Here's the AI diagnosis: How the tech is being leveraged in M'sia's public healthcare system

The Star

7 hours ago

A man in his 60s showed up for a routine health screening at a ­private hospital in Klang Valley last year. Despite being a smoker, he wasn't displaying any ­symptoms that might have ­indicated a health issue.
On that day, he also underwent a chest X-ray augmented by artificial intelligence (AI) software, which was designed to detect subtle abnormalities that the human eye may sometimes overlook.
Prof Anand Sachithanandan, a consultant cardiothoracic ­surgeon, recalls how the software detected a small ­shadow in the upper zone of the man's left lung.
'It was something that would most likely be easily missed or overlooked by conventional X-ray imaging,' he says in a statement to LifestyleTech.
Prof Anand adds that further investigations revealed an elevated tumour marker, which could be a sign of cancer in the body. While this may sound alarming, additional tests are needed to confirm the diagnosis.
'He was swiftly investigated with CT thorax scan, PET scan and biopsy – confirming an early-­stage primary lung cancer,' he says.
Prof Anand adds that they ­performed a surgery to resect (remove) the tumour. Three days later, the man recovered well and went home.
AI has the potential to sift through voluminous data very quickly and will pave the way for precision diagnostics and personalised treatments in the future. — Prof Anand Sachithanandan
It was the first known case of AI helping to detect lung cancer in Malaysia, where Prof Anand says the technology involved cloud-based Qure.AI software powered by deep learning ­algorithms.
'The case demonstrated the potential to screen for lung cancer with a chest X-ray and how the technology can aid radiologists.
'It also highlighted the quick turnaround time which contributed in part to the patient completing definitive treatment in less than two weeks,' he adds.
AI and early detection
Malaysia began incorporating AI screening as early as 2020, when private healthcare centres launched initiatives to use AI for detection of retinal diseases. The adoption of AI is now broadening to public healthcare with the Health Ministry driving initiatives to use AI to detect diseases.
Last year, the Health Ministry announced the roll-out of Dr Mata, an AI-driven software ­solution to detect and diagnose diabetic retinopathy, an eye disease caused by diabetes, in a pilot project involving around 140 ­government clinics.
According to Health Ministry deputy director-­general for research and technical support Datuk Dr Nor Fariza Ngah in a report, the technology can help to produce eye test results at a ­faster rate and would lead to ­better outcomes for patients.
Then in May, the Health Ministry announced that an AI-powered lung cancer screening initiative will be rolled out at seven health clinics nationwide – including in Kelantan, Pahang, and Kedah – starting this year.
'With this AI capability, the incident detection rate is significantly improved,' Bernama ­quoted Health Minister Datuk Seri Dr Dzulkefly Ahmad as ­saying during a National Lung Health Initiative 2025-2030 briefing in May.
Based on the latest available National Cancer Registry Report (2017-2021), Prof Anand says lung cancer remains one of the ­leading cancers in Malaysia, with reported cases on the rise: 'It now makes up 10% of all cancer cases, ranks as the second most common cancer among men, and is the third most common among women.'
He adds that the Health Ministry's initiative to prioritise lung cancer screening is timely, as identifying the disease at an earlier stage offers a significantly better chance of successful treatment.
'As the Health Ministry remains the largest healthcare service provider, adoption of AI-enabling screening in government clinics has the potential to create a significant impact in terms of meaningful stage shift to detect more cases at an early-­stage when the cancer is more amenable to curative-intent ­treatments like surgery,' he adds.
While AI can be scaled and deployed in rural areas to assist clinic staff without a radiologist or doctor on-site, Prof Anand points out that strong communication networks are essential for it to work effectively. — Image by freepik
Prof Yeong Chai Hong from the School of Medicine, Faculty of Health & Medical Sciences, Taylor's University, says AI-driven diagnostics screening at government clinics reflects the country's agenda of moving towards preventive data-driven healthcare.
'The efforts can also be seen as an attempt to demonstrate Malaysia's commitment towards bridging the healthcare gap between urban and rural ­communities through scalable, data-driven solutions.
'It signals a shift from reactive to proactive care, where early detection can lead to timely interventions, improved outcomes, and reduced burden on tertiary care facilities,' Prof Yeong says in a statement to LifestyleTech.
Prof Anand says AI is ubiquitous and its impact is now visible in all sectors of people's personal and professional lives.
'The government has been swift to recognise its potential and been proactive in the early adoption of various innovative AI-driven programmes,' he adds.
He also believes social media has helped to normalise the broad concept of AI that 'it seems less daunting and more acceptable' to most people.
He adds there is also a growing body of scientific evidence pointing towards AI improving diagnostics accuracy and enhancing the delivery of precision and personalised medical care: 'Collectively, this may have contributed to the wider acceptance,' he says.
Dealing with data
Despite its potential, the use of AI in disease detection still comes with some limitations that need to be addressed, says Prof Yeong.
AI should remain a support tool, while clinicians must continue to play the central role in decision-making. — Prof Yeong Chai Hong
'One major concern is algorithmic bias. Many AI models are trained on datasets from specific populations, which may not generalise well to Malaysia's diverse demographic landscape without proper local validation,' she explains.
She also warns of the risk of over-reliance on AI outputs, which could compromise clinical judgment if not carefully managed.
'AI should remain a support tool, while clinicians must continue to play the central role in decision-making. This ensures patient safety and maintains professional accountability,' she says.
Another key challenge lies in data governance. Prof Yeong stresses that the safe, ethical, and transparent use of patient data requires strong regulatory frameworks to protect privacy, ensure security, and build public trust in AI-powered healthcare systems.
To ensure responsible and ­ethical use of AI in the public health sector, she calls for a multi-­faceted approach underpinned by clear regulatory oversight and well-defined standards.
'Authorities such as the Health Ministry should establish national guidelines to govern the validation, certification, and post-­deployment monitoring of AI tools in clinical settings,' she says.
She adds that robust measures are needed to protect data privacy and promote transparency.
'This includes building secure digital infrastructures and ensuring that AI models are developed, trained, and tested with full transparency especially regarding the origin and representativeness of the datasets used.'
She also highlights the need for continuous monitoring.
'AI tools must be regularly assessed and updated to ensure they remain accurate in real-world conditions, are properly calibrated, and do not produce unequal outcomes across ­different patient groups ­especially in today's rapidly evolving healthcare landscape.'
While AI can be scaled and deployed in rural areas to assist clinic staff without a radiologist or doctor on-site, Prof Anand points out that strong communication networks are essential for it to work effectively.
'The cost of installing the software, along with maintaining and upgrading server networks – such as 5G infrastructure – can be a limiting factor,' he adds.
Dzulkefly has said in the Bernama report that the cost of deploying AI lung screening at the selected government clinics is relatively modest – just RM70,000 – and worth the investment for the benefits of early disease detection.
'Cautiously hopeful'
As the founding president of the non-governmental organisation Lung Cancer Network Malaysia (LCNM), Prof Anand has been a key advocate for the use of AI in lung cancer screening. In 2021, LCNM launched a free screening initiative using AI-enhanced chest X-rays at health clinics across the Klang Valley, reaching over 10,000 ­participants.
'I am excited and cautiously hopeful with the more widespread adoption of AI screening,' he shares.
Due to high disease burden and poor outcomes due to most cases being detected at the late stage, he says the time is now ripe for a national level lung screening programme. He adds that other countries like the United Kingdom initiated a national level lung screening ­programme back in 2019, while Australia is about to start their initiative in July.
Ultimately, Prof Anand explains that screening is not a one-off test but part of a longer process that involves follow-up scans, biopsies, and treatment – requiring proper funding and workforce planning.
'There must be a well coordinated pathway for anyone with an abnormal finding on an AI-enabled chest X-ray to be ­followed up and quickly further investigated,' he adds.
For AI tools to be integrated effectively into public healthcare, Prof Yeong says digital infrastructure at hospitals, clinics and screening centres must be strengthened.
'This includes expanding ­electronic health record (EHR) systems, securing medical data storage and improving network connectivity, particularly in rural and underserved areas where digital access may be limited,' she adds.
There is also a need to provide healthcare professionals with ­digital literacy training and ­essential knowledge in data ­science.
She also says there should be a ­structured pathway to expand and scale up successful pilot ­projects into 'sustainable, nationwide AI-powered programmes'.
Beyond health screenings
The potential for AI-driven solutions in Malaysia's public ­sector is vast, according to Prof Yeong. Beyond screening, she sees AI playing a significant role in patient management with tools like AskCPG.
'It can help streamline the implementation of national Clinical Practice Guidelines (CPGs) and optimise healthcare workflows by providing real-time recommendations based on patient data,' she adds.
For hospital operations, she says that AI can also help to ­optimise administrative ­workflows by predicting patient volumes, managing bed ­occupancy and even automating appointment scheduling.
'Beyond clinical care, AI is increasingly used in drug ­discovery and clinical trials. It can optimise drug design ­tailored to specific populations or ­individuals, accelerate the ­identification of new drug candidates, and match patients to appropriate clinical trials based on eligibility criteria,' she adds.
Prof Anand also sees the ­potential for AI to shape the future of health screenings in Malaysia.
'AI has the potential to sift through voluminous data very quickly and will pave the way for precision diagnostics and personalised treatments in the future,' he says, adding that there's the possibility of AI predicting and spotting early warning signs on a chest X-ray or CT scan – even before a tumour fully forms – allowing doctors to closely ­monitor high-risk patients and take action much earlier.
While early intervention can significantly improve a patient's chances of survival, recovery also depends on the individual's own commitment to their health.
Prof Anand says the man in his 60s who was diagnosed with lung cancer – through the AI-assisted screening last year – still requires regular follow-ups for the next three to five years.
'He completed his post-surgery adjuvant chemotherapy (as per international protocol) and thankfully remains well and ­cancer-free. His previously ­elevated tumour marker levels are now normalising. He has also quit smoking!'