Characterizing and predicting pathologic spine fractures in myeloma patients with FDG PET/CT and MR imaging

To determine if FDG PET/CT scans can be used to discriminate between old and new pathologic spine fractures in myeloma patients and also to determine if there is any combination of PET/CT and MR imaging findings that may indicate an impending spine fracture.

Retrospective review, after IRB approval, by two musculoskeletal radiologists and one nuclear medicine physician was done. PET/CT database was searched to identify studies performed in myeloma patients, using the search terms myeloma, plasmacytoma, or plasma cell. Records, for this group of 176 patients, were searched for concurrent spine MR imaging and report of fracture. The SUV(max) was recorded for each vertebral level from C2 through L5. All other imaging studies carried out in each patient (spine radiographs, spine CTs, and spine MRIs) and clinical records also were reviewed. Spine fractures were identified using the Genant system. MR imaging findings noted included signal intensity, pattern of involvement, as well as the size and location, within the vertebral body, of individual focal lesions.

31 of the 176 patients had concurrent spine MR imaging and report of fracture. Five of these 31 patients were excluded because they had no documentation of the age of fracture prior to PET/CT imaging. The 26 remaining patients (598 vertebral levels) included 17 men and 9 women. There were a total of 59 PET/CTs, 104 spine X-ray studies, 25 spine CT exams, and 71 spine MRI exams. There were a total of 134 vertebral body fractures: 27 were determined to be acute/subacute and pathologic due to active myeloma, 1 was determined to be new but owing to simple osteoporosis, and 106 were determined to be old. The mean SUV(max) in the acute/subacute pathologic fractures was 2.9 with a range from 1.1 to 4.3 (standard deviation 0.98). The old fractures had a mean SUV(max) of 1.6 with a range from 0.6 to 3.1 (standard deviation 0.55). This was a statistically significant difference with p value <.0001. The 464 vertebral bodies without fracture had a mean SUV(max) of 1.8. SUV(max) ≥2.5 was seen at 49 vertebral levels, including 27 with a fracture and 22 without a fracture. SUV(max) ≥3.5 was found at 9 vertebral levels, including 6 with a fracture and 3 without a fracture. The combination of diffuse or multifocal MR patterns and SUV > 3.5 was seen at 7 levels, all but one with new pathologic fractures.

PET/CT SUV > 3.2 alone can be used in myeloma patients to discriminate between old and new pathologic fractures, just as it has been reported in other types of cancer patients. The combination of PET/CT SUV > 3.5 and MR findings of diffuse or multifocal vertebral body involvement seems to indicate an impending fracture.