ASPO Abstracts
Antibiotic use and clinical response to immune checkpoint inhibitors in NSCLC
Category: Molecular Epidemiology & Environment, Early Detection & Risk Prediction
Conference Year: 2018
Abstract Body:
Preclinical studies have demonstrated that broad-spectrum antibiotics (ATB) may modulate the efficacy of immune checkpoint inhibitors (ICI) by disrupting the gut microbiome. However, the impact of ATB in patients with advanced non-small cell lung cancer (NSCLC) remains unclear. Our study assessed the effect of pre-treatment antibiotic use on the clinical response of NSCLC patients treated with ICI. We retrospectively examined medical records of >400 patients. ATB use was assessed from 2 months before and 1 month after 1st ICI injection and patients were categorized as those who received or did not receive antibiotics. The best clinical response recorded within six and 12 months of ICI start was abstracted and patients were categorized as responders (complete [CR] or partial response [PR]) or non-responders (stable [SD] or progressive disease [PD]. To assess the association between antibiotic use and clinical response, Cox proportional hazards regression was used in addition to Kaplan Meier analyses of progression-free (PFS) and overall survival (OS). Included in this analysis were 257 eligible patients treated at Moffitt Cancer Center with anti- PD-1/PD-L1 alone, or in combination with anti-CTLA-4, between 2011 and 2017. Half of all patients were male (53%), a median age of 67 years, and 94% diagnosed with stage IV disease. Pre-treatment antibiotic use was recorded among 47 (18%) patients and 79 (31%) were considered ICI responders. In univariate analyses, antibiotic users were 20% more likely to be a non-responder compared to patients who did not use antibiotics (HR: 1.20 (0.81-1.76)); however, this association was not statistically significant. Median PFS and OS were higher among patients who did not use antibiotics than those who used antibiotics (112 vs. 90 days and 524 vs. 348 days, respectively), although these differences were not statistically significant (p=0.91 and p=0.20, respectively). To our knowledge, this is the largest study to evaluate the effect of pre-treatment antibiotics on clinical response to ICI among patients with NSCLC. Although not statistically significant, antibiotic use appears to be associated with poorer prognoses among patients with NSCLC. Multivariable analyses will provide additional insight into this association.
Keywords: Lung cancer; immunotherapy; antibiotics; microbiome