parameters derived from DWI and a Dixon MRI sequence in differentiating Schmorl nodes and bone metastases of the spine. MATERIALS AND METHODS. Fifty-one lesions (51 patients) were assessed with DWI and six-echo Dixon sequences for fat fraction (FF) quantification on 3-T MRI. The lesions were divided into two groups: Schmorl nodes (n = 22) and bone metastases (n = 29). The quantitative parameters of the lesions were obtained, including the apparent diffusion coefficient (ADC) value, FF, and FF ratio (defined as the FF of the lesion divided by the FF of normal marrow). The mean values of the parameters were compared between the two groups, and ROC curves were analyzed. For the ROC curves, AUCs were compared. RESULTS. The ADC value, FF, and FF ratio of bone metastases were significantly lower than those of Schmorl nodes (mean ADC value, 916.01 × 10-6 mm2/s vs 1569.78 × 10-6 mm2/s [p < 0.001]; mean FF, 2.89% vs 10.54% [p < 0.001]; mean FF ratio, 0.05 vs 0.21 [p < 0.001]). The ROC AUCs of the ADC value, FF, and FF ratio were 0.94, 0.95, and 0.97, respectively. The AUCs of all three imaging parameters did not show a statistically significant difference. CONCLUSION. ADC value, FF, and FF ratio can be useful for differentiating Schmorl nodes from bone metastases.