Miyuki Takasu MD, Presenter: Nothing to Disclose

Yoko Kaichi, Abstract Co-Author: Nothing to Disclose

Chihiro Tani MD, Abstract Co-Author: Nothing to Disclose

Shuji Date, Abstract Co-Author: Nothing to Disclose

Masao Kiguchi RT, Abstract Co-Author: Nothing to Disclose

Kazuo Awai MD, Abstract Co-Author: Research Grant, Toshiba Corporation Research Grant, Hitachi Ltd Research Grant, Bayer AG Research Consultant, DAIICHI SANKYO Group Research Grant, Eisai Co, Ltd

Yoshiaki Kuroda, Abstract Co-Author: Nothing to Disclose

Akira Sakai, Abstract Co-Author: Nothing to Disclose

Sixty percent of myeloma patients develop pathologic fractures, with the majority occurring in the spine or ribs. Determining the nature of a spine fracture in myeloma patients may be difficult due to the presence of concomitant osteoporosis. The purpose of this study was to determine whether trabecular microstructural analysis can be used to predict new pathologic spine fractures in myeloma patients.

A total of 22 vertebral bodies from 14 patients with multiple myeloma (4 males; mean age, 64.3±4.1 years; 10 females; mean age, 61.2±7.9 years) were examined by 64-detector row computed tomography prior to follow-up CT (mean period: 10.9 months) which showed new pathologic spine fractures. Using a bone mineral calibration phantom and a 3-dimensional image analysis system, bone mineral density (BMD), trabecular parameters, and mechanical properties were calculated for three vertebrae comprising a vertebra that would become fractured and the two adjacent vertebrae. Areas of lytic lesions were also obtained using manually drawn regions of interest in the axial images containing the largest lytic lesions. Trabecular microstructural indices were expressed as ratios to mean values from the three vertebrae. For data analysis, univariate analysis was used to compare indices between vertebrae that would develop fracture and those that would not. Multivariate logistic regression analyses and receiver operating characteristic curves were also used. Values of P < .05 were considered significant.

Univariate analysis demonstrated that area of lytic lesion, trabecular spacing, structure model index, volumetric BMD (vBMD), failure load, and stiffness were significantly associated with occurrence of pathologic fracture. Multivariate analysis identified area of lytic lesion, vBMD, and failure load as significant predictors of pathologic fracture. Area under the curve was 0.779 for failure load, 0.741 for vBMD, and 0.632 for area of lytic lesion.

Trabecular microstructural analysis and finite element modeling can be used to predict new pathologic fractures in myeloma patients. Failure load and vBMD predict pathologic fracture better than existence of a lytic lesion in a vertebra.

Factors predicting pathologic fracture in myeloma include BMD and bone strength estimated by CT-based FEM. Trabecular structural analysis can be used to predict pathologic fracture in vertebrae.

Takasu, M, Kaichi, Y, Tani, C, Date, S, Kiguchi, M, Awai, K, Kuroda, Y, Sakai, A, Predicting Pathologic Spine Fracture in Multiple Myeloma Patients Using Trabecular Microarchitecture and CT-based Finite Element Analysis: A Longitudinal Study.  Radiological Society of North America 2014 Scientific Assembly and Annual Meeting, - ,Chicago IL. http://archive.rsna.org/2014/14012594.html