Risk of Radiation Pneumonitis in Locally Advanced Lung Cancer Patients: Role of Interstitial Lung Disease Subtypes and Severity of Pulmonary Emphysema

Abstract

Interstitial lung disease (ILD) is a known risk factor for radiation pneumonitis (RP). In this study, the Fleischner Society new diagnostic criteria for idiopathic pulmonary fibrosis (IPF), a subtype of ILD, are applied to pretreatment diagnostic CT scans of lung cancer patients to assess their predictive value for grade 3+ RP. In addition, the severity of pulmonary emphysema (PE) is radiographically graded and correlated to RP as well. The pretreatment diagnostic CT chest scans of 121 locally advanced lung cancer patients with definitive chemoradiotherapy between 2009 and 2018 were retrospectively analyzed by a thoracic radiologist. CTs were classified into typical usual interstitial pneumonia (UIP), probable UIP, indeterminate for UIP, and non-IPF. Typical and probable UIP patterns support the diagnosis of IPF in a typical clinical context. In addition, the location (upper lobe, lower lobe, or both), severity (no PE (score 0), ⩽25% of lung (score 1), 26–75% (score 2), ⩾75% (score 3)), and morphology (predominantly centrilobular vs. predominantly paraseptal) of PE were recorded. RP was recorded as RTOG grade 3+ pneumonitis requiring steroid treatment. Univariate Cox (UVA) and logistic regression were applied to predict for an effect of ILD and PE on overall survival (OS) and RP, respectively. Median dose prescribed was 60Gy (33-78Gy), median lung was V20Gy of 27.7% (20.1-30.0). Twenty patients (16%) experienced RP≥G3 and 17 (13.6%) had radiographic evidence of ILD. Most (13/17, 76%) were classified as non-IPF. The four patients with UIP diagnoses were classified as indeterminate for UIP (4/17, 24%). On logistic regression, ILD was associated with RP (p = 0.08). No significant difference in RP between UIP and non-IPF diagnoses was seen (p = 0.36). On univariate analysis, the presence of ILD or subtype of ILD (UIP vs. non-IPF) did not predict for OS (p = 0.19 and 0.85 respectively). 113 patients (93.4%) had radiographic PE, 96 (84.2%) of which had upper lobe emphysema and 110 (89.4%) of which had PE in a predominantly centrilobular pattern. Eight patients (6.4%) were assigned score 0, 52 (41.6%) score 1, 42 (33.6%) score 2, 19 (15.2%) score 3. The diagnosis, location, morphologic subtype, and severity of PE did not predict for RP≥G3 (p = 0.75, 0.29, 0.93, and 0.36, respectively) and were not significantly related to OS (p = 0.5, 0.85, 0.97, and 0.7, respectively). As a control, known dosimetric predictors of RP were significantly correlated (lung V5Gy/V10Gy/V20Gy; p = 0.005, 0.003, and 0.014, respectively). In the current study, there were no significant differences between UIP and non-IPF diagnoses in regard to clinical RP risk, potentially because none of our patients had UIP features more strongly associated with IPF. There was also a trend towards presence of radiographic ILD and development of RP. In addition, our data indicate no increased pneumonitis risk for patients with emphysema regardless of location or severity.