Quality control by portal film analysis in radiotherapy for prostate cancer: A comparison between two different institutions and treatment techniques

Aims and background: Accuracy and reproducibility of patient setup during r adiotherapy for prostate cancer were investigated in two different Institut ions (A and B), within their Quality Assurance programs. The purpose of the study was to evaluate and compare setup accuracy and reproducibility in In stitutions A and B, which adopt different patient positioning and treatment techniques for prostate irradiation. Materials and methods: A retrospectiv e analysis of portal localization films taken during the treatment course w as performed: 30 and 21 patients in Institutes A and B, respectively, enter ed the study. In Institute A, patients were treated in a prone position, ut ilizing an individualized immobilization cast (either an alpha cradle or a heat and vacuum-formed cellulose acetate cast) with an open table top and i ndividual abdominal wall compressor to minimize small bowel irradiation; a B-field conformal technique was used. In Institute B, patients were treated in a supine position without any immobilization device; a 6-field BEV-base d technique (conformal only for patients treated with a radical aim) was ad opted, A total of 598 portal films (420 from Institute A and 178 from Insti tute B) were analyzed. The mean number of films per patient was 12 (range, 4-29). Systematic and random setup errors were estimated utilizing the stat istical method suggested by Bijhold et al. (1992). Results: When patients w ith a mean (systematic) error larger than 5, 8 and 10 mm in craniocaudal, l ateral and posterior-anterior directions, respectively, were compared, no s tatistically significant difference between the two groups was observed. Si milarly, when comparing portal films, a significant difference (P < 0.01) a ppeared only in the craniocaudal direction (errors > 5 mm: Institute A = 24 %; Institute B = 11%). In both Institutes, the SD of random and systematic error distribution ranged from 1.8 to 4.2 mm, with a small prevalence of sy stematic errors, Only for craniocaudal shifts in Institute A was the random error larger than the systematic error, and it was significantly worse tha n in Institute B (1 SD, 4.2 mm in Institute A vs 1.8 mm in Institute B), Co nclusions: Setup errors observed in Institutes A and B were similar and in accord with data reported in the literature. In Institute B, satisfactory g eometrical treatment quality was achieved without patient immobilization. I n Institute A, the goal of minimizing small bowel irradiation and prostate motion through the aforementioned technique, which makes patient position l ess comfortable, did not seem to considerably increase daily setup uncertai nty.