H.P. van der Laan

Radiation-induced fibrosis in the boost area after three-dimensional conformal radiotherapy with a simultaneous integrated boost technique for early-stage breast cancer: A multivariable prediction model

BACKGROUND AND PURPOSE To develop a multivariable prediction model for the risk of grade⩾2 fibrosis in the boost area after breast conserving surgery (BCS) followed by three-dimensional conformal radiotherapy (RT) with a simultaneous integrated photon boost (3D-CRT-SIB), five years after RT. MATERIAL AND METHODS This prospective cohort study included 1,030 patients treated with RT for breast cancer (stage 0-III), after BCS. Data regarding physician-rated fibrosis and dose-volume parameters were available in 546 patients. A multivariable logistic regression model for grade⩾2 fibrosis was generated. RESULTS At 5years, grade⩾2 fibrosis was observed in 13.4% of the patients. The multivariable analysis resulted in a prediction model for grade⩾2 fibrosis in the boost area including three independent variables: patient age, breast volume receiving⩾55Gy (V55 CTV breast) and the maximum radiation dose in the breast (Dmax). CONCLUSIONS A multivariable prediction model was developed including age, V55 CTV breast and Dmax for grade⩾2 fibrosis in the boost area after breast cancer RT using a 3D-CRT-SIB technique. This model can be used to estimate the risk of fibrosis and to optimize dose distributions aiming at reducing this risk.

PD-0092: The dose to the larynx elevation and tongue retraction muscles has a large impact on post- radiation dysphagia

Materials and Methods: The inclusion criteria for this multi-center prospective study are (1) biopsy-proven NPC, (2) no distant metastasis, (3) in clinical remission after curative treatment (radiotherapy ± chemotherapy), (4) within 3 year after finishing radiotherapy. pEBV DNA concentration is monitored every 3 months. Detailed staging workups areperformed when abnormal pEBV DNA detected. All tumor recurrences are documented by imaging studies along with pathological verification if the lesions are accessible and patients agree. Results: From August 2011 to October2012, 252 patients were enrolled and 33 patients had abnormal pEBV DNA during follow-up visit. Thirty of 33 (91%) patients with elevated pEBV DNA have been proven as tumor relapse, whereas the remaining 219 patients with normal pEBV DNA level are showing no evidence of disease (P<0.0001). In addition, two-thirds (20/30) relapsed patients were detected in a symptomless state. Conclusions: pEBV DNA assay is a very encouraging tool in monitoring NPC patients after treatment.

PO-0723 POSSIBLE ADVANTAGES OF PROBABILISTIC PLANNING FOR HEAD AND NECK CANCER

To investigate possible advantages of uncertainty based probabilistic planning (PP) compared to conventional margin- based planning applied to five head and neck (HN instead, both systematic and random errors are explicitly handled during optimization and evaluation. We hypothesized that thanks to a PP approach we would be able to maintain the dose to the tumour target while decreasing the dose to the Organs at Risks (OAR)