Forty-four patients with advanced colon or rectal cancer were treated with methyl-GAG on a weekly schedule. Of the 40 evaluable patients, 35 (87%) had received prior chemotherapy. Objective tumor regression was seen in six patients (one CR, five PR's). An additional nine patients had stable disease for a median of 42 weeks. The median survival (42+ weeks) for responding and stable disease patients was significantly better (p = 0.0001 Wilcoxan test) than those with progressive disease (11 weeks). Toxicity was reversible and included mild to moderate mucositis, nausea, vomiting, diarrhea, and thrombocytopenia. Responses observed in this study warrant further trials in patients with colon cancer who have no prior chemotherapy.