ESTRO 2020 Abstract book

S330 ESTRO 2020

use of immobilization devices, incorrect computation, anatomical variations and unknown causes. Results The three systems use different approaches and customized alert indices, based on institutional preferences. For VMAT treatments T% mean values were 11.8 and 18.0% for DC, 16.0% for PF systems, and up to 8.9% for SD system. Errors due to “anatomical variations” for Head and Neck treatments were up to 9.0% for SD, 9.0% for DC and 8.0% for PF systems, while for Abdomen Pelvis/Prostate treatments were up to 9.0% for SD, up to 17.1 for DC and up to 9.0% for PF. Comparison among various techniques can be obtained analyzing mean SD results: 1.5% (range 0-3.0%) for SBRT; 7.0% (range 4.7 – 8.9%) for VMAT; 10.4% (range 7.0 - 12.2%) for IMRT; 13.2% (range 8.8 – 21.0%) for the 3D CRT.

Carlo studies have demonstrated that magnetic focusing of proton beams can produce minibeam dose distributions with enhanced delivery efficiency and less primary beam interaction with beam modifying devices. The work presented here experimentally tests the ability of a single quadrupole magnet to focus an incident proton beam and deliver an acceptable minibeam distribution. Material and Methods A proton beam of 9.8 cm range in water was focused by a single permanent quadrupole magnet in this study. The quadrupole magnet was constructed from segments of radiation resistant Sm2Co17 rare-earth permanent magnetic material adhered into a Halbach cylinder with a nominal field gradient of 250 T/m and a length of 6.8 cm. Minibeam distributions were evaluated as a function of depth using EBT3 Gafchromic film located within a custom- made water tank. Both single and multiple minibeam distributions were assessed, with the latter delivered through the use of a laboratory jack and precision height gauge shifting the water tank vertically with respect to the incident proton beam (Figure 1). Lastly, modulated minibeam distributions were delivered by shifting the water tank vertically and modulating individual beamlets using appropriate range shifts and weightings.

Figure 1: Experimental setup for proton minibeam delivery. From left to right components are: 1C=first collimator, FC=functional cone, 2C=second collimator, VS=vertical spacers, M=magnet, FH=film holder, LJ=laboratory jack, HG=height gauge, WT=water tank, DI=digital imager. Results The Halbach cylinder allowed for the delivery of focused proton beamlets with a narrow FWHM of 2.43 mm at entrance. Additionally, the experimental setup allowed for the delivery of these beamlets with precise lateral displacement (+/-0.02 mm) and individual weighting or modulation. At entrance, the peak to valley dose ratios were 9.8 and 9.1 for the multiple beamlet and modulated multiple beamlet cases respectively (Figure 2). Both the multiple beamlet and modulated multiple beamlet cases delivered a merged dose at the level of the Bragg peak (98 mm) with a FW90M of 15.47 and flatness of 4.25% and 2.78% respectively.

Table: Number of patient (P), test (T), and percentage of tests out of the tolerance level (T%), divided for source of error, Center, and EPID dosimetry software. Conclusion The implementation of IVD improved dosimetric accuracy and treatment reproducibility in all Centers participating to this study.

PD-0550 Esophageal tumor identification by deep learning on treatment planning CT images

Abstract withdrawn

PD-0551 Experimental investigation of magnetically focused proton minibeams A. Teran 1 , G. McAuley 1 , J. Slater 1 , A. Wroe 1 1 Loma Linda University, Department of Radiation Medicine, Loma Linda, USA Purpose or Objective Proton minibeam radiation therapy (pMBRT) potentially allows for dose escalation at the target through normal tissue sparing via spatial fractionation. Proton minibeams are typically generated via collimation, however Monte