AI system developed in Singapore could help predict liver cancer recurrence
Dr Joe Yeong is a co-author of the study published in the peer-reviewed scientific journal Nature.
SINGAPORE - A scoring system powered by artificial intelligence (AI) that was developed by researchers here could help predict the recurrence of a common form of liver cancer.
Developed by scientists from A*Star's Institute of Molecular and Cell Biology (IMCB) and Singapore General Hospital (SGH), the system can predict relapses of hepatocellular carcinoma (HCC).
Affecting more than 20 out of every 100,000 people here annually, HCC is the fourth-most-common cancer among men in Singapore, as well as the third-most-common cause of cancer deaths among males here.
The Tumour Immune Microenvironment Spatial (Times) score analyses the precise spatial distribution, or the exact location, of immune cells called natural killer (NK) cells and five specific genes within liver tumour tissues.
NK cells are a type of white blood cell that destroys diseased cells, such as cancer cells, and a lower number of NK cells is an indicator of a higher risk of relapse.
'Instead of just counting how many NK cells are present, we analyse exactly where they are positioned relative to the cancer cells,' said Dr Joe Yeong, a principal investigator at both A*Star IMCB and SGH's anatomical pathology department.
'By understanding how NK cells are distributed and how they interact with cancer cells – for example, whether they are close enough to attack the cancer effectively – we can predict the likelihood of cancer returning after surgery,' said Dr Yeong, one of the study's authors.
Top stories
Swipe. Select. Stay informed.
Singapore Mindef, SAF units among those dealing with attack on S'pore's critical information infrastructure
Asia How China's growing cyber-hacking capabilities have raised alarm around the world
Asia At least 34 killed as tourist boat capsizes in Vietnam's Halong Bay
Singapore 1 dead, 1 injured after dispute between neighbours at Yishun HDB block
Singapore Vessels from Navy, SCDF and MPA to debut at Marina Bay in NDP maritime display
Asia Autogate glitch at Malaysia's major checkpoints causes chaos for S'porean and foreign travellers
Asia SIA, Scoot, Cathay Pacific cancel flights as typhoon nears Hong Kong
Singapore A deadly cocktail: Easy access, lax attitudes driving Kpod scourge in S'pore
Identifying these spatial patterns allows Times to predict the risk of recurrence with about 82 per cent accuracy.
Existing staging methods – frameworks that evaluate how advanced the cancer is, allowing doctors to recommend the most suitable treatments – vary in their accuracy.
Dr Yeong – who is also director of immunopathology at the SingHealth Duke-NUS Pathology Academic Clinical Programme – noted that up to 70 per cent of liver cancer patients experience recurrence within five years.
An accurate prediction method would allow doctors to more easily identify patients at greater risk of cancer recurrence, he said.
'Times offers a significant advancement in predicting these outcomes, enabling clinicians to intervene at the earliest possible stage. This can significantly enhance patient care and improve survival outcomes.'
He noted that the precision medicine approach seeks to tailor treatments that meet patients' specific needs, based on factors such as genetics and lifestyle.
Even though a large proportion of liver cancer patients experience relapses, it is not viable to offer all of them therapeutics as it would be very costly to them, Dr Yeong said. Getting an accurate prediction of their risks would help optimise the costs to patients, he added.
AI was used to analyse vast amounts of data, with more than 100 trillion data points gathered from liver tumour samples, as well as create a scoring system tailored to Asian patients.
It is also currently being used to refine the Times score for clinical use, Dr Yeong said.
'By automating the analysis of patients' surgical tissue samples, AI ensures the process is efficient and consistent, paving the way for integration into routine clinical workflows, such as through a potential diagnostic test kit,' he added.
The study was the cover story for the April 15, 2025, issue of the peer-reviewed scientific journal Nature.
Ms Denise Goh, the study's co-first author, said the Times scoring system 'transforms routine tissue slides into powerful predictive tools'.
'By identifying patients at higher risk of relapse, we can proactively alter treatment strategies and monitoring, potentially saving more lives,' said the senior research officer at A*Star IMCB.
The researchers validated the accuracy of the Times system using samples from 231 patients across five hospitals in Singapore and China.
The technology is now accessible through a free web portal for research use, with plans under way to integrate Times into routine clinical workflows.
The team is planning further studies at SGH and the National Cancer Centre Singapore to validate the technology, scheduled to begin later in 2025.
Discussions are ongoing with partners to develop Times into a clinically approved diagnostic test kit.
Hashtags

Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles

Straits Times
3 hours ago
- Straits Times
Air quality remains concerning in parts of Peninsular Malaysia
Find out what's new on ST website and app. Malaysia's Department of Environment website reported unhealthy Air Pollutant Index readings in multiple locations as of 11am on July 20. PETALING JAYA — Air quality in several parts of Peninsular Malaysia remains concerning, with the Department of Environment (DOE) website reporting unhealthy Air Pollutant Index (API) readings in multiple locations as of 11am on July 20. Nilai in Negri Sembilan recorded the highest API at 160, placing it firmly in the 'unhealthy' category. This was followed by Banting in Selangor, with an API of 153, also categorised as unhealthy. Putrajaya recorded an API of 111, while Temerloh in Pahang posted a reading of 104. Both areas fall within the 'unhealthy for sensitive groups' category, indicating that individuals with respiratory issues, the elderly, and young children may face adverse effects. In the Klang Valley and surrounding areas, air quality remained moderate. Kuala Lumpur recorded an API of 96, followed by Alor Gajah in Melaka at 94. Top stories Swipe. Select. Stay informed. Singapore 1 in 3 vapes here laced with etomidate; MOH working with MHA to list it as illegal drug: Ong Ye Kung Singapore HSA extends hotline hours, launches new platform to report vaping offences Asia Tearful relatives await news of victims in Vietnam boat capsize Singapore ComfortDelDro to discipline driver who flung relative's wheelchair out of taxi Multimedia How to make the most out of small homes in Singapore Asia Over 380,000 people affected by autogate glitch at JB checkpoint over 2 days Singapore Minor Issues: Why I didn't send my daughters to my brand-name primary school Opinion I thought I was a 'chill' parent. Then came P1 registration Port Dickson and Seremban, both in Negri Sembilan, recorded readings of 92 and 91, respectively. Bukit Rambai and Bandaraya Melaka followed closely with API readings of 89 and 88. Over in East Malaysia, air quality remained moderate as well. In Sabah, Kota Kinabalu recorded an API of 59, while Keningau posted 58. In Sarawak, Miri recorded a reading of 62, Bintulu 61, and Kuching 55. According to the DOE, an API reading of 0–50 is considered 'good', 51–100 is 'moderate', 101–200 is 'unhealthy', 201–300 is 'very unhealthy', and anything above 300 is categorised as 'hazardous'. THE STAR/ASIA NEWS NETWORK


New Paper
7 hours ago
- New Paper
One in three vapes here laced with etomidate: Ong Ye Kung
As an interim measure to tackle the vaping scourge in Singapore, the Health Ministry is working with the Ministry of Home Affairs to list etomidate - a drug used in vapes - under the Misuse of Drugs Act, Health Minister Ong Ye Kung said on July 20. He added that vapes seized by the authorities have shown that one in three vapes contain etomidate, a psychoactive substance that can cause hallucinations and organ damage. Listing etomidate under the Misuse of Drugs Act means those who possess, abuse or sell vapes with etomidate may soon be treated the same way as those found with drugs like cannabis. This will pave the way for harsher penalties for abusers, including mandatory rehabilitation, and jail time for repeat offenders. Etomidate's use is currently regulated under the Poisons Act, which carries lighter penalties for users. Speaking on the sidelines of the grand opening of a dog run in Sembawang on July 20, Mr Ong said the Ministry of Health is working with the Ministry of Home Affairs to do this in the coming weeks. The Health Sciences Authority will also be extending the hours for its reporting hotline. ST launched its anti-vaping campaign, Vaping: The Invisible Crisis, on July 13. Mr Ong praised the initiative, saying it was a useful campaign to raise awareness. When ST launched the campaign, Vaping: The Invisible Crisis, several Telegram channels selling vapes became significantly less active. The campaign aims to raise awareness of the vape scourge. Of particular concern is the rise of etomidate-laced vapes, also known as Kpods. Etomidate is a medicinal ingredient used in clinical practice as an anaesthetic agent and is classified as a poison under the Poisons Act. It was designed to be injected directly into the veins under clinical supervision and was never meant to be inhaled directly into the lungs. When vaped, it can trigger spams, breathing difficulties, seizures and psychosis. Kpods have reportedly been pushed in Singapore by organised crime groups, with local authorities now intensifying crackdowns. Public healthcare institutions have also been told to record all Kpod cases. Possessing, using or buying vapes carries a maximum fine of $2,000. Anyone who distributes, imports or sells vapes and their components can be jailed for up to six months and fined up to $10,000. Those found in possession of or using pods containing etomidate can be jailed for up to two years and fined up to $10,000.


New Paper
7 hours ago
- New Paper
Singapore AI system may help predict liver cancer recurrence
A scoring system powered by artificial intelligence (AI) that was developed by researchers here could help predict the recurrence of a common form of liver cancer. Developed by scientists from A*Star's Institute of Molecular and Cell Biology (IMCB) and Singapore General Hospital (SGH), the system can predict relapses of hepatocellular carcinoma (HCC). Affecting more than 20 out of every 100,000 people here annually, HCC is the fourth-most-common cancer among men in Singapore, as well as the third-most-common cause of cancer deaths among males here. The Tumour Immune Microenvironment Spatial (Times) score analyses the precise spatial distribution, or the exact location, of immune cells called natural killer (NK) cells and five specific genes within liver tumour tissues. NK cells are a type of white blood cell that destroys diseased cells, such as cancer cells, and a lower number of NK cells is an indicator of a higher risk of relapse. "Instead of just counting how many NK cells are present, we analyse exactly where they are positioned relative to the cancer cells," said Dr Joe Yeong, a principal investigator at both A*Star IMCB and SGH's anatomical pathology department. "By understanding how NK cells are distributed and how they interact with cancer cells - for example, whether they are close enough to attack the cancer effectively - we can predict the likelihood of cancer returning after surgery," said Dr Yeong, one of the study's authors. Identifying these spatial patterns allows Times to predict the risk of recurrence with about 82 per cent accuracy. Existing staging methods - frameworks that evaluate how advanced the cancer is, allowing doctors to recommend the most suitable treatments - vary in their accuracy. Dr Yeong - who is also director of immunopathology at the SingHealth Duke-NUS Pathology Academic Clinical Programme - noted that up to 70 per cent of liver cancer patients experience recurrence within five years. An accurate prediction method would allow doctors to more easily identify patients at greater risk of cancer recurrence, he said. "Times offers a significant advancement in predicting these outcomes, enabling clinicians to intervene at the earliest possible stage. This can significantly enhance patient care and improve survival outcomes." He noted that the precision medicine approach seeks to tailor treatments that meet patients' specific needs, based on factors such as genetics and lifestyle. Even though a large proportion of liver cancer patients experience relapses, it is not viable to offer all of them therapeutics as it would be very costly to them, Dr Yeong said. Getting an accurate prediction of their risks would help optimise the costs to patients, he added. AI was used to analyse vast amounts of data, with more than 100 trillion data points gathered from liver tumour samples, as well as create a scoring system tailored to Asian patients. It is also currently being used to refine the Times score for clinical use, Dr Yeong said. "By automating the analysis of patients' surgical tissue samples, AI ensures the process is efficient and consistent, paving the way for integration into routine clinical workflows, such as through a potential diagnostic test kit," he added. The study was the cover story for the April 15, 2025, issue of the peer-reviewed scientific journal Nature. Ms Denise Goh, the study's co-first author, said the Times scoring system "transforms routine tissue slides into powerful predictive tools". "By identifying patients at higher risk of relapse, we can proactively alter treatment strategies and monitoring, potentially saving more lives," said the senior research officer at A*Star IMCB. The researchers validated the accuracy of the Times system using samples from 231 patients across five hospitals in Singapore and China. The technology is now accessible through a free web portal for research use, with plans under way to integrate Times into routine clinical workflows. The team is planning further studies at SGH and the National Cancer Centre Singapore to validate the technology, scheduled to begin later in 2025. Discussions are ongoing with partners to develop Times into a clinically approved diagnostic test kit.