[Diffusion weighted MRI of spine tumors]

To evaluate the contribution of diffusion weighted MR imaging in malignant spine pathology. Materials and methods. Between February 2004 and January 2005, 49 patients (43 to 86 years old) were included. Three groups were made: osteoporotic collapses (n = 13), malignant collapses (n = 15) and malignant spine lesions (n = 21). The MRI (Symphony 1.5T) allowed SENSE imaging. After conventional MRI examination (T1, T2 fat sat, T1 with Gadolinium), all patients underwent diffusion weighted imaging (Spin Echo) with variable b values: 0, 250, 500, 750 and 1000. The diffusion sequence lasted 2 min 29 s. The Apparent Diffusion Coefficient (ADC) was calculated automatically. The analysis was qualitative (signal study b = 1,000 mm2/s) and quantitative (ADC measurement).

The image quality was good except for some cervical examinations. Qualitative analysis did not show a difference between benign and malignant lesions. Quantitative results are: malignant spine lesion (mean ADC = 0.826 10-3 s/mm2), malignant spinal collapses (mean ADC = 0.912 10-3 s/mm2) and benign spinal collapses (mean ADC = 1.497 10-3 s/mm2). There was overlapping results between benign and malignant lesion. The statistical study showed a significant difference (t test with p < 1/10 000). For an ADC threshold value of 1.089 (malignant lesion ADC < 1.089), ROC curve showed a specificity = 80% and a sensitivity = 83.3%.

Performing diffusion weighted imaging of the spine is easy with new MR technology. The ADC measurement of spine lesion provides important additional information, but does not serve as a substitute for the routine MRI sequences. In the future, it could become an important point in this difficult diagnosis.