Usefulness of radiography in differentiating enchondroma from central grade I chondrosarcoma

OBJECTIVE. The purpose of this study was to evaluate clinical symptoms and radiographic features that allow radiologists to differentiate between enchondroma and central grade 1 chondrosarcoma Such differentiation is important because of differences in management. MATERIALS AND METHODS. Clinical symptoms and location and size of 35 enchondromas and 43 central grade 1 chondrosarcomas were analyzed. Radiographic features were assessed independently by three observers. The chi-square test and linear discriminant analysis were used to identify features with discriminating strength. Kappa values were calculated to validate the consistency of observations among observers. A consensus diagnosis made by histology and long-term follow-up was used as the standard. RESULTS. No statistically significant correlation was found between clinical symptoms and the benign or malignant nature of the neoplasms. Grade 1 chondrosarcomas were more likely to be found in the axial skeleton and in flat bones. Also, chondrosarcomas were significantly larger than enchondromas (p < .001). Ill-defined margins and lobulated contours were the only morphologic features seen on radiographs that allowed significant discrimination (p = .004 and .009, respectively). An optimal combination of four radiographic features still left 72 of the 78 lesions with a 10-90% probability of malignancy, indicative of poor discriminating power. Kappa values generally showed poor to fair agreement. CONCLUSION. Location in the axial skeleton and size greater than 5 cm are the most reliable predictors of central grade 1 chondrosarcoma. Morphologic features seen on radiographs and clinical symptoms do not improve the ability to differentiate between enchondromas and central grade 1 chondrosarcomas.