University of Philippines Manila

Artificial Intelligence Can Help Doctors Spot Brain Tumor in Minutes

Text by: Jennifer Manongdo
Photo by: Joseph Bautista

A STEP INTO THE FUTURE OF HEALTHCARE. Attendees and organizers of the forum pose for posterity. In the photo are (from left): Dr. Alvin Marcelo, Dr. Kathleen Joy Khu, Dr. Jaime Caro, Dr. Kai-Wei-Yu (Dept. of Radiology, Taipei Veterans General Hospital), Dr. Gerardo Legaspi, Dr. Ethan Tu (founder and CEO of Taiwan AI Labs), Chancellor Michael Tee, Dr. Wan Yuo-Guo (Dept. of Radiology, Taipei Veterans General Hospital), Dr. Hsin Mei Wu (Dept. of Radiology, Taipei Veterans General Hospital), Dr. Marlo Nievera, Dr. Iris Isip-Tan, Dr. Veeda Michelle Anlacan, Dr. Dominic Jamora, Dr. Franco Cañal, Dr. Darah Urbano.

The use of artificial intelligence (AI) in medical imaging has shortened the waiting time for detecting brain tumors.

Representatives from Taiwan AI Labs introduced DeepMets Software, a pioneering technology in AI-driven medical imaging at a forum discussing the future of Magnetic resonance imaging (MRI) that uses AI algorithms in detecting tumors even at an early stage. The revolutionary software, approved by the Taiwan Food and Drug Administration in 2019, presents significant advancements in brain tumor detection with its capability to produce results in just five minutes.

“This software was initially intended to help radiologists detect metastasis, especially during Gamma Knife radiosurgery because if you miss one tiny lesion, there are likely consequences…in the case of outpatients, patients normally need to wait at least seven days for the result. But now, we can immediately get the result,” Dr. Hsin Mei Wu, head of the radiology department of Taipei Veterans General explained in the forum held on Dec. 13, 2024, held recently at the University of the Philippines Manila: Museum of a History of Ideas.

Ethan Tu, founder and CEO of Taiwan AI Labs assured the safety and confidentiality of patient information as DeepMets Software works by analyzing uploaded data or images through Federated Learning (FL). This machine-learning approach allows collaborative training between multiple devices or systems while keeping raw data on each device. FL does not send data on a single server. It preserves privacy by sharing model updates and not the raw data.

Philippine General Hospital (PGH) Director Dr. Gerardo Legaspi acknowledged AI’s significant benefits in healthcare. He said PGH officials will make a cost-benefit analysis of using DeepMets Software once it gains approval from the Philippine FDA.

DeepMets Software was jointly developed by Taiwan AI Labs and a team at Taipei Veterans General Hospital (TVGH), and its subsidiary, TAIMedImg. Six tertiary hospitals in Taipei contributed to the initial data collection for training the AI model that analyzes brain tumors, explained Dr. Kai Wei Yu from the Department of Radiology at Taipei Veterans General Hospital. It is now already being used in some medical centers in Vietnam.

The forum was organized as part of the 2024 Manila Economic and Cultural Office (MECO) –   Taipei Economic and Cultural Office in the Philippines (TECO) Joint Research: Health, Agriculture, and Training (HAT) initiative project.#

Ethan Tu previously worked at Microsoft during the early stages of what is now popularly known as ChatGPT. In 2017, he formed Taiwan AI Labs with support from Taiwan’s Ministry of Science and Technology.