:::

Artificial intelligence-assisted atherosclerotic lesion localization and grading decision support system - National Science and Technology Council


Published:2024-10-30  08:53

【Technical introduction】
Literature indicates that activated macrophages that express the chemokine C-X-C receptor type 4 (CXCR4) are gathered in large quantities at the site of inflammation lesions. In order to develop small molecule CXCR4 nuclear medicine diagnostic drugs, NARI has designed a new atherosclerotic radiopharmaceutical APD based on the CXCR4 antagonist TIQ-15 through artificial intelligence computer simulation technology. (Figure 1), the chemical structure has been patented in the United States (2024), Japan (2023) and Taiwan (2022). This research had won the Excellence Award in the oral paper at the 2021 Taiwan Nuclear Medicine Annual Meeting and won the 2023 "Top Ten Highlight Technologies" in National New Innovation Award and have been invited to give a speech at large-scale domestic conferences. The results of the pre-toxicity test found that APD did not show any toxicity even at a dose that was 1,500 times that of the human body.


▲Figure 1. Based on the structure of TIQ-15, a new atherosclerotic radiopharmaceutical APD had designed through computer simulation technology.

【Project Planning/Technology Application】
 
Using atherosclerotic model (ApoE-/-) mice for PET/CT imaging, Ga-68-APD can accumulate in atherosclerotic lesions within 1 hour, and is rapidly excreted out through kidneys and bladder. The target/background ratio (TBR) > 10, which is better than the currently used radiopharmaceutical Ga-68-Pentixafor in human clinical trials internationally. Comparing with the clinical radiopharmaceutical F-18-FDG, it produces large accumulation in the myocardium. F-18-NaF is currently only used in clinical trials for angiographic diagnosis of calcification in intermediate and advanced lesions, so it cannot use for non-calcified patients with vulnerable plaques.
 
NARI has successfully applied Ga-68-APD to evaluate the effects of diabetes medication (sGLT2i) and bromelain, non-invasively. Subsequently, NARI has also used Ga-68-APD to evaluate the effects of common cardiovascular-related health foods (Nattokinase, Red yeast rice, Lumbrokinase). After cross-comparison of the blood lipid data (cholesterol, triglycerides, etc.), it was found that various health foods are effective in improving blood lipids and vascular plaques. If they are mixed consume, the effects will show different additive or inhibitory results. This technology platform will be an effective tool for the rapid screening of compound health foods or clinical treatment drugs in the future.
 
This study also collaborated with the Institute of Engineering, industry, academia, and hospital to develop an artificial intelligence-assisted atherosclerosis lesion location and classification decision support system (ADSS), uses the Taiwanese open-source WEB medical image browser system (Bluelight) with clinical experience in domestic medical institutions to develop the system's core program. In collaboration with a cardiology medical team, the atherosclerosis image data of different atherosclerosis programs used to mark the lesion area and at the same time define the atherosclerotic grading index to provide a reference sample for artificial intelligence learning. After AI training, an accurate and effective model will be established and built into the system to provide AI identification of atherosclerotic lesions in raw images. The identification results will be able to mark the atherosclerotic area, 3D positioning, and expert-recommended indicators for the degree of atherosclerosis. This system will be designed with relevant modules that can be linked to the hospital system, provide report printing functions, and print reports with relevant chart information after AI recognition in combination with patient data, providing medical teams with evaluation. This system uses the latest global commercial 3D MR technology, and uses a head-mounted display to display the AI ​​training recognition results and the 3D image display of the location of arteriosclerotic lesions through a 3D MR display platform, providing the medical team with preoperative communication, intraoperative assistance, and postoperative tracking. (Figure 2) 


▲Figure 2. Client VIEWER 3D MR simulation screen

【Future layout】
Cardiovascular disease is the leading cause of death worldwide. Atherosclerotic vulnerable plaques formed by fat accumulation can cause narrowing or blockage of arteries, which can easily lead to coronary artery disease, stroke and other diseases. Currently, non-invasive CT, MRI, ultrasound and commonly used nuclear medicine drugs (such as Tl-201, etc.) are only suitable for angiography of mid- and late-stage atherosclerosis, and are not suitable for the detection of systemic vulnerable plaques.
Ga-68-APD is currently undergoing GLP toxicology testing and has begun to conduct CMC documents, PIC/S GMP process, and analytical validation technology. In combination with the absorption and metabolism of pharmacokinetics, the relevant eCTD documents required for clinical trial application. It is estimated that IRB application and subsequent Phase 0~1 human clinical trials will be carried out by the end of 2026. Considering the convenience of international transportation, the arteriosclerotic radiopharmaceutical APD will develop to sterilize liquid kit. It can label with Ga-68 in hospitals around the world, without restricted by the half-life of radioactive isotopes.
In the future, NARI will establish a knowledge database of atherosclerosis symptom classification exclusively for Chinese people, which will apply to longitudinal research and prevention related to atherosclerosis, and achieve the effect of intelligent diagnosis and prevention warning. NARI will also complete the product registration and technology transfer procedures for the artificial intelligence-assisted atherosclerosis lesion location and classification decision support system (ADSS), and complete the TFDA artificial intelligence medical equipment inspection and registration application. It is expected that through introduction of the AI-assisted decision support system will improve the technical level and service quality of medical institutions, increase market competitiveness, attract outstanding medical talents and improve the level of medical teams. Moreover, by applying this system to the diagnosis of various types of cardiovascular diseases, expanding the service scope of medical institutions, more business opportunities can be created.
According to the data of the World Health Organization (WHO), the number of deaths from cardiovascular diseases will reach 23.6 million in 2030. With the aging population, changes in lifestyle and the improvement of cardiovascular disease awareness, the global market for cardiovascular imaging is growing rapidly at 6% per year. Ga-68-APD can use to diagnose mild to severe systemic vulnerable plaques that cause vascular obstruction, and the radiation dose received by the human body is only 2.8% of that of Tl-2010. This chemical structure has been patented and will promote mainly in Europe and the United States, where there are more markets. The scope of application will gradually expand to create a more efficient business model, and early detection, early treatment and efficacy tracking will ultimately achieve the substantive goal of improving the health and well-being of people around the world.
 
【Contact information】
Name: Chien-Chung Hsia. Researcher
Tel: 03-4711400 ext 7159
E-mail: hsiacc@nari.org.tw