Grant: Project Grant

Recent progress in cancer detection and treatment has led to a new issue. People who have survived cancer for many years now have a higher chance of getting a completely different type of cancer later on in life. This is called "multiple primary cancers." Figuring out if a person has a recurrence of their old cancer or a new one is tough. Getting it right matters because the treatment given depends on the answer, and this makes a big difference to how well the cancer responds to treatment. A similar problem occurs with a type of cancer called "cancer of uncertain primary" (CUPS). This is the name for a situation in which a patient's cancer is discovered after it has spread to other parts of the body, but the doctors can't find the original location where it started. Recent studies have shown that between 2% and 17% of people with a history of cancer eventually develop multiple primary cancers, and between 2% and 5% might have CUPS. This means around 10,000 people with cancer in Canada each year face these challenges. To help with these issues, we have developed a computer program called DeepTumour. It uses an advanced AI technique called deep learning to predict where a tumor comes from. It looks at signals from changes in the tumor's DNA and can tell apart 28 different types of cancer using data from the patient's tumor DNA. It's right about 88% of the time for tumours that have not yet spread, and 83% of the time for tumors that have. This project seeks to improve DeepTumour by applying a method called deep plasma whole genome sequencing (pWGS) to measure small amounts of tumor DNA in the circulating blood. This will allow Using DeepTumour to identify the source of a cancer from a non-invasive blood test, and by guiding doctors to the potential source of the cancer, to help them choose the therapy that has the greatest chance of success.