Quality control review for radiotherapy of small breast cancer: analysis of 708 patients in the GBSG I trial. German Breast Study Group (GBSG).

Within six years between November 1983 and December 1989 the German Breast Cancer Study Group (GBSG) conducted a prospective multicenter trial on the treatment of pT1 pN0 M0 breast carcinoma. Out of 1036 eligible patients from 69 hospitals, a total of 733 underwent breast preservation surgery and postoperative radiotherapy. A detailed quality control (QC) screening of 708 available radiotherapy records was performed by the radiotherapy reference center. The following quality control scoring system was used for all radiotherapy variables: Radiotherapy treatment "completely according to study protocol" (QC score = 0); "acceptable protocol deviations" (QC score = 1); "unacceptable protocol violations" (QC score = 2); the "overall QC score" was the worst judgement for any single quality control criterion, while the "total QC score" was the summation of all single quality control scores (range: 0 to 14). 292 (41.2%) patients were treated per protocol; 290 (41.0%) had acceptable protocol deviations; and 126 (17.8%) had unacceptable protocol violations. 107 (84.9%) single violations, 15 (11.9%) combinations of two, three (2.4%) combinations of three, and one (0.8%) combination of four unacceptable violations were observed. 564 (79.4%) patients achieved a total QC score of < or = 2, 124 (17.5%) of 3 to 5, and 20 (2.8%) of > or = 6. The criterion "radiotherapy treatment duration" had the highest rate of protocol violations, 67/708 (9.5%); other quality control criteria with protocol violations were "radiotherapy reference dose" (27 = 3.8%), "radiotherapy treatment initiation" (15 = 2.1%) and "other protocol violations" (33 = 4.7%). No protocol violations were observed for the criterion "target volume coverage". The protocol compliance improved slightly during the conduct of the study. Institutional differences in protocol compliance were observed depending upon the accrual rate, the institutional treatment preference and between academic and non-academic hospitals. After a short median follow-up of 48 months, differences in protocol compliance, i.e. protocol violations, have not resulted in a reduced disease-free survival; however, the final prognostic evaluation of the quality control review may require longer follow-up and eventually different treatment endpoints to be analyzed.