109O - A multi-modal nanopore-based liquid biopsy assay in cancer: Initial findings

Background

Current liquid biopsy assays are generally restricted to a single aspect of detection: somatic SNV/Indels, sCNA, SV, fragmentomics or methylation. These have been used in diagnosis, therapy, minimal residual disease detection (MRD) and screening - each use case relies on a specific asset of an assay. Utilising a custom mult-modal Nanopore sequencing assay, we present the initial results of patients sequenced using this assay and demonstrate that it can successfully detect all variations to a high degree of sensitivity and accuracy.

Methods

Per participant, 1- 4 ml of plasma was extracted, and DNA was library prepared using an in-house protocol utilising components of the Oxford Nanopore LSK114 kit to enable maximum ligation of DNA; libraries were sequenced and library product was recovered to allow for panel sequencing of a 54-gene DNA panel. Bioinformatic analysis was performed using a proprietary pipeline, which calls somatic SNV/Indels, CNV, SV, tumour fraction (TF) and tumour type classification, fragment size metrics and methylation, and cell of origin detection. Available short-read sequencing of the same samples was used as a reference.

Results

In total, 354 samples underwent sequencing. Median genome-wide read depth varied between 0.01-4x and correlated to DNA input mass. TF varied between 4-85%, and histological subtypes could be differentiated using genomewide patterns of mutation, methylation and sCNA. Comparison between samples undergoing low pass WGS sequencing for sCNVs detected concordance. Significant differences (p< 0.05) in fragment length were seen between tumours & controls, irrespective of TF. MRD detection of recurrence was successful compared to reference time-course samples. Benchmarking using controls demonstrated > 95% sensitive & specific detection of SNV and Indels to a VAF of < 1%.

Conclusions

Nanopore sequencing-based multi-modal assays are comparable to current standard-of-care assays and offer advantages in the detection of multiple classes of variants over competitor assays, including input from significantly lower DNA mass input (5ng) and plasma volumes. The cost per sample is below $100, making this assay feasible for screening as well as therapy selection and treatment tracking.

Editorial acknowledgement

Clinical trial identification

Legal entity responsible for the study

The authors.

Funding

Cancer Research UK Medical Research Council, UK.

Disclosure

A.D. Beggs: Financial Interests, Institutional, Research Grant: Oxford Nanopore; Financial Interests, Personal, Invited Speaker: Oxford Nanopore, Illumina; Financial Interests, Personal, Advisory Role: Oxford Nanopore. All other authors have declared no conflicts of interest.