A multicenter study to quantify systematic variations and associated uncertainties in source positioning with...

The reconstruction of radiation source position in the treatment planning system is a key part of the applicator reconstruction process in high dose rate (HDR) brachytherapy treatment of cervical carcinomas. The steep dose gradients, of as much as 12%/mm, associated with typical cervix treatments emphasize the importance of accurate and precise determination of source positions. However, a variety of methodologies with a range in associated measurement uncertainties, of up to ±2.5 mm, are currently employed by various centers to do this. In addition, a recent pilot study by Awunor et al. ["Direct reconstruction and associated uncertainties of (192)Ir source dwell positions in ring applicators using gafchromic film in the treatment planning of HDR brachytherapy cervix patients," Phys. Med. Biol. 58, 3207-3225 (2013)] reported source positional differences of up to 2.6 mm between ring sets of the same type and geometry. This suggests a need for a comprehensive study to assess and quantify systematic source position variations between commonly used ring applicators and HDR afterloaders across multiple centers.

Eighty-six rings from 20 European brachytherapy centers were audited in the form of a postal audit with each center collecting the data independently. The data were collected by setting up the rings using a bespoke jig and irradiating gafchromic films at predetermined dwell positions using four afterloader types, MicroSelectron, Flexitron, GammaMed, and MultiSource, from three manufacturers, Nucletron, Varian, and Eckert & Ziegler BEBIG. Five different ring types in six sizes (Ø25-Ø35 mm) and two angles (45° and 60°) were used. Coordinates of irradiated positions relative to the ring center were determined and collated, and source position differences quantified by ring type, size, and angle.

The mean expanded measurement uncertainty (k = 2) along the direction of source travel was ±1.4 mm. The standard deviation associated with the source position reproducibility was within ±1.0 mm for all afterloaders. Maximum source positional variations of 2.1 and 3.9, 1.8 and 5.4, and 2.3 and 3.4 mm were observed at standard treatment positions for the Ø26, Ø30, and Ø32 mm sized 45° and 60° rings, respectively. Mean positional differences between a majority of the rings were within ±1.0 mm. Mean positional differences between a majority of the intracenter ring sets were within the expanded measurement uncertainty. When comparing the 45°-60° source paths, mean differences of 1.6, 0.9, and 0.9 mm were observed across the Ø26, Ø30 (MicroSelectron), and Ø32 mm (GammaMed) rings, respectively. When comparing to manufacturer source path models, maximum offsets of 1.9 and 2.1, 2.6 and 2.3, and 0.8 and 1.6 mm were observed for the Ø26, Ø30 (MicroSelectron), and Ø30 mm (Flexitron) sized 45° and 60° rings, respectively. When comparing the audit to ring commissioning data of participating centers, mean differences of up to 2.4 mm were observed.

A majority of the audited rings showed a good degree of manufacturer consistency; however, substantial positional variation observed between some rings emphasizes the importance of commissioning each ring before clinical use. Differences observed between audit and commissioning data also indicate some variation in source treatment positions across centers.