ESTRO 2025 - Abstract Book

S2615

Physics - Detectors, dose measurement and phantoms

ESTRO 2025

2762

Digital Poster Can you trust AAA and AXB to calculate lens dose in VMAT treatments? Angharad E Ganguli 1 , George Thickett 2 , Alison Starke 3 , Jackie Poxon 2 , Niall MacDougall 2

1 Radiotherapy Physics, University College London Hospitals NHS Foundation Trust, London, United Kingdom. 2 Radiotherapy Physics, Barts Health NHS Trust, London, United Kingdom. 3 Radiotherapy Physics, NHS Greater Glasgow, Glasgow, United Kingdom Purpose/Objective: A custom-made anthropomorphic head phantom with a detachable eye was used to measure the dose to the retina and lens in complex VMAT treatment plans. These measurements were used to assess the accuracy of AXB and AAA when reporting doses to these critical OARs in two scenarios: when dose is delivered to directly the optics, and when all dose is from scatter. Material/Methods: The anthropomorphic head phantom contains a removable right eye with an internal cylinder designed to hold thermoluminescent detectors (TLD) chips at varying depths below the surface of the eye. The TLDs representing the lens and retina of the eye were positioned at a depth of 5 mm and 21mm respectively. Two historical patient plans were calculated on the phantom. These included a 60Gy/30# single dose level sinus plan and a 24Gy/1# three lesion metastatic brain SRS plan. The sinus plan delivered dose directly through the optics region. In the SRS plan, the eye received only scattered dose. Each plan was calculated in Varian TPS Eclipse version 17.1 using the AAA and AXB beam models. Monitor units, jaw settings and MLC settings were identical when calculating with AXB and AAA. Calculations using a grid size of 2.5 mm and 1 mm were performed for both algorithms. This was done for 6FFF and 10FFF. The measurement for each patient plan was delivered three times on a Varian TrueBeam linear accelerator. Average doses were computed for each plan at the retina and lens locations. TLD values were corrected for machine output. TLD doses were compared with calculated doses.

Results:

Measured and calculated TLD doses are provided in Table I. Good agreement was seen between both algorithms and measured dose for the sinus plan, when dose was delivered through the optics. Poorer agreement was noted for the SRS plan, with calculated doses higher than measured doses. In this plan, the dose to the retina and lens was scatter. Reducing the grid resolution from 2.5 mm to 1.0 mm dramatically improved the agreement with measured data for AXB, while negligible improvement was observed for AAA. This effect is very pronounced for the SRS brain plan.