[OA077] Modelling the risk of fatal acute toxicity following radiotherapy of lung cancer. (August 2018)

• Record Type: Journal Article
• Authors: Stervik, Louise; Pettersson, Niclas; Scherman-Rydhög, Jonas; Behrens, Claus F.; Ceberg, Crister; Vogelius, Ivan; Bäck, Anna
• Is Part Of: Physica medica. Volume 52(2018)Supplement 1
• Page Start: 30
• Publication Date: 2018-08
• Subjects: Medical physics -- Periodicals; Biophysics -- Periodicals; Biophysics -- Periodicals; Imagerie médicale -- Périodiques; Radiothérapie -- Périodiques; Rayons X -- Sécurité -- Mesures -- Périodiques; Physique -- Périodiques; Médecine -- Périodiques; 610.153
• DOI: 10.1016/j.ejmp.2018.06.149
• Languages: English
• ISSNs: 1120-1797

Abstract

Abstract : Purpose: To model the risk of fatal acute toxicity after conventionally fractionated curative radiotherapy of patients with non-small-cell lung cancer (NSCLC). Methods: Scripting was used to automatically extract treatment-related data for all patients treated for NSCLC between 2008 and 2016 from three hospitals. Inclusion criteria were conventionally fractionated curative radiotherapy and no prior treatment in the thorax region. Maximum likelihood estimation and logistic regression (LR) were used to model the risk of fatal acute toxicity defined as death within 90 days from treatment start. Mean lung dose (MLD), patient age and the volume of the gross tumour volume (GTV) were investigated as predictors in a univariable LR analysis. We performed analysis on the data from the three hospitals separately and merged. Predictor variables were considered statistically significant if p < 0.05. All predictor variables with a p < 0.1 were further analysed in multivariable LR models. Confidence intervals (CIs) for the predictors were calculated using the likelihood ratio test. CIs for the multivariable models were calculated by bootstrapping. Results: Data was extracted for 848 patients. The incidence of death within 90 days from treatment start was 3.8% (32/848) for the merged data set and varied from 1.8% to 5.4% between hospitals. For the hospital with the highest incidence, a statistically significant relationship between MLD and the risk of fatal acute toxicity ( p = 0.020) was found. The model parameters and their 95% CIs were D 50 = 42.8 (31.4–167.7) Gy and γ 50 = 1.20 (0.77–1.68). In the univariable LRs with patient age and GTV volume, patient age had p-values < 0.1 and GTV volume p-values > 0.2. Multivariable LR with MLD and patient age resulted in a statistically significant multivariable model (p = 0.005) for the merged data. The calculated risks and their 95% CIs for a patient with MLD = 20 Gy were 1.6% (0.5–3.3%), 2.8% (1.3–4.5%), 4.9% (3.1–6.8%), and 8.6% (4.8–13.5%) at 50, 60, 70, and 80 years of age, respectively. Conclusions: A statistically significant multivariable model quantifying the risk of fatal acute toxicity with MLD and patient age as predictor variables was found.