Can we distinguish between benign versus malignant compression fractures of the spine by magnetic resonance imaging?

Abstract

Study design: The authors investigate the usefulness of magnetic resonance imaging in differentiating benign versus malignant compression fractures by reviewing patients and a fracture model in a canine model.

Objectives: To determine the sensitivity and specificity of magnetic resonance imaging in differentiating benign versus malignant compression fractures of the spine and to obtain distinguishing features in magnetic resonance imaging.

Summary of background data: The differentiation between benign and abnormal compression fractures of the thoracolumbar spine has important implications regarding patient treatment and prognosis. Plain radiographs, bone scans, and computed tomography are not accurate imaging modalities for this purpose.

Methods: Magnetic resonance imaging scans of 22 patients with confirmed lesions of the thoracolumbar spine were studied. There were 11 malignant and 11 benign lesions. Two experienced neuroradiologists blindly reviewed the magnetic resonance imaging scans and determined benign or malignant lesions. A canine study was performed to simulate a compression fracture model with a vertebral osteotomy in two dogs, and serial contrast-enhanced magnetic resonance imaging scans were performed 15, 30, 60 and 90 days after surgery.

Results: The correct interpretation between two neuroradiologists was 77% and 95%. The combined sensitivity rate was 88.5%, and the specificity rate was 89.5%. Magnetic resonance imaging reliably distinguished benign versus malignant lesions based on the anatomic distribution and intensity of signal changes of bone and adjacent tissues, contrast enhancement characteristics, and changes over time. Only one malignant lesion was misinterpreted by both neuroradiologists as benign, whereas there was one additional missed malignant lesion and three misinterpreted benign lesions by one radiologist. In the canine study, signal changes and enhancement were found 60 days after surgery, but no signal changes or enhancement were noted on the scan 90 days after surgery.

Conclusions: Magnetic resonance imaging scans can detect malignant vertebral lesions early, but acute healing compression fractures may mimic the findings of metastatic lesions. The use of contrast-enhanced magnetic resonance imaging scans and serial magnetic resonance imagings are helpful for additional differentiation between benign and malignant compression fractures. In addition to magnetic resonance imaging scans, other diagnostic tests and clinical findings should be correlated before biopsy or surgery of the suspected lesion.