Originally published on DIGITIMES / Report by Lin Pei-Ying

According to the latest Cancer Registry Report released by the Ministry of Health and Welfare in November 2023, lung cancer has surpassed colorectal cancer to become the most prevalent cancer in Taiwan. Beyond Taiwan, lung cancer remains one of the leading causes of death worldwide. One of the primary reasons for its high fatality rate is late detection—most cases are diagnosed at advanced stages, missing the optimal window for treatment. This underscores the critical importance of early screening.To address this, DeepRad.AI has developed the DeepLung-CAC multimodal imaging health platform, which integrates multi-modal AI technology to analyze low-dose computed tomography (LDCT) scans. This technology not only detects lung nodules but also predicts coronary artery calcification levels, supporting the advancement of preventive medicine.

DeepRad.AI, a spin-off startup from Taipei Medical University, was founded by Dr. Chen Chen-Yu, former Vice President of the university and an expert in artificial intelligence research. The company focuses on developing AI-driven technologies for early disease detection and prevention, aiming to reduce mortality rates and optimize the use of healthcare resources.  According to DeepRad.AI CEO, Dr. Chang Yao-Chi, the earlier lung cancer is detected, the higher the survival rate. However, diagnosing lung lesions is challenging. Traditionally, X-ray screenings often fail to detect lung cancer until stages 3 or 4, when the survival rate drops to just 10–30%. In contrast, low-dose computed tomography (LDCT) can identify lung cancer as early as stage 1 or even stage 0, significantly increasing survival rates to over 90%, demonstrating the clear benefits of early screening.  

Currently, LDCT lung cancer screenings require interpretation by radiologists. However, as CT scanning technology advances, modern 640-slice CT scanners generate vast amounts of imaging data, leading to a dramatic increase in radiologists’ workload and impacting efficiency. To address this challenge, DeepRad.AI developed the DeepLung-CAC multimodal imaging health platform, which utilizes AI technology and optimized screening report workflows to enhance efficiency and accuracy.

Dr. Chang Yao-Chi explains that DeepLung-CAC can generate preliminary screening results within just five minutes after an LDCT scan, while also comparing new and previous scans to track nodule changes. This assists physicians in explaining results to patients during consultations. Additionally, DeepLung-CAC automatically generates lung screening reports compliant with Taiwan’s National Health Administration standards, significantly reducing the time required for radiologists to analyze scans and prepare reports. This not only enhances diagnostic accuracy but also creates a more efficient work environment for medical professionals.

Beyond lung cancer screening, the DeepLung-CAC multimodal imaging health platform also assesses cardiovascular calcification risk. Since cardiovascular calcification is a key indicator of cardiovascular disease risk, DeepRad.AI trained DeepLung-CAC using paired datasets—data from simultaneous lung cancer screenings and cardiovascular calcification assessments. This enables the platform to detect both early-stage lung cancer and cardiovascular calcification risk from a single low-dose CT (LDCT) scan. Over four years of research and analysis, DeepLung-CAC has been developed into a comprehensive screening tool for the top two leading causes of death in Taiwan. 

The platform is currently undergoing clinical trials in preparation for TFDA (Taiwan Food and Drug Administration) approval, with certification expected in 2024. Clinical trials are already underway at Taipei Medical University Hospital and Shuang Ho Hospital, demonstrating its ability to significantly improve CT image interpretation efficiency. This makes DeepLung-CAC a highly effective AI-assisted diagnostic tool for lung cancer screening and general health checkups, helping to optimize healthcare service quality.

Regarding market potential, Taiwan has an estimated 500,000 high-risk individuals for lung cancer, yet in 2022, only 50,000 underwent government-funded screenings. If preventive awareness improves, the total screening volume could increase tenfold. However, Taiwan currently has only 1,600 radiologists, highlighting the urgent need for AI solutions to reduce workload and enhance diagnostic accuracy.  DeepRad.AI’s strategy is to leverage AI to assist radiologists, streamline the entire screening workflow, and ultimately improve lung cancer detection rates. By enhancing screening efficiency and boosting early-stage lung cancer diagnosis, the company aims to achieve its vision of advancing lung cancer prevention and treatment.

DeepRad.AI after achieving significant advancements in lung cancer and cardiovascular calcification detection, DeepRad.AI is now focusing on developing AI-powered solutions for osteoporosis and dementia prediction. The goal is to introduce breakthrough innovations in preventive healthcare and diagnostics.

According to DeepRad.AI CEO Yao-Chi Chang, the company’s success is built on two key pillars: the dedication of its team and Taiwan’s supportive startup ecosystem. In addition to cutting-edge R&D, precise and effective market strategies are essential for product success. With government support and policy initiatives, DeepRad.AI aims to attract international investors, successfully commercialize its AI technologies, and expand its global impact. Looking ahead, DeepRad.AI hopes to see further regulatory improvements and policy support in Taiwan. By leveraging the country’s strengths in technology and healthcare, the company aspires to enhance Taiwan’s competitive edge in smart medical solutions and drive innovation in AI-powered healthcare.

Yao-Chi Chang, CEO, DeepRad.AI