CT imaging of bone and bone marrow infiltration in malignant melanoma--Challenges and limitations for clinical staging in comparison to 18FDG-PET/CT.

Rationale of this study was the evaluation of the diagnostic value of computed tomography (CT) in the detection of bone marrow infiltration in comparison to PET/CT. Fifty patients (age 61 ± 15.12 years) with metastatic malignant melanoma underwent 18F-FDG-PET/CT, including contrast-enhanced CT. 2 readers evaluated the CT images in consensus for bone and bone marrow lesions focusing on lesion location, type and size. PET/CT was used as reference standard to estimate sensitivity, specificity, negative and positive predictive value. Moreover, the bone marrow density was estimated in the long bones and the sacral bone. Serum hamoglobin, thrombocyte level and S100 protein were correlated with the presence or absence of bone and bone marrow lesions. According to PET/CT as standard of reference, of 594 bone and medullary lesions 495 were considered malignant. Of these 77.8% were medullary, 20.4% lytic, 1% sclerotic and 0.8% mixed lytic/sclerotic. Contrast-enhanced CT yielded a lesion-based sensitivity of 36.8% and a specificity of 87.9% (PPV 93.8%; NPV 21.8%). Patient-based sensitivity and specificity were 78.8% and 82.4%, respectively. Of the missed lesions, most were medullary (95.8%). A disseminated bone marrow involvement (defined as >10 bone marrow lesions or diffuse infiltration of a whole body segment) was described in 11 cases, in 6 cases the disseminated involvement was underestimated or missed on CT. In cases with disseminated bone marrow involvement the bone marrow density was significantly higher in the humerus (p=0.04), but not in the femur or sacral bone (p=0.06). Multivariate analysis revealed no isolated effect of bone metastases on S100 serum and hemoglobin level, but both were significantly altered in patients with disseminated bone marrow involvement (p<0.05). In conclusion, the diagnostic value of computed tomography for the detection of bone marrow metastases in patients with melanoma, is limited. Especially in cases with disseminated bone marrow involvement about 50% of the cases were missed or underestimated.