ESTRO 38 Abstract book

S1076 ESTRO 38

Material and Methods We propose a pixel-based exponential model to relate the measured transit images ( P ij ), the pre-patient fluence ( F ij ) and the mass attenuation thickness map ( x ij ) for each field: P ij = P 0 F ij exp(– µx ij ), where P 0 and µ are free scalar parameters. A first-order approximation of the directional derivatives of this model gives us the BEV shift estimation maps ( s ij ) as: s ij ≈ ( P ij / P ij 0 – 1)·[ µ ( ∂x / ∂s ) ij ] –1 , where P ij 0 is a convenient reference image (after IGRT corrections). Finally, the gantry angle is used to transform s ij from the BEV reference frame to patient/couch coordinates s ij ’. We validated the model for the following sites and techniques using a RANDO anthropomorphic phantom (Eclipse v13.5, Varian):  A head-and-neck IMRT plan with 7 sliding- window 6 MV fields (1.6/1.8/2.12 Gy/fx @ low/intermediate/high risk areas). We delivered these plans on a Clinac 2100C/D equipped with a Millenium 120 MLC (Varian). We shifted the phantom between fractions by moving the couch top from s = -10 to 10 mm in 1 mm steps along the three axes. We acquired transit images ( P ij ) in integrated mode by placing the EPID at 140(150)cm for the head-and-neck(breast) plans. We obtained F ij , x ij and ( ∂x / ∂s ) ij using the Eclipse Scripting API v13.5. We developed an in-house code (MATLAB v2015b, MathWorks) to fit P 0 , µ and to obtain s ij . Finally, we compared the per-beam pixel-averaged shift estimations < s ij ’> against the real shifts s for all plans and irradiations. Results Figure 1 compares the 723 model shift estimations against the real shifts. Estimations showed good linearity up to 5 mm (slope = 0.854±0.016, RMS of residuals = 1.03 mm). Linearity decreased for larger shifts due to the first-order nature of the model (slope = 0.820±0.011, RMS of residuals = 1.65 mm) but, even in this setting, estimations led to minimize positioning errors after correction. Figure 2 shows that only 12% of the estimations differed from the real shifts by more than 3mm (only 1% corresponded to shifts ≤5mm). Therefore, cumulative corrections after each field irradiation may rapidly converge to target position.  A whole-breast IMRT plan with 7 quasi- tangential sliding-window 6 MV fields (2 Gy/fx). A whole-breast 3DCRT plan with 2 tangential + 2 field-in-field 6 MV fields (2 Gy/fx). 

Results All three corsets reduced the pancreatic motion by a similar amount, mainly in inferior-superior direction (Table 1). As expected, the CT scan showed that the two PE corsets were made of homogeneous material that had a constant thickness of 2.5±0.1mm and 4.9±0.1mm, respectively. The PU corset was inhomogeneous due to the presence of small air inclusions throughout the whole corset. Furthermore, its thickness varied between 8.0 – 24.2mm. The WER of the two PE corsets was 0.990 and 0.956, while the average WER of the PU corset was 0.298.

Conclusion While all three corsets significantly reduce the respiratory-induced pancreatic motion, the PU corset is not suitable for PT due to its irregular material structure and thickness. The two PE corsets both show stable material conditions which could easily be included in treatment planning. Preferences for any of the two PE corsets will mainly depend on differences in fit, flexibility, cost and the time required for preparation. EP-1972 Transit-Guided RadioTherapy (TGRT): a novel intra-fraction patient monitoring approach A. Latorre- Musoll 1 , P. Delgado-Tapia 1 , M. Lizondo 1 , N. Jonet 1 , P. Carrasco 1 , A. Ruiz-Martínez 1 , I. Valverde- Pascual 1 , M. Barceló 1 , M. Ribas 1 1 Hospital de la Santa Creu i Sant Pau, Servei de Radiofísica i Radioprotecció, Barcelona, Spain Purpose or Objective IGRT is an effective tool to reduce patient positioning uncertainties. However, intra-fraction motion component is not completely removed by conventional daily imaging protocols. We propose a novel approach to correct intra- fraction patient positioning shifts based on transit portal images (TGRT) which does not add extra time nor imaging dose.