Multiparametric spinal MRI for differentiating benign and malignant vertebral compression fractures

This study aimed to assess the efficacy of multiparametric spinal MRI, utilizing conventional techniques, dynamic contrast-enhanced MRI, and diffusion MRI, in distinguishing between benign and malignant vertebral compression fractures. This retrospective study included 42 consecutive patients who underwent biopsy and pre-biopsy multiparametric spinal MRI. Patients under 18 or with non-diagnostic imaging were excluded. Images were evaluated blinded by two radiologists. Quantitative DWI and DCE-MRI parameters were analyzed using the Tofts model. Statistical tests included ANOVA, Mann–Whitney U, chi-square, and ROC analysis. A total of 50 vertebral fractures from 42 patients (18 men, 24 women; mean age 55.3 ± 3.3 years; range 31–82) were included in the study. Among these lesions, 23 were classified as malignant and 27 as benign based on histopathological analysis. A statistically significant difference was found between benign and malignant groups in terms of in perfusion MR curve type, volume transfer constant (Ktrans), flux rate constant (Kep), and apparent diffusion coefficient (ADC) parameters (p = 0.000–0.02). In the ROC analysis, the highest sensitivity and specificity values were 73.9% and 77.8% for Ktrans, 88% and 77.8% for Kep, and 63% and 87% for ADC in distinguishing between benign and malignant vertebral compression fracture. In the multiparametric evaluation made according to ADC, Ktrans, and Kep cut-off values, the highest sensitivity and specificity values were 91.30% and 96.29%, respectively. Combining the information of quantitative ADC, Kep, and Ktrans analysis, multiparametric spinal MRI can differentiate benign from malignant compression fracture with high diagnostic accuracy. In this way, the need for invasive procedures to be performed only for diagnostic purposes can be reduced.