Purpose: To evaluate the accuracy of trabecular analysis of the lumbar spine by using multidetector computed tomography (CT) in differentiating multiple myeloma patients without spinal fractures from control subjects.
Materials and methods: This retrospective single-institution study was approved by the institutional review board, with a waiver of informed consent. Spinal microarchitecture was examined by using 64-detector row CT in 29 patients with multiple myeloma and 51 control subjects. The patients were 13 men and 16 women (mean ages, 71.1 and 69.2 years, respectively). By using multidetector CT data, eight parameters were calculated for the L3 vertebral trabeculae with a three-dimensional image analysis system. The χ(2) test was used to select a preliminary set of predictors for multiple myeloma. A multivariable generalized linear model was constructed to identify parameters that could be used to differentiate between patients and controls. Parameters with findings of P < .05 were included in the multivariable model.
Results: Generalized linear models showed that mean trabecular thickness (TbTh) (patients, 703.7 μm ± 46.7 [standard deviation]; controls, 661.1 μm ± 35.7) and fractal dimension (FD) (patients, 2.23 ± 0.24; controls, 2.41 ± 0.15) in men (P = .05 and .03, respectively) and degree of anisotropy (DA) (patients, 1.63 ± 0.23; controls, 1.38 ± 0.10) in women (P = .02) had significant effects in differentiation between patients and controls.
Conclusion: Significant differences were observed in spinal microarchitecture between control subjects and multiple myeloma patients without pathologic fractures. Independent predictors of multiple myeloma included TbTh and FD in male patients and DA in female patients.
© RSNA, 2011.