A method for validating a transperineal ultrasound system for intrafraction monitoring of the prostate during external beam radiotherapy

Abstract

Introduction: The Clarity Autoscan 40 transperineal ultrasound (TPUS) system (Eiekta, Sweden) for prostate motion management employs a vertically-oriented 20 ultrasound array that is continuously swept mechanically to repeatedly produce 30 images containing the prostate [1]. The target position relative to a pre-fraction reference scan is determined multiple times per second. Other investigators have studied the tracking accuracy of the system using displacements of ~1 0 mm from the initial normalisation point typical to a clinical treatment [1-4]. The primary aim of this work was to utilise clinically available equipment to compare the target positions reported by the Clarity Autoscan system to known target positions over the full imaging volume. A scanning dosimetry water tank was used, however refraction in the 20 mm PMMA wall of the tank presented a significant complication. The potential variation in target dose due to intervention based on the Clarity prostate motion management was also investigated. Method: A prostate analogue was mounted to the scanning mechanism of a MP3-XS scanning water tank "' (PTW, Germany). The Clarity probe was positioned externally against the wall of the scanning tank in the treatment orientation. The scanning mechanism was programmed to make in-plane, cross-plane and diagonal 'profiles' in the horizontal plane ranging approximately ±30 mm from the isocentre. Seven sets of these four 'profiles' were acquired between ±30 mm in the vertical direction yielding data throughout a 60 cm-sided cube centred on the isocentre. A bi-layer 30 refraction correction algorithm was derived to account for refraction caused by differences between the speed of sound in both PMMA and water from the speed of sound in soft tissue assumed by the Clarity system. The prostate analogue was then replaced with a Farmer-type ionisation chamber and monitored by the Clarity system during beam delivery. Programmed movements of the chamber triggered manual or automatic suspension of the beam and the resulting measured doses compared. Results: Without refraction correction the maximum difference in the reported positions from the programmed positions was 9.3 mm and the mean(±SD) difference was 4.0±1.8 mm. Refraction correction reduced this to a maximum of 3.4 mm, and a mean(±SD) of 1.0±0.5 mm. The worst results were at the peripheries of the imaged volume and near the transducer where the Clarity system had difficulty maintaining tracking due to narrowing of the swept imaging volume. At the lateral (left-right) and vertical (anterior-posterior) extremities, the prostate analogue images were visibly distorted which may have affected the accuracy of the Clarity centroid position calculation. There was no significant difference in measured dose between manual and automatic beam suspension in a 10x10 cm2 field when the target moved along the beam axis. Furthermore, there was only a minimal difference in measured dose to the centre of the 'prostate' between intervention and no intervention when the 'prostate' was programmed to move ±20 mm along the beam axis during a 180 MU 1 Ox1 0 cm2 field beam. However, it was found that there was a delay of 5.4±0.9 s between threshold crossing and beam suspension which could become significant at higher dose rates. Conclusions: The target positions reported by the Elekta Clarity Autoscan system can be validated using a programmable scanning water tank by employing a refraction correction if care is taken in the initial positioning of the transducer. Further improvement might be achieved by using a smaller target analogue and associated volume to reduce the effect of the refraction-induced distortion on the Clarity centroid calculation. Intervention following detected prostate motion along the beam axis will have minimal effect on the dose to the centre of the prostate; however, motion in any direction will compromise target coverage and dose minimisation to healthy tissue.