AI's New Edge: Thailand's Deep GI Fights Cancer and a Specialist Crisis

BANGKOK, THAILAND – December 01, 2025 – In a significant move that underscores the transformative power of artificial intelligence in healthcare, Chulalongkorn University has unveiled "Deep GI," an AI-powered diagnostic system poised to revolutionize the detection of gastrointestinal (GI) cancers. The innovation, which has already received Thai FDA approval, is not merely a technological leap; it represents a strategic countermeasure to one of Thailand’s most pressing public health challenges: a rising tide of GI cancers coupled with a severe shortage of medical specialists.

Developed through a collaboration between the university's Faculties of Medicine and Engineering, Deep GI functions as an AI "co-pilot" for physicians during endoscopic procedures. As organizations globally grapple with leveraging technology to augment human expertise, this system provides a compelling case study in human-AI collaboration, offering a glimpse into a future where advanced algorithms become indispensable tools in the fight against critical diseases.

The AI Co-Pilot in the Endoscopy Suite

At its core, Deep GI is a sophisticated computer-aided detection (CADe) system designed to enhance, not replace, the diagnostic capabilities of a physician. The system analyzes endoscopic video feeds in real time, highlighting suspicious lesions—primarily polyps—that might otherwise be missed by the human eye. This is particularly crucial for detecting subtle or flat lesions characteristic of early-stage gastric and bile duct cancers, which are notoriously difficult to diagnose.

The system's performance metrics are impressive. In randomized controlled trials involving 1,200 patients across five hospitals, Deep GI demonstrated a remarkable ability to improve diagnostic accuracy. The Adenoma Detection Rate (ADR)—a key quality indicator in colonoscopies—increased by a substantial 17%, rising from 38% to 55%. For more difficult-to-reach proximal adenomas, the ADR saw a 16% jump. These figures are not just statistical improvements; they translate directly into more cancers caught at an earlier, more treatable stage. For colorectal cancer detection, the AI has achieved an accuracy rate of up to 97%, a performance level on par with that of seasoned endoscopy experts.

The technology’s design facilitates seamless integration into existing clinical workflows. Deep GI is external hardware that connects to any standard endoscope, processing the image feed and displaying its analysis on a monitor alongside the primary view. For bile duct cancer, which can manifest as tiny cracks or irregular grooves, the AI generates a "heatmap" overlay, guiding the physician's attention to potential malignancies much like a fighter pilot's radar highlights targets.

Further advancing its capabilities, developers are now working on a computer-aided diagnosis (CADx) feature. This next-generation function aims to not only detect a polyp but also classify it as benign or potentially precancerous. If successful, this could significantly reduce the number of unnecessary biopsies, lowering both patient risk and healthcare costs, and providing another layer of actionable intelligence directly at the point of care.

A Strategic Response to a National Healthcare Challenge

The development of Deep GI is deeply rooted in Thailand's specific healthcare landscape. Colorectal cancer is the nation's third most common cancer, with over 15,000 new cases diagnosed annually. Early detection through screening is the most effective strategy for improving survival rates. However, the country has only about 1,000 endoscopy specialists, a number woefully inadequate to serve a large and aging high-risk population. This gap creates significant backlogs and unequal access to care, particularly in rural areas.

Deep GI is engineered to directly address this specialist shortage by empowering general physicians to perform endoscopies with expert-level support. By standardizing the quality of detection, the system promises to broaden the reach of effective screening programs nationwide. The Thai government, recognizing this potential, is actively supporting its deployment. Through the Board of Investment (BOI), a pilot program is set to deploy 35 Deep GI units in hospitals across the country, with plans for a broader rollout in public hospitals through the Ministry of Public Health.

A critical aspect of the system's efficacy is its training data. Prof. Dr. Rungsun Rerknimitr, a key leader in the project, noted, "AI trained on Thai medical data is more accurate for Thai patients." This focus on localized data sets is a crucial differentiator, ensuring the algorithm is finely tuned to the genetic and physiological nuances of the local population, a factor often overlooked in globally developed AI models. This approach not only enhances accuracy but also builds trust among clinicians who are assured the tool is relevant to their patient base.

Navigating the Competitive and Regulatory Landscape

While groundbreaking, Deep GI enters a competitive and complex global market for AI in medical diagnostics. Established players like Medtronic with its GI Genius™ system have already made significant inroads in computer-aided polyp detection for colonoscopies. Other specialized AI tools, such as Ibex's Galen Gastric for pathology and various systems for detecting gastric or bile duct cancers, are also emerging from research institutions and startups worldwide.

However, Deep GI has carved out a unique position with its claim as the world's first single AI system designed to detect cancers across all three major GI organs: the colorectum, stomach, and bile duct. This multi-organ capability, launched in its Phase 2 iteration in June 2025, represents a significant leap in consolidating diagnostic AI. Instead of requiring multiple specialized systems, a single hospital could potentially leverage one platform for a broader range of GI screenings, offering significant efficiency and cost advantages.

Widespread adoption, however, is not without its hurdles. Like many nations, Thailand is grappling with challenges related to AI integration in healthcare. These include fragmented healthcare data systems, a need for clearer data-sharing guidelines under the country's Personal Data Protection Act (PDPA), and the financial investment required for new technology. Furthermore, ensuring that medical professionals are adequately trained and willing to trust and incorporate AI into their established workflows remains a universal challenge. The success of the 35-unit pilot program will be a critical test of the system's real-world usability and its ability to overcome these barriers.

From University Lab to a Regional Beacon

Chulalongkorn University's strategy for Deep GI extends beyond domestic public health improvement; it serves as a powerful model for converting academic research into a commercially viable enterprise with regional and global potential. With Thai FDA approval secured, the university is moving to spin the project off into a startup. The commercialization plan targets both a Business-to-Government (B2G) model, partnering with the Ministry of Public Health for national deployment, and a Business-to-Business (B2B) model for private hospitals.

This transition from lab to market, supported by government bodies like the BOI and industry partners, could become a blueprint for other research institutions in Southeast Asia. It demonstrates a pathway for developing cutting-edge technology that addresses local needs while also possessing the potential for international scalability. The success of the Deep GI startup could catalyze further investment and innovation in the region's burgeoning medical AI ecosystem.

By building a world-class AI on localized data, solving a critical domestic problem, and charting a clear path to commercialization, Chulalongkorn University is not just creating a life-saving tool. It is strategically positioning Thailand as a key player in the future of AI-driven healthcare, proving that impactful innovation can emerge from anywhere to address universal challenges. As Prof. Dr. Rungsun stated, the hope is that Deep GI will serve as a powerful tool for "preventing severe illness, improving quality of life, and reducing cancer-related deaths/cost in Thailand," a mission with resonance far beyond its borders.