NextFin News - In late 2025, Chulalongkorn University, based in Bangkok, Thailand, unveiled a pioneering artificial intelligence system called Deep GI, developed collaboratively by its Faculty of Medicine and Faculty of Engineering. Deep GI is trained on hundreds of thousands of endoscopic images to assist physicians in detecting gastrointestinal cancers—colorectal, gastric, and bile duct cancers—with accuracy comparable to expert specialists. The system received Thai FDA approval and is set for pilot deployment across multiple hospitals nationwide, backed by investment support from the Thai Board of Investment.
The impetus for Deep GI arises from Thailand’s critical challenge: colorectal cancer ranks as the third most common cancer among Thais, particularly impacting the over-50 demographic. With approximately 15 million Thais aged above 50 but only about 1,000 trained endoscopists to screen this population, there exists a massive gap in early cancer detection services. Deep GI functions like a "co-pilot" during endoscopic procedures, highlighting abnormal lesions such as polyps in real time, thereby enhancing diagnostic speed and confidence for both seasoned and less-experienced physicians.
Phase 1 of Deep GI, completed in 2022, focused on colorectal cancer detection, achieving up to 97% accuracy. This success catalyzed Phase 2, launched in June 2025, which expanded detection capabilities to include early identification of gastric and bile duct cancers—conditions notorious for subtle and flat lesions that often elude human observers, even specialists. The AI system’s capacity to delineate these difficult lesions represents a world-first milestone in AI-assisted gastrointestinal oncology.
Technologically, Deep GI employs deep learning models trained on localized Thai medical imaging datasets, optimizing precision for the country’s population characteristics. This localized training circumvents the limitations seen in imported AI systems that lose accuracy when deployed in different ethnic and regional contexts. Integrated hardware units connect seamlessly with existing endoscopic equipment, augmenting procedures without additional patient burden or extended examination time.
The anticipated impacts of Deep GI extend beyond diagnostic enhancement. By empowering a broader cohort of general physicians to perform effective screenings, Deep GI addresses the severe human resource bottleneck in gastroenterology. This democratization may substantially increase nationwide screening coverage, vital for early cancer interception and improving survival rates. Moreover, real-time diagnostic aids could reduce unnecessary biopsies and pathology costs, streamlining care pathways and lowering healthcare expenditures.
Strategically, the Thai government and healthcare institutions foresee Deep GI as a cornerstone innovation to mitigate cancer mortality—a leading cause of death where approximately 227 Thais succumb daily. The pilot program rollout in partnership with regional hospitals aims to establish scalable workflows, enabling data accumulation to further train and refine the AI. Future iterations plan to integrate Computer-Aided Diagnosis (CADx) functionality, enabling AI to classify polyps as benign or precancerous, moving from detection to diagnostic interpretation, which would be a critical advance in precision medicine.
From an industry perspective, Deep GI’s transition toward a startup commercialization model underscores a trend where academic innovations evolve rapidly into market-ready healthtech solutions with regulatory clearances. This facilitates faster diffusion across private and public healthcare sectors, fostering AI adoption while potentially catalyzing ancillary developments in medical devices and AI algorithms tailored to Southeast Asia’s clinical needs.
Looking forward, the integration of AI systems like Deep GI in low- and middle-income countries represents a transformative shift in oncology diagnostics. By overcoming human resource constraints and enhancing diagnostic accuracy, these AI tools can significantly reduce late-stage cancer presentations, which are often incurable and cost-intensive. However, successful widespread adoption requires robust training programs for clinicians, continuous AI validation with diverse datasets, and strong healthcare policies promoting technology integration while addressing ethical considerations and data privacy.
In sum, Deep GI exemplifies how cutting-edge AI can be harnessed locally to meet urgent public health challenges—a scalable model for other nations grappling with increasing cancer burdens amid medical workforce limitations. Its ongoing evolution and deployment will be closely watched as a case study in AI-driven cancer screening innovation within emerging economies, with profound implications for improving global health equity and cost-effective care delivery.
Explore more exclusive insights at nextfin.ai.
