Medical software startup develops AI to support Japanese surgeons
As Japan battles a dwindling workforce and rising cancer cases, a tech firm designs software to guide doctors through complex operations.
- A Tokyo-based startup, Direava, is developing an AI system with government support to address Japan's projected shortage of surgeons.
- The software uses generative AI to analyze images of organs and blood vessels, creating written, step-by-step surgical instructions that highlight areas of caution.
- Initially launching as an instructional tool, the AI was tested in a trial where it demonstrated an 85-90% accuracy rate when reviewed by medical experts.
- Widespread clinical adoption of the technology faces significant regulatory hurdles, as specific safety and evaluation standards for this type of AI do not yet exist in Japan.
By 2043, the headcount of Japanese gastroenterological surgeons aged 65 or under is projected to plummet by 50 per cent from the roughly 16,000 recorded in 2023, according to data from the Japanese Society of Gastroenterological Surgery.
This looming deficit stems from a gruelling professional environment that deters recruits, even as the country's ageing population drives cancer cases toward an expected peak around 2040.
Such a scarcity of younger clinicians threatens to disrupt the traditional mentorship model, where surgical techniques are traditionally passed down through the close observation of seasoned specialists.
In response to this impending healthcare emergency, the Japanese government is actively funding private enterprises to innovate artificial intelligence systems capable of alleviating the intense pressures on medical staff.
Because pioneering these computational platforms demands massive capital injections and high-performance computing resources, the nation's industry ministry has provided essential backing to a Tokyo-based startup named Direava.
The firm is building an AI solution aimed squarely at tackling these systemic hurdles.
Direava's proprietary software leverages generative artificial intelligence to evaluate digital imagery of internal organs and blood vessels.
It then translates this visual data into comprehensive, written, step-by-step operational workflows that highlight specific areas demanding extreme caution.
To train the algorithm, developers are feeding it clinical data harvested from operations performed by highly proficient surgeons both domestically and globally.
The firm intends to launch the platform initially as an instructional resource, targeting a rollout as early as this year.
A practical trial conducted this February saw medical students utilise the system to pose real-time queries while watching a gastric cancer operation.
When medical experts subsequently scrutinised the AI's feedback, they verified an accuracy rate hovering between 85 and 90 per cent.
To refine this performance, the startup plans to process a vastly expanded pool of surgical data.
"We hope to turn AI into an assistant who supports surgeons by reducing the stress of surgeries and patients' complications as much as possible," stated Masashi Takeuchi, a practising surgeon who serves as Direava's founder and chief executive.
Deploying automated intelligence within life-critical medical environments remains bound by strict regulatory constraints.
While Japan's health ministry is currently engineering an evaluation framework for AI-driven medical apparatus, official assessment benchmarks specifically for intraoperative guidance software do not yet exist due to a lack of historical precedents.
Consequently, significant roadblocks persist before widespread clinical adoption can occur, particularly regarding rigorous safety validation and the creation of enabling legislation to foster the technology.
[Copyright The Jiji Press, Ltd.]
