ESTRO 37 Abstract book

S656

ESTRO 37

Intensity modulated RT (IMRT) is a modern RT technique that has initially been developed to reduce this toxicity. Actually, it is considered standard of care based on level I evidence of reduction in xerostomia. A simultaneous integrated boost (SIB) IMRT allows the irradiation of different target volumes to different dose levels within a single treatment session and to increase the radiation dose without an increase in toxicity. This is done in an attempt to achieve higher local and regional control rates, as locoregional recurrences are very frequent and known to often appear within the high dose region. We report the Jules Bordet Institute experience relative to the outcome and toxicity of patients treated by definitive or postoperative SIB IMRT for early and locoregionally advanced squamous cell carcinoma in head and neck cancer (SCCHN). Material and Methods One hundred and six patients with SCCHN of the oral cavity (OC), oropharynx (OP), larynx (L) and hypopharynx (HP), consecutively treated at our cancer center between 3/2012 and 3/2014 were retrospectively analyzed. The prescribed doses SIB IMRT were in the postoperative setting (group A): 60–66 Gy and 53 Gy in 30-33 fractions for PTV high risk and PTV elective, respectively; and when given as primary treatment (group B): 66–70 Gy and 56 Gy in 35 fractions for PTV high risk and PTV elective, respectively. Toxicity was consistently graded by the same radiation oncologist. Results Seventy two percent of patients received definitive IMRT and 28% were postoperatively irradiated. In 56% At 5 years, locoregional control, distant control and overall survival were 77%, 79%, and 53% in the postoperative setting and 60%, 76% and 41% in the definitive SIB IMRT group respectively. The observed acute grade 3 toxicities were dysphagia (40%), mucositis (40%) and dermatitis (18%). Late toxicity were predominantly xerostomia (47%), dysgeusia (28%), dysphagia (10%) and dysphonia (7%). Conclusion SIB IMRT is feasible technique, safe and effective in the treatment of head and neck cancer patients. concomitant chemotherapy was given. Median follow-up was 31, 2 months.

Figure 2. Outcome in patients treated with definitive SIB IMRT EP-1172 Characterization of recurrence origin using pre-treatment PET/CT for head and neck cancers J. BEAUMONT 1 , O. Acosta 1 , A. Devillers 2 , X. Palard- Novello 2 , E. Chajon 2 , R. De Crevoisier 2 , J. Castelli 2 1 INSERM U1099, LTSI, Rennes, France 2 INSERM U1099, Centre Eugenes Marquis, Rennes, France Purpose or Objective In a context of locally advanced head and neck cancer (HNC) treated with definitive chemoradiotherapy (CRT), the aim of this study was to characterize the origin of recurrence from pre-treatment 18 F-fluorodeoxyglucose (FDG) positron emission tomography (PET) images . Material and Methods Twenty-six patients with a local recurrence after definitive CRT for locally advanced HNC were included in this study. PET and Computed Tomography (CT) images acquired to plan the treatment and at the recurrence were available for each patient. The primary tumor was segmented on both PET pre- treatment and at recurrence . Pairwise co-registration between pre-treatment and recurrence PET/CT images was performed using an automatic ROI-based rigid registration algorithm. Recurrence origin (GTVfailure) was defined as the anatomical region where the pre- treatment and the recurrent primary tumors intersected after co-registration . Conversely, the responder volume (GTVr) was defined as the anatomical region where the pre-treatment tumor was not overlapped by the recurrent tumor. Registration accuracy was assessed by a senior radiation oncologist. Based on pre-treatment PET, Metabolic Tumor Volume (MTV) was computed with thresholds ranging from 0% to 100% of SUVmax . The rate of GTVfailure and GTVr that compose the MTV was computed for each threshold. Histogram and texture features (Haralick) were extracted from pre-treatment PET images. Wilcoxon signed-rank test was used to assess whether features computed within GTVfailure were significantly different from features computed within GTVr. Results were considered as significantly different if their p-values were lower than 0.05 (95% interval). Results Sixteen pairwise registrations out of twenty-six proved to be anatomically accurate enough to identify recurrence origin . Thus, ten patients were removed from the study for further analysis. None of the different thresholds of MTV allowed to distinguish GTVr from GTVfailure (Figure 1). Table 1 presents the p-values of histogram and texture features after a Wilcoxon signed-rank test. The mean SUV (SUVmean) of GTVfailure was significantly higher than SUVmean of GTVr (p= 0.007).

Figure 1. Outcome in patients treated with postoperative SIB IMRT