A retrospective clinical review of drug induced interstitial lung disease at a tertiary cancer centre.

Abstract

Introduction: Drug induced interstitial lung disease (DIILD) is a well-recognised complication of cancer therapy with an unknown incidence. There are no standardised diagnostic criteria and management is variable. We retrospectively reviewed the clinical presentation, aetiology, management and outcomes for consecutive cases presenting to The Christie NHS Trust to inform future practice and areas for research.

Methods: CT thorax requests and reports between 1/1/15 and 31/12/16 were screened for the terms ‘pneumonitis’ or ‘drug induced’. Patient demographics, clinical symptoms, diagnosis, cancer treatment, DIILD management and outcome data were collected. CT images were independently reviewed by a specialist radiologist and cases re-classified as ‘probable’, ‘possible’ or ‘unlikely’ for DIILD based on imaging features.

Results: 199 CT thorax reports were identified. Duplicates (n=69) and radiotherapy-only treated cases were removed (n=11). 119 cases were re-classified as probable (n=22), possible (n=16) or unlikely (n=81. DIILD was clinically suspected in ~50% of cases, but symptoms were non-specific. Everolimus, bleomycin, pembrolizumab, gefitinib and paclitaxel were the most common causative drugs. Management varied widely. 90% of DIILD patients had treatment withdrawn, 50% were hospitalised and 74% received corticosteroids. Of symptomatic patients, 63% resolved, mostly within 3 months. All-cause mortality was 16%.

Conclusions: Results highlight important unmet clinical and research needs to develop management protocols to unify practice and CT imaging biomarkers to improve diagnostic accuracy and clinical outcomes.