Advancements in medical AI usage in Japan have previously been staggered by significant regulatory challenges, those posed by years-long quality assurance, the difficulties of achieving proper certification, and the gauntlet of pharmaceutical approval. Recent ordeals such as the pandemic-era influx of patients, as well as the implications of a rapidly ageing population and its burden on the existing, human-driven medical system, have demanded the arrival of AI tools to support healthcare across the country.

It's no surprise that clinical and hospital spaces benefit from—in fact, require—the fastestresponse times, quickest thinking, and most consistent analyses. Since the first Japanese AI-based Software as a Medical Device (SaMD), called EndoBRAIN, made to recognise colorectal lesions, was approved for use in 2018, Japan’s many regulatory conventions have quickly changed, leading to faster implementation of AI SaMD’s such as Nodoca, an influenza detection system that can correctly diagnose patients in seconds. Other AI SaMD’s based on cardiac arrest prediction already outperform traditional practices, and AI Medical Services (AIM)’s endoscopic cancer detection SaMD boasts a 94% accuracy rate at two-hundred times the speed of the average physician.

These regulatory changes, though exciting, are just the beginning. In Japan in 2021, the Pharmaceuticals and Medical Devices Agency (PMDA), which works directly with Japan’s Ministry of Health, Labour, and Welfare, established its official SaMD review office and enacted its “DASH for SaMD” program, an enhanced review system meant to facilitate even earlier realisation and adoption of devices, including AI, whether for simple monitoring or full diagnosis.

7The PMDA’s 2021 statement on these changes, itself a promise of “regulatory agility”, was a promising new beginning, and judging by Japan’s rapidly evolving, razor-sharp diagnostic tech and expanding number of medical AI start-up companies since, the AI SaMD revolution’s own diagnosis is looking to be a healthy one.