AI Blood Analysis Paves the Way for Early Cancer Detection

Hong Kong people could soon benefit from affordable cancer screening technology, according to Professor Dennis Lo Yuk-ming of the Chinese University of Hong Kong, a pioneer in prenatal testing for Down syndrome.

Over the next three years, breakthrough advances in AI technology promise to make detecting malignant tumors easier and faster than ever before.

Professor Dennis Lo Yuk-ming said that AI has been helping researchers identify unique patterns in gene fragments found in plasma. This is a real revolution in the field of epigenetics, the study of the effects of non-genetic factors on genes.

The AI ​​approach allows researchers to decode epigenetic signals from DNA samples without the need for complex chemical treatments. This not only preserves up to 90% of genetic material but also opens up deeper insights into the mechanisms of cancer development.

"Our team wondered if some of the AI ​​methods developed for facial recognition could be applied to 'see' the signs we couldn't yet detect. And surprisingly, the answer was yes," said Professor Dennis Lo Yuk-ming.

Professor Dennis Lo Yuk-ming's research team at the Novostics Centre at Hong Kong Science and Technology Park has developed an in-house AI system based on Convolutional Neural Networks - a deep learning model specifically designed for image recognition.

Furthermore, with the advent of powerful generative AI tools like ChatGPT, the company has also integrated advanced neural network architecture into its system.

Like many biomedical companies pioneering drug discovery, diagnostics and therapeutics, Novostics is actively applying AI to the fight against cancer.

However, like many other companies in Hong Kong and mainland China, Novostics is facing significant challenges in accessing computing power due to US restrictions on access to advanced AI chips.

"Fortunately, we still have a solid system that has been developed over many years. Given the challenges of accessing the latest generation of Nvidia chips, we are forced to look for other manufacturers that can provide what we need," said Professor Dennis Lo Yuk-ming.

Dennis Lo Yuk-ming's research team is focusing on developing early cancer detection tests, based on a similar approach to his pioneering work using maternal plasma to screen for Down syndrome and other genetic disorders.

Dennis Lo Yuk-ming’s first target was nasopharyngeal carcinoma (NPC), a common tumor in southern China’s Guangdong province. Recognizing a suspected link between the disease and the Epstein-Barr virus, in 2013 he developed a test that could detect traces of the virus in blood plasma.

After four years of collecting 20,000 samples, the study concluded in July 2023 that virus screening is an effective tool for early cancer detection, reducing mortality from 40% to just 3.7%. The test is now available in Hong Kong at a cost of HK$1,500 (US$192.50).

"Although Hong Kong's medical system is relatively advanced, before this test was introduced, 75% of nasopharyngeal cancer patients were only detected at a late stage," said Professor Dennis Lo Yuk-ming.

He also stressed that AI has opened up the possibility of detecting many other types of cancer, even when no virus is directly causing the disease. "Finding individual cancer mutations is like hunting for a grain of sand in the desert, but AI has made it easier than ever."

After developing the NPC test, Cirina – a company founded by Professor Dennis Lo Yuk-ming in Hong Kong – merged in 2017 with Grail, a California-based company specializing in cancer detection.

The result of this merger is the Galleri test, which screens for more than 50 types of virus-related cancers. This achievement was named one of the Best Inventions of 2022 by Time magazine.

In 2021, Grail was acquired by San Diego-based biotech giant Illumina for $8 billion. Currently, Grail’s multi-cancer test is only available in the US market for $949, ​​but is not yet covered by most health insurance plans.

Not stopping there, Professor Dennis Lo Yuk-ming's team is looking to develop a lower-cost version of the test to reach mainland China and other markets around the world.

Novostics, the company he founded, created Fragma, a test that can detect DNA fragments from lung and liver cancer cells in plasma and urine.

To commercialize this technology, in 2023, Professor Dennis Lo Yuk-ming partnered with Prenetics Group, a US-based company listed on the world's largest stock exchange Nasdaq, to ​​establish a $200 million joint venture, Insighta.

In October 2023, Chinese tech giant Tencent Holdings invested $30 million in Insighta, valuing the startup at $200 million.

Based in the Hong Kong Science and Technology Park, Insighta is currently conducting large-scale clinical trials with the Fragma test in Hong Kong and mainland China.

The company's ultimate goal is to offer the test for $200, making it accessible to more patients.

Other products, such as a liver cancer test, are expected to be launched in the next two to three years, said Professor Dennis Lo Yuk-ming. “If we want to bring this technology to areas with limited resources, the most important thing is to make the cost more affordable.”