Incidence and prognostic factors in severe drug-induced interstitial lung disease caused by antineoplastic drug therapy in the real world

PURPOSE: Investigate the frequency and prognostic factors of severe drug-induced interstitial lung disease (DILD) caused by antineoplastic drugs regardless of cancer types or type of drugs. METHODS: From 2014 to 2018, we reviewed patients with a history of antineoplastic agents administration in the real-world database of our hospital's electronic medical record and extracted patients who experienced "severe" DILD, requiring hospitalization with treatment or developed during hospitalization and required treatment. We collected patients' backgrounds, clinical and radiological features, laboratory data, treatment, and survival outcomes. RESULTS: 19,132 cancer patients received antineoplastic drug therapy during the study period, and 120 (0.62%) experienced severe DILD. The incidence of severe DILD in patients with thoracic cancer was highest among the patients included in this analysis (2.52% vs. 0.34% other cancers). Diffuse alveolar damage (DAD) pattern on CT was associated with higher mortality in patients with severe DILD compared with non-DAD pattern (hazard ratio [HR], 11.24; 95% CI, 4.82–26.2). Multivariate analysis revealed that the DAD pattern at diagnosis as severe DILD (HR, 3.59; 95% CI, 1.17–11.03), concurrent/previous interstitial lung disease (HR, 3.20; 95% CI, 1.27–8.10), and ECOG performance status of 2–4 (HR, 3.81; 95% CI, 1.10–13.17) were independent risk factors for mortality in patients with severe DILD. CONCLUSIONS: The frequency of severe DILD was highest in patients with thoracic cancer. The DAD pattern was associated with a poor outcome. From the perspective of DILD, special attention should be paid when administering antineoplastic agents to patients with thoracic cancer.