Margriet Kwint

Acute esophagus toxicity after IMRT and concurrent chemoradiation 117 6 Scoring of acute esophagus toxicity AET was scored using the Common Toxicity Criteria 3.0 from start of treatment, until 3 months after. Grade 2 was scored in case of symptomatic and altered eating and intravenous fluids indicated for a period shorter than 24 hrs. Grade 3 included symptomatic and severely altered eating/swallowing and intravenous fluids, tube feedings, or total parenteral nutrition indicated ≥24 hrs; and grade 4 included life- threatening consequences. The patients were examined and toxicity was scored at baseline and weekly during treatment, until 3 weeks after treatment by the treating physician. Thereafter the patients were followed with 2 months intervals or more frequently if indicated. All patients consulted a dietician at least twice during treatment. Dosimetric analysis The physical RT-dose was converted to Normalized Total Dose (NTD) for 2 Gy per fraction with an α/β-ratio of 10 Gy for acute toxicity. With the NTD corrected dose, esophageal dose-volume-histograms (DVH) were computed and dose-volume- parameters were derived in steps of 5 Gy from V5 to V70, as well as the D mean and D max . For comparison with 3DCRT, DVH parameters of the current study using IMRT were compared to the data of 36 of the 37 patients treated with CCRT in the historical dataset (data for 1 patient was missing) [2]. Statistical analysis To evaluate the introduction of IMRT in the CCRT-protocol, we compared the incidence of grade 2 and 3 AET with the historical patient data [2] using a chi-squared test. The statistical analysis of AET predictability was performed in two steps. First, the V5- V70, D mean and D max were analyzed for correlation with the AET grade using Spearman’s rank correlation coefficients. In the second step, the best dosimetric predictors for grade ≥2 and ≥3 AET were estimated, using a stepwise logistic regression method. The stepwise regression was done in a forward selection fashion, which involves starting with all candidate variables and testing them one by one for statistical significance, deleting variables that were least significant until the best predictor remained. The resulting logistic function is expressed as: