Preoperative differentiation of spinal multiple myeloma and osteolytic metastasis using diffusion-weighted imaging–based habitat

Motivation: Multiple myeloma (MM) and osteolytic metastases (OM) are difficult to distinguish due to similar radiographic features and clinical symptoms, yet they require different treatments. Goal(s): Develop a non-invasive, preoperative method to differentiate MM from OM using diffusion-weighted imaging (DWI) habitat imaging. Approach: Study involved MM and OM patients undergoing multi-b DWI exams. Tumors were segmented into three habitats based on diffusion coefficient and mean kurtosis coefficient. Logistic regression identified risk factors, and nomogram models were developed. Results: Habitat 3 fraction was higher in MM, and habitat 2 fraction was lower, both significant predictors. The nomogram model showed the highest diagnostic accuracy. Impact: DWI-based habitat imaging shows clinical potential for noninvasively and preoperatively differentiating between spine MM and OM.