Purpose: To determine if portal setup films are an accurate representation of a patient's position throughout the course of fractionated tangential breast irradiation.
Methods and materials: Thirteen patients undergoing external beam irradiation for T1-T2 infiltrating ductal carcinoma of the breast following excisional biopsy and axillary dissection were imaged using an on-line portal imaging device attached to a 6 MV linear accelerator. Medial and lateral tangential fields were imaged and a total of 139 fractions, 225 portal fields, and 4450 images were obtained. Interfractional and intrafractional variations for anatomical parameters including the central lung distance (CLD), central flash distance (CFD), and inferior central margin (ICM) were calculated from these images. A pooled estimate of the random error associated with a given treatment was determined by adding the interfractional and intrafractional standard deviations in quadrature. A 95% confidence level assigned a value of two standard deviations of the random error estimate. Central lung distance, CFD, and ICM distances were then measured for all portal setup films. Significant differences were defined as occurring when the simulation-setup difference was greater than the 95% confidence value.
Results: Differences between setup portal and simulation films were less than their 95% confidence values in 70 instances indicating that in 90% of the time these differences are a result of random differences in daily treatment positioning.
Conclusions: In 90% of cases tested, initial portal setup films are an accurate representation of a patients daily treatment setup.