Early Detection

Methylation Based Early Detection: The Future of Cancer Detection

It has become increasingly clear in the scientific community that our ability to treat and combat cancer in a patient changes dynamically as cancer develops over time. Though cancer’s development is complex, warranting ever intricate solutions for treatment strategy, one fact is evident in this battle: detecting cancer at its earliest stages positions us to have the greatest capability for responding effectively.

Burning Rock’s Approach

We understand that every cancer has a distinctive signature in the body, one that is multiplexed in biology, genetics, epigenetics, and chemistry. Thus, our approach for identifying cancer reflects that complexity. We target the genome utilizing Next Generation Sequencing to distinguish methylation patterns among cancers and non-cancers in order to guide the diagnostic modality of a patient based on suspected tumor origin.

The Determination and the Results

The task of distinguishing the cancerous from the non-cancerous is made difficult by cancer’s sparsity in relation to the entirety of a human body; the endeavor to identify cancerous loci in an otherwise healthy genome is comparable to that of finding a needle in a haystack. We began this uphill battle in the Thunder Project, involving over 7000 subject enrollments. The goal was to generate a statistical model-based methylation pattern. Utilizing high throughput bioinformatics and machine learning for training and validation, a high dimensional modeling methylation pattern database was generated. This database allows for subsequent methylation patterns of unknown clinical information to be queried to determine the best fit.

We demonstrated that our assay is capable of blood-based multi cancer detection with a specificity of greater than 98%. We also demonstrated the ability of our assay to accurately predict the diseased organ offering guidance for subsequent diagnosis. Currently, we are completing the Phase III of the project and are in the process of starting commercial clinical testing in China. We are concurrently running further studies such as the PREDICT Project to expand the assay to 9 cancer types and we seek to expand the testing to other populations around the world.

Assay performance:

Our current ELSA-SEQ platform has shown promising results. Our assay is competitive with other early detection assays such as Grail’s assay and even outperforms it in the detection of similarly staged colo-rectal and liver cancer.

Challenges/Requirements for an effective Early Detection Assay at Scale

In order to be an effective tool, an early detection assay must be non-invasive, have both high sensitivity and specificity, as well as identify the correct origin of the tumor. The assay must cover a wide array of cancers to offer useful data for researchers, doctors, and patients alike. Current early detection assays in the field struggle with detecting the very early stages of cancer that they are designed to detect. Often falling below a 50% sensitivity and specificity for stage 1 cancer. However, as the amount of data gathered on methylation patterns in the population increases, the use of early detection assays will become common practice.

Contact Us

We look forward to hearing from you. Whether you are an academic collaborator or a potential biopharma partner, feel free to reach out and our team will get back to you as soon as possible.

Call Us


Send Us A Message