Medical AI Moves Beyond Detection to Reporting and Follow-Up Management... "Complementary to Physicians"
Annemiek Snux, Professor at Antwerp University Hospital, Belgium
"Presenting Strategies for Report Generation and Follow-Up Management Is Essential"
"Currently, a significant portion of AI-based imaging interpretation programs are somewhat overestimated. While they are effective at detection, there has never been evidence provided showing that they save patients or reduce medical costs."
Annemiek Snux, Professor of Radiology at Antwerp University Hospital in Belgium, is giving an interview on the 12th at Oakwood Premier Coex in Gangnam-gu, Seoul, during the World Lung Cancer Conference (WCLC) CorelineSoft Conference. CorelineSoft
View original imageAnnemiek Snux, Professor of Radiology at Antwerp University Hospital in Belgium, made this statement during an interview with a reporter on the 12th at Oakwood Premier Coex in Gangnam-gu, Seoul, at the World Lung Cancer Conference (WCLC) CorelineSoft Conference. She voiced a note of caution regarding the method of assessing the value of medical AI solely based on how well it detects lesions (physical changes caused by disease). At the same time, she stressed the need for investment in automating the process of delivering results to patients and linking findings to follow-up exams after interpretation is complete. Professor Snux is an expert in medical AI imaging interpretation and has participated in discussions on the introduction of lung cancer screening in Europe.
Professor Snux stated that AI's lesion detection performance has already reached a sufficiently high level. "AI performance has improved rapidly over the past few years, so there is not much more to gain in terms of increasing accuracy," she said. "The key issues moving forward will not be accuracy, but rather how this technology will be used and who will bear the cost."
She explained that, in terms of application, the convenience of use for doctors needs to be improved further. Tasks such as transcribing examination results into notification documents for patients and contacting those who require retesting are still manually performed. Professor Snux said, "The biggest challenge is achieving seamless integration where AI-analyzed results are transferred directly from the interpretation report into the electronic medical record (EMR)." She also explained that for expansion across multiple hospitals, it should be possible for doctors to complete analysis and reporting with just a few clicks on familiar screens.
She also believes that the entity responsible for the cost burden should be clearly defined. If AI is used to reduce overall healthcare costs, the government should be responsible; if AI is used to reduce hospital staff workload and save time on image interpretation, the hospital should bear the cost. Professor Snux said, "Policymakers do not invest based solely on claims that an algorithm is 'good.' There must be clear evidence of what specifically benefits patients in terms of reduced mortality, disease burden, or costs."
Hot Picks Today
"Two Spend 500,000 Won in a Flash, Yet Wait Two Hours?" ... 'Dreamland' Shunned by Kids and Foreign Tourists
- The Era of Getting 200 Million Won After 30 Years Is Over... A Japanese Bank's "Perpetual Department Head" Who Just Clocked In Cries Out
- Spain Hit by Its Worst Tragedy... Lee Kangin, Moreno and Real Madrid Join Soccer World in Mourning
- They Said, “Make It Your Home for Life” - Then Suddenly Kicked 200 Passengers Off... What Happened to the Around-the-World Cruise?
- "130 Rice Bowls Sold Out in a Day"... 'Unfiltered Face' Adds 8 Million Won to Night Market Sales: China's Bungeunwoo
Professor Snux pointed out that 95-97% of screening tests return normal results. "If the algorithm can reliably identify normal cases, doctors can focus on the small number of cases that actually require clinical judgment," she said. This means that AI and radiologists should be viewed as complementary, not substitutes. She also noted that the direction of medical staff training should change. "It is extremely difficult even for experienced physicians to overturn the decision of a high-performance algorithm," she said. "Going forward, professional expertise will not be about knowing what AI does well, but about understanding where it can make errors."
© The Asia Business Daily. All rights reserved. Unauthorized AI training and use prohibited.