We conducted a feasibility study of sublingual methadone for breakthrough cancer pain, to determine whether a larger, randomized trial was warranted, and to identify a study design that would be likely to succeed. From approximately 1930 patients in the initial pool, nine patients were enrolled. Five patients completed the study, generating data on 83 discrete episodes of breakthrough cancer pain at optimal dose. Mean pain intensity dropped by 1.7 points (on a 10 point numerical scale) within 10 min of sublingual methadone administration, and by 3.2 points after 15 min. No serious or severe toxicity was encountered. Based on the results of this feasibility study, a larger randomized clinical trial of sublingual methadone for breakthrough cancer pain using this trial model would not be successful. Extensive information obtained from small numbers of carefully studied patients provides proof of concept that sublingual methadone is effective, safe, and well tolerated.