As a physician and a scientist, I have spent over a decade researching new and more effective ways of treating cancer. I’ve worked with some of the best clinicians and oncology experts in the world – across Oxford, Toronto, and the World Health Organization. Despite this, there was little I could do to save my uncle from cancer. Despite facing many hardships in his life, he was the kind of man who was unerringly kind to those around him. His death made me realize I wanted to dedicate my life to developing lifesaving innovations in the field of cancer research, helping those who needed it the most.  

I’m not alone in my loss. According to the Canadian Cancer Society, almost half of Canadians will develop cancer at one point in their life. Nearly a quarter of Canadians will die from it. However, despite great progress having been made over the past century to manage this terrible disease, with greater treatment options available every year, more people – not fewer – are dying from cancer in Canada. 

Why aren’t we saving more family, friends and loved ones? As the population gets older, cancers are still being diagnosed too late, when they have already spread across the body and become incurable. For example, lung cancer – the leading cause of cancer deaths worldwide, resulting in a quarter of Canadian cancer deaths – is often asymptomatic, with few signs until it has spread. 

Being able to diagnose cancers earlier when they can still be treated would drastically improve patient’s survival rates and their quality of life during treatment. This fact led me to co-found Oxford Cancer Analytics (OXcan), a company that uses machine learning to develop new blood tests capable of detecting cancer-signaling proteins, earlier and better than other forms of testing.  

A person’s bloodstream holds a wealth of information that can be used to diagnose disease. However, our ability to unravel and find patterns within that information has historically been limited: Some scientists have attempted to detect lung cancer early by analyzing DNA shed into the blood by tumours. However these tests have sometimes been found to give inconclusive results, potentially because less DNA is shed during the early stages of lung cancer than in later stages.

At OXcan, we’ve decided instead to focus on developing protein-based blood tests: Proteins are found at higher rates in our blood and can provide detailed insights at all stages of cancer. Our new tests allow us to take a patient’s blood and take a closer look at approximately 20 proteins in their blood, which can determine their risk of developing lung cancer. We determined that these proteins were well-positioned to signal cancerous growth by first examining over 5,000 kinds of proteins in other blood samples – an abundance of data that gave us great opportunities to identify novel and measurable indicators of disease. It’s the difference between watching a movie on a 4K flat screen versus a Game Boy screen: the higher resolution broadcasts more information, allowing for a clearer picture to form. High-quality, large datasets like these allow us to deploy machine-learning models in our research to find the most suitable proteins for cancer detection. Machine learning, i.e. computer programs that recognize patterns in data to improve their performance over time, excel at recognizing complex patterns that are not immediately obvious to human analysts or other forms of testing. It’s perfect for pinpointing the right information for disease detection. 

When it comes to people’s health, it’s normal to be apprehensive about the black box that is AI and machine learning, but being able to explain our model’s conclusions is critical to our mission. Tests we’ve run show that our machine learning models are able to detect lung cancer biomarkers with over 90 percent accuracy – rates up to three times better than reported DNA-based blood tests. 

When OxCan was first starting out, we decided that joining the private sector was our best means of funding the development of our blood tests. We also felt that incorporating as a private company was most conducive to partnering with the medical and diagnostic companies needed to distribute our products. Despite this, securing financial investors for a medical product like ours was hard because while we preferred to partner with those who had medical or scientific experience, those we pitched weren’t always experts in our field. In those early days, we pitched well over 100 investors and successfully landed a few across the United States, United Kingdom and Canada, including Eka Ventures, and Civilization Ventures. 

We also received grants from agencies such as Innovate UK and Cancer Research UK to support this endeavour, and we’d like more Canadian government grants, as we believe they could be pivotal to expediting medical innovation for those who need it the most. We are glad that our investors agreed with our vision because they’ve helped us attract over seven million in funding and transform our prototype blood tests into products that can save lives. We’re now working to launch in North America, the European Union and the UK, with eyes on Asia and Africa soon after. 

We’re specifically working to get our products into the hands of primary care networks, family doctors, and diagnostic labs because our blood tests are designed for seamless clinical implementation: While the technology behind our tests is unique, they’re administered by a doctor or diagnostic lab like any other blood tests and don’t require any new resources or equipment. 

Blood tests for lung cancer are only the first step. This workflow can be applied to other cancers and diseases to re-invent the diagnostic landscape: at costs lower to produce than other forms of testing like complex imaging, we not only lower the barrier to diagnosis – testing more people earlier, regularly, and affordably – we also alleviate pressure on our healthcare systems because treating cancer in its late stages is more expensive than in its earlier stages. 

The ultimate vision is to drastically transform modern disease management from a reactive to proactive approach, identifying signs of early disease development for more effective treatment and prevention. I believe we can fundamentally transform cancer detection and management to save lives and create hope for millions impacted by cancer.