Quality control involving surgical treatment in multi-institutional cancer trials is important because the results of postoperative adjuvant therapy might be obscured by inadequate surgery or pathologic examination of the specimen. In 1975, the Southeastern Cancer Study Group (SEG) initiated a randomized clinical trial of adjuvant immunotherapy (Corynebacterium parvum vs. bacillus Calmette-Guerin) in melanoma patients with nodal metastases. During the course of reviewing the results several years later, 20 of 136 patients (15%) entered into this study were judged as surgically ineligible. The reasons were: (1) biopsy of a metastatic node only without any subsequent regional lymph node dissection (12 patients), (2) partial lymph node dissection (six patients), or (3) too few nodes surgically removed or pathologically identified in the specimen (six patients). All 20 patients were entered into the study by medical oncologists. Thirteen of these 20 surgically ineligible patients have relapsed so far; many were taken off the study as 'immunotherapy failures,' when, in fact, they were surgical failures. Compared to the 116 surgically eligible patients, the 20 ineligible patients had a shorter median survival (4 months vs. 25 months) and a lower 1-year disease-free survival rate (36% vs. 62%, p = 0.01). The two groups were balanced equally with respect to prognostic factors. Because of these findings, minimum surgical and pathologic guidelines were established for each adjuvant therapy protocol in the SEG. Surgical quality control was reviewed by a surgeon in each institution prior to randomization and again by a surgical investigator centrally. Pathologic criteria were also defined more precisely. The problems with surgically ineligible patients have since been virtually eliminated. Quality control measures for surgical patients entered into cooperative group trials is an essential part of the protocol design and data review. In order to evaluate properly the impact of adjuvant therapy, each clinical trial must comprise a uniform group of surgically treated patients.