SINGAPORE -- By Kim Ji-hye / Korea Biomedical Review correspondent
Hospitals are not short of MRI scanners. But they are short of people who can use them well.
At Singapore General Hospital (SGH), the country’s largest public hospital, staff say demand for imaging is rising, while scans are becoming more complex and time-consuming. That creates pressure on radiographers, the specialists who operate MRI scanners, and on patients who often have to wait or stay still for long periods.
During a Thursday tour of the SGH and Philips MRI Training Centre, staff showed how each scan must be carefully planned before it begins -- deciding which parts of the body to image, how thin each slice should be, and how long each sequence should run.
Radiographers at the SGH-Philips MRI Training Centre demonstrate scan planning using AI-enabled imaging software in Singapore. (KBR photo)
“Patients come in different sizes… so we use these initial scans to plan exactly how much we need to cover,” one SGH radiography staff member told reporters during the visit.
The process is precise. A routine scan can involve multiple sequences, each taking several minutes. Some complex scans can take close to an hour.
SGH and Philips set up a dedicated training center to allow radiographers to practice with simulation tools and software outside the pressure of a live clinical setting.
The center, launched in 2025, is designed to help staff learn advanced MRI techniques and adopt newer tools, including AI that can automate parts of the scanning process.
Training for more complex systems
The new system does not replace radiographers. Instead, it aims to reduce the manual work involved in scanning and make the results more consistent.
Staff demonstrated how AI can help shorten scan times. In some cases, a scan that previously took about 2.5 minutes can now be completed in under 2 minutes, depending on the settings used.
The benefit is not just speed. Shorter scans can make the experience easier for patients, particularly those in pain or those who struggle to remain still.
“The good thing is that it helps patients who cannot keep still for very long,” the staff member said during the tour.
The training center includes several workstations where staff can practice scanning without the pressure of a waiting patient.
“In the actual scan room, there’s always the next patient waiting,” the staff member said. “It’s very difficult for them to learn.”
Why hospitals are focusing on skills
Dr. Liew Tau Ming, senior consultant psychiatrist at Singapore General Hospital, presents the hospital’s PENSIEVE-AI dementia screening tool. (KBR photo)
Dr. Luke Tay Hsien Ts'ung, a senior consultant vascular surgeon and clinical innovation lead at SGH, said the challenge for hospitals is not simply buying AI systems, but having the infrastructure, data, and expertise to use them.
“You need investment in infrastructure for your data… [and] investment in the know-how,” Ts'ung said in an interview with Korea Biomedical Review on Wednesday night on the sidelines of the
Philips APAC Innovation Summit
.
He said model-building also depends on large data sets, meaning large hospitals or hospital networks are better placed to develop useful tools.
“A lot of hospitals in Singapore want to do AI, I’m sure in Korea as well,” Ts'ung said. “But the actual ability to do it well is different.”
Ts'ung said hospitals must also prove that new systems add clinical value. In Singapore, he said, infrastructure such as monitoring systems is generally paid for by the hospital, not reimbursed to patients as a separate item.
“There’s no reimbursement code for a hospital purchase of an infrastructure,” he said.
Earlier detection as part of the same shift
Dr. Luke Tay Hsien Ts'ung, senior consultant vascular surgeon at Singapore General Hospital, speaks during a Philips APAC Innovation Summit session in Singapore. (KBR photo)
The same question -- whether hospitals can turn technology into practical care -- also came up during SGH’s Thursday showcase of PENSIEVE-AI, its tablet-based dementia screening tool.
Dr. Liew Tau Ming, a senior consultant psychiatrist at SGH, said in response to questions from Korea Biomedical Review that many patients with early cognitive decline are still detected too late.
“In Singapore, many cases only come very late… by which time it’s very little we can do,” Liew said during the showcase and media Q&A.
PENSIEVE-AI uses drawing tasks that can be completed in a few minutes and do not rely on language. It is designed to help detect cognitive impairment and dementia earlier, especially among older adults from different language and educational backgrounds.
“What we actually capture is the whole process of how you draw,” Liew said, explaining that the system analyses both the drawing and the way it is produced.
The tool has shown around 93 percent accuracy in identifying cognitive impairment in older adults, according to SGH materials, and is designed to help identify patients who may need further assessment.
The work fits into a broader SGH plan to bring care, training, and research closer together. By 2035, its new 43-hectare campus is expected to become Singapore’s largest academic medical campus, bringing together SGH, five national specialty centers, Duke-NUS Medical School, research facilities, and primary care.
Kim Ji-hye
댓글
0