MRI versus

purposeTo compare the diagnostic performance of MRI and MATERIALS AND

methodsThis retrospective study was approved by our Institutional Review Board. Two radiologists and two nuclear medicine specialists independently and blindly reviewed 84 pairs of MRI and PET/CT scans obtained in 73 MM patients. Readers assessed the presence and patterns of BMI. The best valuable comparator (BVC) for BMI was established by a panel review of all baseline and follow-up imaging, and biological and pathological information. Intra- and inter-reader agreement and correlation between MRI and PET/CT were assessed using the prevalence-adjusted bias-adjusted kappa (k) coefficient. Diagnostic performance of MRI and PET/CT in detecting BMI was evaluated from ROC characteristics. Association between imaging and biological, pathological, and clinical findings was assessed using Wilcoxon rank-sum and chi-square tests.

resultsIntra- and inter-reader agreement was very good for MRI (k = 0.90 [0.81; 1.00] and 0.88 [0.78; 0.98]). Intra- and inter-reader agreement was good for PET/CT (k = 0.80 [0.69; 0.91] and 0.71 [0.56; 0.86]). The sensitivity of MRI to detect BMI (97% [90%; 100%]) was significantly superior to that of PET/CT (76% [64%; 85%]) (p < 0.001). The specificity of MRI (86% [57%; 98%]) was lower than that of PET/CT (93% [66%; 100%]), without reaching statistical significance (p = 0.32). There was a strong correlation between decisions regarding patient management and PET/CT findings (p < 0.001).

conclusionMRI is significantly more sensitive than PET/CT to detect BMI in MM. Patient management is more strongly correlated with PET/CT findings. KEY POINTS: • MRI and PET/CT have very close diagnostic value for the detection of bone marrow involvement in multiple myeloma. • MRI has a significantly higher sensitivity and better reproducibility. • PET/CT findings appear to have a higher impact on clinical decisions.