Between August 1989 and July 1993, a controlled randomized trial was performed in the Netherlands to compare the therapeutic efficacy of Dl resection with limited (N1 level) lymphadenectomy versus D2 resection with extended lymph node dissection (N1 and N2 level). In this trial, a total of 1,078 patients were randomized, of whom 996 (92%) were eligible for analysis. Of these, 711 (71 %) were operated on with curative intent and 285 (29%) underwent a non-curative procedure. The actual distinction between limited and extended lymphadenectomy was analyzed. Nonadherence to the protocol was evaluated on the basis of noncompliance (no yield of lymph nodes from indicated stations) and contamination (finding of one or more lymph nodes for which dissection was not indicated). In the first period, data of 237 patients (115 D1 and 122 D2) operated on with curative intent were available. Noncompliance occurred in 84% of both Dl and D2 resections, while contamination was seen in 48% of D1 and 52% of D2 resections. Contamination and noncompliance indicate a partial loss of distinction between the two types of resection, showing the tendency of performing an intermediate type of resection, D1.5. Additional measures were taken, such as stricter perioperative quality control, centralized lymph node retrieval and evaluation of protocol deviations per participating team to improve standardization of the surgicopathological team performance. After implementing these measures, in the second period, noncompliance did not decrease significantly; 236 (138 Dl and 98 D2) patients (p = 0.08), while in D2 the decrease was significant (p = 0.03). The contamination rate decreased significantly in both Dl (p < 0.05) and D2 (p = 0.02) resections. Furthermore, our results proved that extended lymphadenectomy induced a substantial stage migration. In conclusion, the implementation of measures improved standardization of the surgicopathological team performance. Stage migration is also demonstrated in our randomized controlled trial. Not only major changes in clinical practice, such as routine application of extended lymphadenectomy, but also more subtle changes in the performance of the surgicopathological team, such as contamination, noncompliance and diligence, strongly affect stage-specific survival rates.