Background Three months after radical radiotherapy for lung cancer, 50–60% of patients have radiation pneumonitis (RP) on CT thorax. Our aim was to assess the clinical and dosimetric factors associated with radiologically-defined RP. Our primary endpoint was the development of new infiltrates on CT thorax at 3 months following radiotherapy.
Methods 161 patients with lung cancer were referred for radical radiotherapy during 2009–2010. Exclusion criteria were previous thoracic radiotherapy or surgery, palliative radiotherapy, or missing dosimetric or CT data.
Information on medical history, lung function and date of death were taken retrospectively from electronic notes. Dosimetric parameters V20-Lung (percentage normal lung exposed to more than 20Gy), V5-Lung and Mean Lung Dose were derived from treatment planning dose-volume histograms. Development of RP was defined as an increase in the percentage lung volume occupied by consolidation or ground glass on post-radiotherapy CT. Student’s t-test and Fisher’s Exact Test were used to define variables which were associated with RP prior to logistic regression analysis.
Results 98 cases were included in analysis. 86% had non-small cell lung cancer, 44% had chronic obstructive pulmonary disease (COPD), and 27% smoked. 49/98 (50%) patients developed RP on CT at median 90 days post-radiotherapy.
The factors which had a significant positive correlation with RP on univariate analysis were V20-lung, V5-lung and MLD: these were best represented using V20-Lung ≥22%. Current smoking, poor performance status and having COPD had a significant inverse correlation with RP. Use of statins or inhaled Long Acting β2 Agonists, and the presence of moderate-severe radiological emphysema also approached significance: these were included in regression analysis.
After logistic regression, the factors which had a significant correlation with RP were V20≥22% (OR 6.45, 95%CI 2.22–18.08), current smoking (OR 0.23, 95%CI 0.07–0.79), and statin use (OR 0.30, 95%CI 0.102–0.863).
Neither RP nor any other variable was associated with post-radiotherapy mortality.
Conclusions This study confirms that V20≥22% is associated with the radiological development of RP. In addition, patients who smoked, and those taking statins were significantly less likely to develop RP. A potential role for statins in modifying radiotherapy side effects deserves further attention.