Abstract 117P
Background
The intestinal microbiome has a substantial influence on the cancer related response to immune checkpoint inhibitor (ICI ) therapy. Biome Diagnostics (BiomeDx) has developed a microbiome-based prediction test to ICI response. The biomarker was developed using datasets from stool samples of previously studied patients with non-small-lung-cancer (NSCLC), renal cell cancer (RCC), melanoma, and a healthy cohort (n=8000).
Methods
Stool samples were taken prior to ICI treatment initiation with a sampling kit using Norgen® tubes. Genetic sequencing of intestinal microbiota was performed with Illumina MiSeq® at a central laboratory. 1400 species were identified which were then transformed into a robust low-dimensional representation using an autoencoder. Machine learning (ML) algorithms were applied to determine a specific microbial signature, named BiomeOne®. Clinical response was assessed at the end of first-line therapy. In this study, responders were classified as complete and partial responders. Best response was compared with the outcome of BiomeOne® analysis.
Results
In this multi-centric study, 67 patients (29 f, 38m) encompassing NSCLC (42), RCC (15) and Melanoma (10) patients were enrolled. Patients were treated with ICIs targeting CTLA-4, PD-1 or PD-L1 as well as combinations with TKIs. Antibiotic therapy within 30 days prior to treatment was not allowed. Overall, 63 including 42 responders and 21 non-responders completed first-line treatment. BiomeOne® was able to identify responders to ICI treatment with a sensitivity of 81% and a positive predictive value of 77%. ML analysis did not depict significant differences between the 3 tumour types, suggesting that the biomarker is tumour agnostic.
Conclusions
The presented research demonstrates the potential of BiomeOne® as a novel, non-invasive test to predict the outcome of ICI therapy. Further research will be necessary to expand the scope of BiomeOne® to other tumour entities.
Clinical trial identification
Editorial acknowledgement
Legal entity responsible for the study
The authors.
Funding
Biome Diagnostics GmbH.
Disclosure
A. Valipour: Non-Financial Interests, Personal, Advisory Board: Biome Diagnostics. All other authors have declared no conflicts of interest.