Background and purpose: To evaluate the impact of quality assurance on treatment compliance, we compared the outcome of the two individual case reviews (ICR) conducted early and late during the accrual period of a large prospective multi-centre trial. Patients and methods: At the onset of the trial, medical files of five patients from each participating centre were evaluated for the compliance to the protocol for eligibility, surgery, pathology and radiotherapy and for the quality of reporting of the data on the case report forms. In nine major centres, this procedure was repeated near the end of the trial. Results: Both in the early and the late ICR, we found a very limited number of deviations from the guidelines for eligibility, staging, surgery, and pathology. Compliance to radiotherapy requirements was good with the exception of a too low minimal dose in 30% and the lack of target volume delineation in the majority of the evaluated patients. The comparison of the late with the early ICR demonstrated an improvement of the quality of data reporting by 6% and of target volume delineation from 33 to 53%. Conclusions: The initial ICR has lead to the identification of a number of parameters, which needed a clarification in the protocol. These items have been corrected and the individual institutions have been made aware of the necessary adaptations. The evaluation at the end of the trial period showed that there was an improvement but also showed that continuous monitoring is necessary, especially for institutions which have the most deviations in the first ICR. © 2005 Elsevier Ireland Ltd. All rights reserved.