Impact of antibiotics, corticosteroids, and microbiota on immunotherapy efficacy in patients with non-small cell lung cancer

Abstract

Lung cancer is a leading cause of morbidity and mortality globally, with its high mortality rate attributed mainly to non-small cell lung cancer (NSCLC). Although immunotherapy with immune checkpoint inhibitors (ICI) has revolutionized its treatment, patient response is highly variable and lacking predictive markers. We conducted a prospective study on 55 patients with NSCLC undergoing ICI therapy to identify predictive markers of both response and immune-related adverse events (IrAEs) in the airway microbiota. We also analyzed the clinical evolution and overall survival (OS) with respect to treatments that affect the integrity of the microbiota, such as antibiotics and corticosteroids. Our results demonstrated that respiratory microbiota differ significantly in ICI responders: they have higher alpha diversity values and lower abundance of the Firmicutes phylum and the Streptococcus genus. Employing a logistic regression model, the abundance of Gemella was the major predictor of non-ICI response, whereas Lachnoanaerobaculum was the best predictor of a positive response to ICI. The most relevant results were that antibiotic consumption is linked to a lower ICI response, and the use of corticosteroids correlated with poorer overall survival. Whereas previous studies have focused on gut microbiota, our findings highlight the importance of the respiratory microbiota in predicting the treatment response. Future research should explore microbiota modulation strategies to enhance immunotherapy outcomes. Understanding the impact of antibiotics, corticosteroids, and microbiota on NSCLC immunotherapy will help personalize treatment and improve patient outcomes.

Keywords: Airway microbiota; Antibiotics; Corticosteroids; Immunotherapy; Lung cancer.

Fig. 1

Oral microbiota. A: Bacterial phyla distribution for ICI responders and non- responders. B: Alfa and beta diversity statistical análisis among ICI responders and non- responders.

Fig. 2

LEfSe analysis of the oral microbiota between ICI responders (in green) and non-responders (in red). (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Fig. 3

Percentage of the most relevant genera with statistical differences between ICI responders and non-responders. A: Streptococcus; B: Porphyromonas; C: Fusobacterium.

Fig. 4

Kaplan–Meier survival curve graphs. A) Survival curves based on Eastern Cooperative Oncology Group scale. B) Survival curves based on tumor stage. C) Survival curves based on the type of immunotherapy. D) Survival curves based on the best-achieved response. The numerical data and p-values are specified in the text above.

Fig. 5

Kaplan–Meier overall survival curve graphs. A) Survival curves based on use or non-use of antibiotics. B) Survival curves based on cycles of antibiotics. C) Survival curves based on timing of use.

Fig. 6

Kaplan–Meier overall survival and progression-free survival curve graphs. A) Survival curves based on the use or non-use of corticosteroids. B) Survival curves based on reasons for its use. C) Survival curves based on timing. D) Progression-free survival based on use or non-use of corticosteroids. E) Progression-free survival based on reasons for its use. F) Progression free survival based on timing.

Fig. 7

Correlation between immune checkpoint inhibitor response and immune-related adverse events. Chi-square test p = 0.06.

Fig. 8

Machine learning results for ICI response prediction. A: Confusion matrix, B: Linear discriminant analysis (LDA) representing the significant differences in the abundance of each species.