NEWS
Feature Story
- Publish Date:2024-09-04
AI’s Precision Strike, Leaving Brain Tumor Cells Nowhere to Hide - An Interview with Professor Yu-De Wu of the Institute of Biophotonics
Accurate delineation of brain tumors significantly impacts the treatment of lesions and postoperative outcomes. The detection capabilities of the “AItewan BioMedical Technology Inc.” system offer highly effective support, benefiting patients. (Photo provided by Yu-De Wu)
By NYCU Alumni Voice
Translated by Hsuchuan, Edited by Chance Lai
______
Translated by Hsuchuan, Edited by Chance Lai
______
According to statistics from the Ministry of Health and Welfare, brain tumors, although not among the top ten cancers, can invade surrounding brain tissues regardless of being benign or malignant, leading to symptoms that impact vital functions. The nature of brain tumors is highly variable; some are hard, while others contain cystic components, which can cause blurred boundaries in imaging scans. This makes it challenging for doctors to delineate the tumors, complicating their complete removal accurately.
AI Enhances Brain Tumor Diagnosis, Ensuring Precision Treatment Through Experience and Technology
A groundbreaking AI system is now poised to change the game. A team led by Professor Yu-De Wu from the Institute of Biophotonics at National Yang Ming Chiao Tung University (NYCU) has collaborated with radiology experts Dr. Wan-Yuo Kuo and the Gamma Knife team from Taipei Veterans General Hospital (TVGH). Together, they have developed an AI system using dual-parameter MRI imaging (T1W+C and T2W) to quickly and accurately assist doctors in identifying and outlining major brain tumors, such as acoustic neuromas, meningiomas, and metastatic brain tumors.
The “DeepBT (Deep Brain Tumor) AI-Assisted Brain Tumor Detection System,” which combines AI technology with healthcare, was initially developed as part of a research project led by Yu-De Wu under the “Big Data Research Project for Medical Imaging” from the National Science and Technology Council (NSTC). Following this, Wu applied for two more NSTC research and entrepreneurship projects, gradually leading Yu-De Wu and Wan-Yuo Kuo down the entrepreneurial path, resulting in the founding of “AItewan BioMedical Technology Inc.” in 2023. Wu humorously remarked, “We aim to use artificial ‘intelligence’ to ensure that brain tumor patients ‘receive’ the excellent medical care of Dr. Wan-Yuo Kuo.”
AI Enhances Brain Tumor Diagnosis, Ensuring Precision Treatment Through Experience and Technology
A groundbreaking AI system is now poised to change the game. A team led by Professor Yu-De Wu from the Institute of Biophotonics at National Yang Ming Chiao Tung University (NYCU) has collaborated with radiology experts Dr. Wan-Yuo Kuo and the Gamma Knife team from Taipei Veterans General Hospital (TVGH). Together, they have developed an AI system using dual-parameter MRI imaging (T1W+C and T2W) to quickly and accurately assist doctors in identifying and outlining major brain tumors, such as acoustic neuromas, meningiomas, and metastatic brain tumors.
The “DeepBT (Deep Brain Tumor) AI-Assisted Brain Tumor Detection System,” which combines AI technology with healthcare, was initially developed as part of a research project led by Yu-De Wu under the “Big Data Research Project for Medical Imaging” from the National Science and Technology Council (NSTC). Following this, Wu applied for two more NSTC research and entrepreneurship projects, gradually leading Yu-De Wu and Wan-Yuo Kuo down the entrepreneurial path, resulting in the founding of “AItewan BioMedical Technology Inc.” in 2023. Wu humorously remarked, “We aim to use artificial ‘intelligence’ to ensure that brain tumor patients ‘receive’ the excellent medical care of Dr. Wan-Yuo Kuo.”
Yu-De Wu’s team has received recognition for their innovative medical device and diagnostic technology, the “DeepBT Intelligent Precision Brain Tumor Medical System,” winning the 19th National Innovation Award. (Photo provided by Yu-De Wu)
Professor Wu elaborates on the intricacies of brain tumor delineation, which heavily relies on experience. Accurate mapping of the tumor’s growth is critical to avoid damaging healthy cells or missing parts of the tumor during surgery. At TVGH, pre-surgery protocols involve multi-disciplinary consensus meetings across radiology, neuroradiology, and neurosurgery, requiring the evaluation of numerous MRI scans to ensure precision in manual outlining. This process demands significant time and human resources.
Furthermore, each physician's delineation of tumor boundaries varies based on their experience and skills, which can affect post-surgical outcomes. Yu-De Wu noted that if there are discrepancies among individual doctors, the differences between various hospitals and urban and rural areas could be even more pronounced. To address this, Professor Wu's team developed an AI-assisted system to minimize human variability in clinical practice, making delineation results more consistent. This ensures that all patients receive the same high-quality medical care as if there were a replica of Dr. Wan-Yuo Kuo assisting in every operating room.
To realize this ideal, the team collaborated with TVGH, which boasts 30 years of clinical experience with brain tumors. Over three million imaging data points were used to train the AI algorithms. Wu highlights a critical insight, “The volume of data is crucial, but data quality is equally important.” The team meticulously verified each image’s accuracy and tackled challenges posed by different MRI machines before training the AI model, ensuring the system’s reliability. Without these preparatory steps, the AI model risks becoming a case of “garbage in, garbage out.”
The Limitless Potential of Smart Healthcare: Challenges and Opportunities in Commercialization
The meticulous preparation required for the AI model feels like academic research to Yu-De Wu, a process he’s well accustomed to. However, now that he has transitioned from being a scholar to becoming a company founder, he must focus on commercialization, moving away from his previous habit of being entirely research-focused. The challenge of effectively marketing their excellent product is often the first hurdle that these professor-turned-novice entrepreneurs face. Wu couldn’t help but sigh, “The real difficulties often have nothing to do with the technology itself—being a professor was simpler!”
No matter how outstanding a product is, knowing how to sell it is key. This is especially true for startups in the medical device sector, which require significant capital for long-term investment. Since its inception and development, “AItewan BioMedical Technology Inc.” has accumulated over NT$30 million in R&D costs. The pressure on entrepreneurs, driven by financial demands, is undeniable, especially before obtaining official FDA certification, which is necessary for market sales and applications.
As one of the professors engaged in university-derived startups, Wu often shares his experiences and the associated hardships with fellow professors who share similar aspirations. He frequently emphasizes that one of the key conditions for entrepreneurship is having the personal traits of flexibility and resilience to handle various pressures and setbacks.
It might be about money or age, but Yu-De Wu believes that “this is a once-in-a-lifetime opportunity for entrepreneurship.” If they fail, it will be hard to start over. Therefore, he is determined to go all in, collaborating with key figures in the field to ensure success.
In addition to inviting Dr. Wan-Yuo Kuo, a trusted colleague of over 20 years, to serve as Chairman of the company, Wu also brought in Fu-Ming Wang, a seasoned business strategist, as the Chief Strategy Officer. This move has given the company’s development a significant boost. The company has also set up operations at NYCU’s Innovation Center (currently under the jurisdiction of the Office of Industry-Academia Cocreation), leveraging university resources to unlock the full potential of its original research. By forming strategic alliances with major medical institutions and information systems companies, they are actively extending their reach to ensure the practical application of their innovations.
Additionally, Yu-De Wu, who previously served as a board member of the Spring Foundation of NCTU and as Director of Research and Development at National Yang Ming University, is well aware of the influence the alumni network of NYCU holds in the corporate world. From his perspective, it is somewhat regrettable that, since the merger, many alumni with technological backgrounds still lack a deep understanding of the medical field, even under the broader framework of promoting “smart healthcare.” He hopes that fostering stronger interactions between the university and its alumni will lead to a proliferation of industry-academic achievements, realizing the value of synergy where the whole is greater than the sum of its parts.
As of July 2023, the “DeepBT” system has received TFDA certification, and the application for FDA certification in the United States is underway, with approval expected by mid-2025. Yu-De Wu’s team has no plans to slow down during this waiting period. Building on the foundation laid by their AI-based brain tumor detection system, they are actively developing more neuroscience and technology applications. Wu playfully hints, “There is still so much more the brain can do.” They intend to continue exploring and deepening their work in this field to benefit patients in the future.
Please refer to the original text (Mandarin) for details.
Related Image(s):