ESTRO 2021 Abstract Book

S1365

ESTRO 2021

implants J. Tait 1 , T. Williams 1 , C.F. Stacey 2 , M. Hussein 3 , C.H. Clark 1 , U. Johnson 1

1 University College London Hospitals NHS Foundation Trust, Radiotherapy Physics, London, United Kingdom; 2 North Middlesex University Hospital, Radiotherapy Physics, London, United Kingdom; 3 National Physics Laboratory, Metrology for Medical Physics Centre, London, United Kingdom Purpose or Objective To investigate the accuracy of the Varian Eclipse Acuros XB (AXB) photon algorithm in the presence of high- density materials within the treated volume in comparison with the Varian Analytic Anisotropic Algorithm (AAA). Materials and Methods A RANDO phantom was adapted to include titanium inserts replicating surgical implants as seen in spinal reconstruction surgery. Two 6MV photon plans of varying complexity (single static field and Rapidarc VMAT) were planned with 2Gy prescription and delivered to the phantom. Dose measurements were made using EBT3 Gafchromic film in a plane of the phantom bisecting an array of titanium rods. Profile evaluation and gamma analysis for the film measurements was undertaken using a commercial (Omnipro I’mRT) and an independent (Vigo, NPL) software and compared with AAA and AXB calculated distributions. Analysis was made for both AXB Dm (dose to medium) and Dw (dose to water) calculations. Results for regions both including and excluding the titanium in the gamma analysis were compared using Vigo. Results For a single field exposure, film measurements confirmed the presence of hotspots anterior/lateral to the titanium rods that were predicted by AXB but not by AAA. This resulted in a 24% under-prediction of the dose deposited in this region by AAA. AXB Dm successfully predicted the hotspot to within 3.3% as measured with film. For a Rapidarc VMAT exposure, a plan sum subtracting the AAA distribution from the AXB Dm distribution can be seen in Figure 1 to show the hotspots surrounding the titanium predicted by AXB Dm but not AAA. These are approximately 4.5-5% of the 2Gy prescription dose. Film measurements validated the presence of these hotspots.

Gamma analysis was performed in a region encompassing the titanium pins and surrounding ‘tissue’ (table 1). AAA had 80.31% gamma pass rate for the Rapidarc plan comparing to Gafchromic film measurement with a 2%/2mm passing criteria. AXB had 89.05% gamma pass rate for the same plan. Using Vigo and excluding the regions of titanium, the gamma pass rate increased by 1.8% for the AXB Dm Rapidarc plan. Removing the titanium from the gamma analysis for the AAA plan made no significant difference to the pass rate.

Conclusion The AXB Dm was able to calculate a dose distribution that more accurately represented the dose distribution for Rapidarc treatment plans in the presence of titanium spine implants in an anthropomorphic phantom, as measured with Gafchromic film. The primary indications are that the algorithm improves calculation accuracy in these conditions giving more reliable dosimetric information to the clinician and planner. This should then allow visualization of hotspots, indicating that the treatment plans should be adjusted to protect OARs. Therefore, plans could be more confidently delivered to effectively treat the disease whilst ensuring OARs and normal tissues do not receive consequential, local hotspots.