Prediction of Long Bone Fractures in Multiple Myeloma Patients in an Advanced Imaging World

Introduction: Bone disease occurs in 90% of multiple myeloma patients. A highly active lesion is locally destructive to the bone. The advent of hybrid PET/CT scanners has offered metabolic and morphological information. There is no published study that correlates these two factors to long bone fractures in multiple myeloma (MM) patients.

Methods: We retrospectively reviewed the medical records of all bone marrow biopsy proven multiple myeloma patients with long bone lesion from 1.1.2010-12.31.2015 at a single institution. Data collection included anatomical location of the lesion, PET/CT SUVmax, cortical involvement, Mirel classification (MC), fractures, and indication for fracture. The relationship between SUVmax and other variables were calculated with P value<0.05.

Results: 119 patients (59 males/60 females) were identified with 256 long bone lesions. Mean age at diagnosis was 58. Most lesions were in the femur (n = 150, 59%). 13 lesions (10 patients) required surgery for impending (n = 9) or actual fracture (n = 4). Mean MC score was 10 in fracture group vs 7 in non-fractures. Higher median SUVmax was measured for those with cortical involvement (8.05, range 0-50.8) vs no involvement (5.0, range 2.1-18.1). Those with cortical involvement had 34 times the odds for developing a fracture (p = 0.0018). Higher median SUV uptake was measured in fractured vs non-fractured lesions (6.9 vs 5.4, p-value = 0.38). No significant correlation was found between SUVmax and pain or fracture (p = 0.43).

Discussion: With the use of routine PET/CT scan in MM patients, additional information as SUVmax and better imaging may guide the surgeon for surgical treatment. The strongest predictors for fracture we found were cortical involvement and Mirel classification. We have not found the SUVmax to be a significant direct predictor for fracture of long bones however; a larger study may show a more significant prediction.