Abstract Background and purpose: Radiation pneumonitis and fibrosis are limiting factors in radiation treatment (RT) of thoracic tumors. The objectives of this study were to quantify the influence of irradiated lung volume and dose on lung response, and to evaluate the influence of other prognostic factors. Material and methods: Treatment histories of 55 thymoma patients were evaluated retrospectively for radiation pneumonitis and fibrosis. Complications were scored as pneumonitis if observed within 6 months of completion of RT, and as fibrosis if observed after 6 months. Complications were classified as ‘all pneumonitis’ and ‘all fibrosis’ if a patient either showed symptoms (such as chronic cough and dyspnea) or radiographic changes in lung. The second group scored as ‘symptomatic pneumonitis’ and ‘symptomatic fibrosis’ consisted of patients that exhibited clinical symptoms. Dose–volume data were estimated using representative anatomies combined with available individual dose data. The Lyman NTCP model was used to assess the dependence of lung complication incidence on dose and volume. Results: The derived values of the parameters governing dose–volume dependence for symptomatic complications agreed with currently accepted and recently published values within the margins of error. Dose–response curves for complications that included radiographic changes were less steep than for symptomatic complications. The volume dependence for symptomatic fibrosis was more pronounced compared to all fibrosis. A strong correlation was observed between developing pneumonitis and developing fibrosis. Conclusions: The long survival allowed the assessment of lung complication data in thymoma patients for both acute and late response. Mean dose in lung strongly correlated with lung complications that manifest clinically. The determination of the dose–volume dependence is affected by the choice of endpoints (i.e. whether complications are scored based on clinical symptoms or radiographic changes not accompanied by clinical symptoms).