Evaluating surgical margins during composite resection of carcinoma is limited by analysis of the bone margins. The standard pathologic evaluation of bone is by decalcification. A method of analysis was devised based on histologically proven methods of cortical invasion and subsequent spread. Frozen section analysis (FSA) of the cancellous bone was investigated as a rapid method of evaluating adequacy of the mandibular resection. Subjects consisted of 29 patients undergoing full-thickness mandibular resection, with 16 cases of histologically proven mandibular invasion qualifying for evaluation. Results of FSA were then compared to the permanent section analysis of the cancellous bone and to the decalcified specimen. Complete correlation was found between frozen and permanent section results. Frozen section analysis was able to correctly predict adequacy of resection in 32 (97%) of 33 margins.