ESTRO 37 Abstract book

ESTRO 37

S528

Oncology Department, Cambridge, United Kingdom 5 Cambridge University Hospital's NHS Foundation Trust, Department of Medical Physics and Clinical Engineering, Cambridge, United Kingdom 6 University of Cambridge, Department of Engineering, Cambridge, United Kingdom Purpose or Objective Delivered radiation dose (D A ) to critical OARs in patients undergoing curative radiotherapy for squamous cell carcinoma of the head & neck (HNSCC) can differ from planned dose (D P ), potentially increasing long-term morbidity risk. This study aims to quantify these differences, and examines their relationship to anatomical changes during treatment. Material and Methods 87 HNSCC patients were selected (72 oropharynx, 8 larynx, 5 oral cavity, 2 hypopharynx). All were treated with IMRT and daily MVCT-IG on TomoTherapy HiArt units, with a 3-dose level technique (65Gy/30#/6wks). All patients were weighed before fraction 1 and in the final treatment week. Patient separation (lateral neck dimension - LND - the transverse skin-to-skin distance at the level of the C1 vertebra) was measured on day 1 & 30 MVCTs, and differences noted (ΔLND). An automated function within Prosoma segmentation software defined the skin surface at the same slice on the same images (days 1 & 30); the slice surface area of this contour (SSA) was measured, and the change recorded (ΔSSA). Doses to salivary glands (parotid – PG, submandibular – SMG, ipsilateral & contralateral to primary tumour), and superior pharyngeal constrictor muscles (SPCs) were assessed. All structures were re-contoured on kVCT planning scans by one observer. An open-source deformable image registration algorithm (Elastix) was used to propagate kVCT contours to daily MVCTs. Daily IG scans, TomoTherapy sinogram data and an in-house ray- tracing dose calculation algorithm (CheckTomo) were used to compute daily dose cubes. Individual voxel histories were created and D A DVHs produced. D A data for individual structures was only reported where the full volume was included on every IG scan. To optimise comparison, D P was recomputed on the kVCT with CheckTomo. Mean dose was used for comparison. Differences were assessed with paired t-tests. Relationships between anatomical metrics and dose differences to each organ were assessed with univariate linear regression. Results Mean weight loss was 6.7kg (7.9%). Mean reductions in LND and SSA were 12.8mm (8.3%) and 11.7cm2 (5.1%). Delivered mean dose was higher than planned mean dose for all 5 structures (Table 1, Figure 1). The biggest observed dose difference was for the contralateral PG (7.6% increase); the greatest individual difference was 6.3Gy (D P 23.3, D A 29.6). No relationship between anatomical changes and dose differences was seen for the contralateral PG, either SMG, or the SPCs. A borderline significant relationship between greater reduction in LND and higher D A was observed for ipsilateral PG.

Conclusion Delivered dose was higher than planned for all structures, with the greatest proportional difference for the contralateral PG. Weight loss and anatomical change did not influence dose differences with the exception of the ipsilateral PG, where a relationship with reducing separation is suggested. These data may help to guide patient selection for Adaptive Radiotherapy. PO-0962 CBCT dose calculation in head and neck adaptive radiotherapy: accuracy assessment of four methods A. Barateau 1,2 , N. Perichon 3 , C. Hervé 3 , O. Henry 3 , J. Castelli 1,2,3 , E. Chajon 3 , R. De Crevoisier 1,2,3 , C. Lafond 1,2,3 1 INSERM U1099, LTSI, Rennes, France 2 LTSI, IMPACT, Rennes, France 3 Centre Eugène Marquis, Radiotherapy Department, Rennes, France Purpose or Objective Anatomical variations occur during head and neck (H&N) radiotherapy treatment. Cone Beam CT (CBCT) images could be used for dose calculation in order to quantify the dose variations due to these anatomical changes. The main limits of CBCT dose calculation are the image quality, the Hounsfield Units (HU) consistency and the reconstruction size. This study aims to assess accuracy of four CBCT dose calculation methods in H&N radiotherapy. Material and Methods 20 couples of weekly CT (Bigbore, Philips) and CBCT images (XVI, Elekta) from five H&N patients were used. Each CBCT was acquired the same day as the corresponding CT in a context of adaptive radiotherapy with weekly replanning. Four methods were investigated to calculate the dose on CBCT: i) use of a density to HU relation from phantom CBCT image (phantom HU-D relation), ii) use of a HU-D relation from patient CBCT image (patient HU-D relation), iii) use of a density assignment method of three structures (air, soft tissues and bones) and iv) use of an elastic deformation method (research Admire software, Elekta) deforming the patient CT on the shape of CBCT, creating a deformed-CT whose