ESTRO 36 Abstract Book

S597 ESTRO 36 2017 _______________________________________________________________________________________________

A database of HNSCC patients recruited to the Cancer Research UK VoxTox study between March 2013 and January 2016 was reviewed. All patients had been treated on TomoTherapy units, according to local protocol. Inclusion criteria for this sub-study were: full details of treatment protocol (RT and chemo) available, adequate follow-up, confirmed residual or loco-regionally relapsed disease, and diagnostic quality imaging of this disease (including MRI, PET/CT, CT) available for review. This gave a final cohort of 12. Imaging confirming relapse or residual disease was uploaded to virtual simulation software (Prosoma 3.3), where an in-built DIR package was used to fuse these images with the planning CT and treatment dose-cube. Residual/relapsed disease was contoured on the relapse imaging, named rGTV, and the volume of this structure recorded. Contours describing the 95% isodose of 65Gy, 60Gy, 55Gy and 54Gy where relevant were generated. For each case, the proportion of rGTV covered by these isodoses and the treatment PTV and CTVs were recorded (labelled ‘V’). Relapses were classified according to the 95% isodose of prescription dose contour: ‘in-field’ V > 75%, ‘marginal’ V – 25-75%, ‘out of field’ V < 25%. Results

Figure 1 describes the volume of rGTV covered by pertinent isodose contours and target CTV/PTVs. According to our criteria, 9 of 12 relapses were in-field (75%), 1 (8.3%) was marginal and 2 (16.7%) were in deliberately un-irradiated contralateral neck nodes as described. Conclusion Oral cavity tumours appeared at highest risk of relapse in our cohort. Using DIR to map disease relapse to treatment volumes and dose, we found most relapsed HNSCC (75%) occurred within the high dose volume, in keeping with previous studies, and suggesting that unfavourable biology rather than inadequate RT is the predominant reason for loco-regional HNSCC relapse. Our results will be used to inform review of our neck irradiation policy, particularly for SCCs of the oral cavity. EP-1097 P16 expression: a predictive marker for treatment-related outcomes in oropharyngeal cancer patients? A. Modesto 1 , T. Galissier 2 , A. Lusque 3 , E. Uro-Coste 2 , J. Delord 4 , A. Laprie 1 , J. Sarini 5 , P. Graff 1 , P. Vergez 5 , M. Rives 1 1 Institut Universitaire du Cancer, radiation therapy, Toulouse, France 2 Institut Universitaire du Cancer, Pathology, Toulouse, France 3 Institut Universitaire du Cancer, Biostatistics, Toulouse, France 4 Institut Universitaire du Cancer, Medical Oncology, Toulouse, France 5 Institut Universitaire du Cancer, Head and Neck Surgery, Toulouse, France Purpose or Objective Treatment strategies in oropharyngeal squamous cell carcinomas (OSCC) consist in either surgery followed by adjuvant radio(chemo)therapy or definitive radio(chemo)therapy. P16 overexpression (p16+) is considered as a surrogate marker for HPV-induced tumors that are associated with improved outcome whichever treatment modality is considered in comparison with p16 negative (p16-) OSCC. To date no predictive factors are known to guide treatment decision. Material and Methods All consecutive patients treated for an OSCC with a curative intend at a single tertiary cancer center between 2009 and 2013 were eligible to this study. P16 status was determined by immunochemistry and centrally reviewed. Late toxicities incidence ie: dysphagia, xerostomia, painful shoulder, osteoradionecrosis or nerve paralysis were registrated and graded according to CTCAE v4 in patients alive without loco-regional evolution at least 6 months after treatment completion. Three-year disease free survival (DFS) and late severe toxicity occurrence were compared according to p16 expression and treatment modality: initial surgical treatment or definitive radio(chemo)therapy. Results Among the 167 patients included in this study, 77 (44%) presented a tumoral p16 overexpression (p16+). Initial

Case details are given in Table 1. Oral cavity primary disease accounted for 15% of cases in the original cohort available for review (prior to exclusions), but half of patients who relapsed loco-regionally. Mean time to relapse was 18 weeks (Std. error +/- 2.4 weeks). 10 of 12 patients (83%) had residual or relapsed disease ipsilateral to the primary site. 2 (both lateral tongue SCCs who underwent primary surgery followed by RT +/- chemo to the primary site and ipsilateral neck) relapsed early (8 and 20 weeks respectively) in the contralateral, un-irradiated neck.