ESTRO 36 Abstract Book

S592 ESTRO 36 _______________________________________________________________________________________________

Radiation Oncology - Oncology and Hemato-oncology, Milan, Italy

Purpose or Objective To explore dosimetric predictors of acute dysphagia in head and neck (H&N) cancer patients (pts) treated with definitive radiotherapy (RT). We prospectively examined correlation between doses to swallowing-associated structures and acute radiation-related side effects, in terms of dysphagia and percutaneous endoscopic gastrostomy tube (PEG) requirement. Material and Methods We analyzed all consecutive not previously treated pts with H&N cancer who underwent RT at our Department between May 2010 and March 2011. Exclusion criteria were: baseline dysphagia (functional dysphagia or enteral nutrition) and previous surgery in the H&N region. A nutritional standardized step-wedge protocol was applied. Dysphagia (grade ≥3 according to CTCAE v4.0) and indication to PEG insertion were classified as acute toxicity events. Ten swallowing-related structures were considered for the analysis: pharyngeal axis, base of tongue, constrictor muscles (superior, middle and inferior), cricopharyngeal muscle, soft palate, cervical esophagus, oral cavity and supraglottic larynx. Dosimetric parameters included mean dose (Dmean), near maximum dose (D2%) and the percentage volume exceeding X Gy (Vx) evaluated in 5-Gy steps. The correlation of clinical information along with swallowing-related structure dose parameters related to acute toxicity events was analyzed by means of Spearman’s rank correlation coefficient (Rs). Multivariate logistic regression method using resampling methods (bootstrapping) was applied to select model order and parameters for normal tissue complication probability (NTCP) modeling. Model performance was evaluated through the area under the curve (AUC) of the receiver operating characteristic (ROC) analysis. Results Patient and treatment characteristics are summarized in Table 1. Two pts required PEG, 3 pts had grade 3 dysphagia and 4 pts had both PEG and grade 3 dysphagia. A strong multiple correlation between dosimetric parameters was found. Intra-organ dosimetric parameters were strongly correlated as well as inter-organ dosimetric parameters. Accordingly, the highly correlated variables (Rs>0.75) were not included in the multivariate analysis. A two-variable model was suggested as the optimal order by bootstrap method. The optimal model (Rs=0.452, p<0.001) includes V45 of the cervical esophagus (OR=1.016) and Dmean of the cricopharyngeal muscle (OR=1.057). The model AUC (Fig1a) was 0.82 (95% CI 0.69- 0.95). The comparison of the predicted incidence of acute radiation-related toxcity and the actuarial incidence in the population is shown in Figure 1b.

Conclusion We propose a 2-variable predictive NTCP model including both cervical esophagus and cricopharyngeal muscle dosimetric parameters with a good prediction performance for acute radiation-related toxicity in H&N cancer pts EP-1078 Transient xerostomia in head and neck cancers with significant parotid inclusion in target volume A. Datta 1 , A. Mukherji 1 , E. Thiraviyam 1 1 Jawaharlal institute of post graduate medical education and research, radiation oncology, Puducherry, India