The application of AI in this field allows researchers to recognize patterns in gene fragments found in blood plasma, significantly enhancing the study of epigenetics.

Lo's team is also developing a more affordable version of the Fragma test, which detects lung and liver cancer DNA fragments in blood plasma and urine, aiming for a price point of US$200.

The AI system developed by Lo's team integrates advanced machine learning techniques, including Convolutional Neural Networks and transformer architectures, inspired by generative AI tools like ChatGPT.

Despite Grail being acquired by Illumina for US$8 billion in 2021, its multi-cancer test remains limited to the US market, priced at US$949 and not widely covered by healthcare plans.

Currently, a test for nasopharyngeal cancer (NPC) is available in Hong Kong for approximately US$192.5, and it has dramatically reduced mortality rates from 40% to 3.7%.

Leading this innovation is Professor Dennis Lo Yuk-ming, a pioneer in prenatal testing, who is utilizing AI to analyze gene fragments from blood plasma for early cancer detection.

Challenges persist, particularly in accessing advanced AI chips due to US restrictions, yet Lo's team continues to innovate in cancer detection technologies.

Lo's company, Cirina, merged with Grail in 2017, creating a global entity focused on early cancer detection, leading to the development of the Galleri test for over 50 cancer types.

Market forecasts indicate significant growth in the cancer diagnostics sector, driven by advancements in AI methodologies, which are expected to enhance public health strategies.

This AI-driven method enables the decoding of epigenetic signals from DNA samples without the harsh chemical treatments that typically destroy up to 90% of genetic material.

Clinical tests for the Fragma test are already underway in Hong Kong and mainland China, with a target launch within the next two to three years.

Advances in artificial intelligence (AI) are set to revolutionize cancer screening in Hong Kong, with predictions that affordable options could be available within three years.