MR surpasses CT for detecting bone metastases

A study of whole-body MR and CT as possible alternatives to planar nuclear medicine bone scanning has found that whole-body MRI depicts spinal bone metastases that escape detection with 16-slice CT.

Dr. Andrea Baur-Melnyk, a radiology researcher at the Grosshadern Clinic of Ludwigs-Maximilian University in Munich, Germany, put the two modalities to the test in a retrospective study of 41 cancer patients and 782 vertebral bodies.

The 16-slice MSCT scanning was performed at 120 kV and 250 mAS with a collimation of 0.75. The images were reconstructed as 3-mm axial slices and with sagittal MPR. The MR protocol featured T1- and T2-weighted spin-echo and STIR sequences performed with and without gadolinium contrast. Two experienced radiologists reached a consensus in their interpretation of both sets of studies. They rendered their interpretations for each modality in sessions separated by three weeks to avoid bias.

MRI identified all 424 metastatic spinal lesions and resulted in four false-negative findings No false positives were reported, resulting in a sensitivity and accuracy rate of 99%. On the MSCT images, 195 metastatic lesions were read as normal, and four false positives arose. Although the specificity of MSCT was 98%, its specificity was only 55%, and its accuracy was 75% because of the misinterpreted scans.

MRI's superior performance was exemplified in a case of myeloma, Baur-Melnyk said. MRI revealed tumor infiltration in the thoracic vertebral body. Although no signal intensity was produced on T1 images, the lesion was evident from the high signal intensities characteristic of such metastatic disease on contrast-enhanced STIR images. No metastases or lesions could be identified on the corresponding MSCT study.

Although 16-slice MSCT permits thin collimation and a generally high-resolution display of bony structures, it still cannot match MRI for identifying spinal metastases, Baur-Melnyk said.