Morphological classification for prediction of malignant transformation in multiple exostoses.

AIM: To explore the value of morphological classification in predicting malignant transformation in multiple exostoses (ME). PATIENTS AND METHODS: The imaging data of 116 patients (totally 190 tumors) with ME were retrospectively analyzed. All the tumors were pathology confirmed after surgical resection, including 175 exostoses from 101 patients, and 15 exostotic chondrosarcomas in 15 cases. Based on the ratio of diameter between tumor tip and tumor base (R1), tumors were classified into two types: cauliflower-like tumor (R1 ≥ 1.0) and non-cauliflower-like tumor (R1 < 1.0). In addition, non-cauliflower-like tumors were further classified into two subtypes according to the ratio of tumor height to tumor base diameter: sessile type (R2 < 1.0) and pedunculated type (R2 ≥ 1.0). The relationship between tumor shape and malignant transformation was studied.
RESULTS: Of all the 175 exostoses from 101 patients, 27 were cauliflower-like tumors and 148 were non-cauliflower-like tumors. Of all the 15 exostotic chondrosarcomas in 15 cases, most tumors were cauliflower-like (c2 = 38.0075, p < 0.05). Cauliflower-like tumor for the prediction of exostotic chondrosarcoma, the sensibility, specificity, positive predictive value, negative predictive value were 86.7%, 84.6%, 32.5%, 98.7%, respectively.
CONCLUSIONS: Tumor malignant transformation was more common in cauliflower-like tumors than in non-cauliflower-like tumors. The morphological classification and preventive resection of cauliflower-like tumors maybe helpful in preventing the malignant transformation of ME.