ESTRO 36 Abstract Book

S350 ESTRO 36 _______________________________________________________________________________________________

and 37.5 months in surviving patients, the majority of patients (90%) have died. One-year overall survival and local control were 45% and 30% respectively. However, high-dose SBRT was associated with improved local control (p < .01), and the one-year overall survival and local control were 77.8% and 66.7% respectively in this sub- group. Furthermore, an increased mean interval between initial treatment and SBRT was observed in patients who achieved durable local control (41.9 vs. 13.4 months, p < .01). While treatment was generally well tolerated, there were two cases of radiation pneumonitis (grade 2) and two cases of recurrent laryngeal nerve paralysis (grade 2 and 3), all of which resolved prior to last follow-up or death. No late esophageal toxicity was noted. A patient who received an SBRT dose of 45 Gy (total EQD2: 129.7 Gy) experienced cardiopulmonary death 35 months following treatment which was attributed to radiation toxicity. Conclusion Although the overall prognosis for patients with in-field central NSCLC recurrences following CF-EBRT remains grim, five-fraction SBRT was well tolerated with an acceptable toxicity profile. Dose escalation above 35 Gy may offer improved local control, however caution is warranted when treating high-risk recurrences with aggressive regimens. Our findings support the efficacy of five-fraction SBRT re-irradiation reported by Trovo et al. [Int J Radiat Oncol Biol Phys. 2014 Apr 1;88(5):1114-9]. PO-0669 Models of pulmonary function changes after thoracic radiotherapy A.G.H. Niezink 1 , O. Chouvalova 1 , J.F. Ubbels 1 , A.J. Van der Wekken 2 , J.A. Langendijk 1 , J. Widder 1 1 UMCG University Medical Center Groningen, Radiation Oncology, Groningen, The Netherlands 2 UMCG University Medical Center Groningen, Pulmonary Diseases, Groningen, The Netherlands Purpose or Objective Reproducibly measuring pulmonary toxicity remains challenging in thoracic radiotherapy. Pulmonary function tests may render objective parameters to assess pulmonary radiation toxicity. Our aim was to establish a model predicting post-radiotherapy forced expiratory volume in one second (FEV1) and diffusion capacity (DLCO). Material and Methods Patients with both baseline and follow-up FEV1 and/or DLCO available were included from a prospective data registry (clinicaltrials.gov). Patient and tumour characteristics as well as dose-volume parameters and survival data were available. Changes in pulmonary function tests were calculated using a paired t-test, and univariable and multivariable linear regression models were built predicting pulmonary function test changes. Multicollinearity was tested using the variance inflation factor and the quality of the models were compared using adjusted R-square and the Akaike information criterion (AIC). Results Baseline and follow-up FEV1- and DLCO-data were available for 379 and 283 patients, respectively, who were treated between 2013 and 2015 for (N)SCLC stage I-III with (chemo)radiotherapy or SABR. Both FEV1 and DLCO declined significantly after treatment (p=0.001 and p<0.001). WHO-performance status (2-3 versus 0-1), chemotherapy (yes versus no), smoking (never versus former or current), technique (SABR versus external beam RT), GTV, lung dose-volume parameters (V5, V20, V30, V40, mean lung dose) and heart volume parameters (V5, V40 and mean heart dose) were significant factors predicting follow-up DLCO after adjustment for baseline DLCO. The best model, based on multivariable linear regression for predicting follow-up DLCO, contains baseline DLCO, WHO-performance status and lung V5 (adjusted R-square=0.71, p<0.0001) [Figure 1]. Univariable and multivariable linear regression showed

that baseline FEV1 and lung V40 are significant factors predicting follow-up FEV1 (adjusted R square = 0.21, p<0.0001).

Figure 1 : Baseline DLCO and post-radiotherapy DLCO decline by V5-lung for three different scenarios. Abbreviations: DLCO Fu and DLCO BL = diffusion capacity of carbon monoxide corrected for hemoglobin concentration at follow-up and baseline; WHO-PS=WHO performance score (binary: 2-3 versus 0-1); V5Lung= percentage of lung volume receiving 5Gy or more. p25 / mean / p75 = 25 th percentile, mean and 75 th percentile of the V5 Lung. Conclusion FEV1 and DLCO both decline after thoracic radiotherapy, and DLCO decline is predictable based on a well- performing (adjusted R-square=0.71) linear-regression model including the V5-lung. Limiting post-radiotherapy DLCO decline would require dramatic reduction of low lung dose, which might only be achievable using protons. PO-0670 CPAP ventilation might allow better sparing of normal lung tissue during lung cancer radiotherapy D. Di Perri 1,2 , A. Colot 2 , A. Barragan 1 , G. Janssens 3 , V. Lacroix 4 , P. Matte 5 , K. Souris 1 , X. Geets 1,2 1 Université catholique de Louvain, Center of Molecular Imaging Radiotherapy and Oncology MIRO Institut de Recherche Expérimentale et Clinique IREC, Brussels, Belgium 2 Cliniques universitaires Saint-Luc, Department of Radiation Oncology, Brussels, Belgium 3 Ion Beam Applications, Louvain-La-Neuve, Belgium 4 Cliniques universitaires Saint-Luc, Department of Cardiovascular and Thoracic Surgery, Brussels, Belgium 5 Cliniques universitaires Saint-Luc, Cardiovascular Intensive Care, Brussels, Belgium Purpose or Objective Lung toxicity is a major dose-limiting factor in lung cancer radiation therapy (RT). By increasing lung volume, continuous positive airway pressure (CPAP) ventilation during treatment might allow better sparing of the normal lung parenchyma. However, CPAP might also influence respiration-induced tumour motion amplitude and/or tumour position reproducibility. In this study, taking stage I lung cancer patients as a model, we evaluate the effect of CPAP ventilation on lung volume, tumour motion amplitude, and tumour position reproducibility. Material and Methods Stage I lung cancer patients referred for stereotactic body radiation therapy underwent two 4D-CT scans (with and without CPAP) at two time-points: during the treatment preparation session (T0) and the first day of treatment (T1), resulting in four 4D-CT scans per patient (noCPAP T0 , CPAP T0 , noCPAP T1 , and CPAP T1 ). All images were reconstructed in their time-averaged midposition (MidP) for subsequent analysis. Gross tumour volumes and lungs were delineated on each MidP scan.