Purpose: To determine the detection of diffuse bone marrow infiltration with MRI in comparison with histopathological findings.
Materials and methods: MRI was performed on 45 patients with histologically proven multiple myeloma and on 30 healthy individuals. Three experienced radiologists read separately Tl-weighted SE sequences, STIR sequences and the combination of Tl-weighted SE and STIR sequences of the spine. Additionally, Tl-weighted SE sequences were obtained after gadolinium administration and the percentage increase in signal intensity was calculated. Bone marrow histology was used as gold standard for assessing the grade of infiltration. A dichotomous decision (infiltration yes/no) was made when assessing the MRI examinations.
Results: For the visual detection of diffuse infiltration, the best sensitivity was found with Tl-weighted SE sequences, achieving 71 % on average. The specificity was 89 %. The STIR sequences showed a sensitivity of 61 % and a specificity of 98 %, and the combination of Tl-weighted/STIR-sequences achieved a sensitivity of 65 % and a specificity of 94 %. In comparison with the histological findings, the sensitivity of the Tl-weighted sequences was 35 % for low-grade, 89 % for moderate and 100 % for high-grade infiltration. The application of contrast material with calculation of the percentage signal increase improved the detection by 7 %.
Conclusion: The sensitivity of the visual detection of diffuse multiple myeloma with unenhanced MRI is limited for low-grade or moderate infiltration, whereas the sensitivity for high grade infiltration is reliable. The specificity is high and the diagnostic confidence improves after application of contrast material with calculation of the percentage increase in signal intensity.