Vietnam can warn of early liver cancer through AI technology

Hanoi University of Technology is studying a system to support the diagnosis of liver cancer based on ultrasound imaging and gene interpretation data.

On the afternoon of August 15, at the Vietnam Artificial Intelligence Festival (AI), Assoc. Prof. Vu Duy Hai, director of the Center for Biomedical Electronics (Hanoi University of Technology), said this unit is studying the system. Support diagnosis of primary liver cancer (cancer derived from liver) based on big data on ultrasound imaging and gene interpretation.

Research on the application of AI technology to diagnose liver cancer has been carried out since 2018, based on ultrasound imaging data of 10,000 primary liver cancer patients. These images will be collected from central and local hospitals of all ecological regions throughout the country. Of these, 200 patient records will be used to sequence genes to improve the accuracy of diagnosis.

Medical records from hospitals after being transferred to the research center will be standardized and labeled before storage. To date, the group has collected more than 6,000 ultrasound images.

Assoc. Prof. Vu Duy Hai.(Photo: VT).

After collecting enough images, Hanoi Polytechnic University will develop a tool to diagnose primary liver cancer early that can run on many devices such as computers and mobile phones. Application can integrate into the ultrasound system of many units from local health facilities, district, provincial, central hospitals and clinics.

"The application is integrated in the software that can diagnose and screen primary liver cancer early with an accuracy of about 90%," Hai said.

Liver cancer is one of the diseases with a high incidence and mortality rate in Vietnam. Primary liver cancer accounts for 40% of liver cancer patients. If successful research will bring opportunities for many patients.

However, Dr. Dao Viet Hang (Hanoi Medical University Hospital) wondered, ultrasound is not a method to diagnose liver cancer accurately. Because want to determine liver cancer biopsy or tomography, magnetic resonance."Ultrasound images to detect tumors can not confirm that it is liver cancer. Therefore, this application should only be used to screen for warning of early liver cancer , " Dr Hang said, saying that there should be a conclusion. The combination of health professionals and AI is feasible.

As a unit that is deploying the application of AI to diagnose lung disease through radiographic images, Dr. Tran Quy Ha (Big Data Research Institute, Vingroup) analyzes, currently, the same image but doctors can draw different conclusions about the condition. Even the same picture was concluded by a doctor but several months later there was a difference in the diagnosis.

Therefore, in the future, the combination of doctor council and AI application will provide the most accurate diagnosis. However, he said, the application of AI in medical imaging diagnostics must include the process instead of applying in each small component.

Dr. Tran Quy Ha.(Photo: VT).

"Another challenge is that the images collected must be annotated by the doctor's council, labeled. Every picture must be diagnosed by a doctor, what is the cause, how to treat it , " he said. Ha said.

The big data research institute is developing medical diagnostic imaging databases to help physicians have more basis to make conclusions about the condition. At the same time, Vingroup also aims to decode the genes of 1,000 healthy people, finding genetic variations to build the genetic database of Vietnamese people. When these databases are completed, they will be shared freely for hospitals and communities to serve diagnosis and treatment.

Dr. Tran Quoc Long (Hanoi National University) said that this unit is also collecting data to build AI applications in explanation of echocardiography images.

Currently, to determine the heart's image, the doctor must place 15 probes around it, thereby judging the condition."Although the same echocardiography images, doctors assess the disease status is a big difference , " Mr. Long said.

According to Mr. Long, in the world, there are many countries applying AI in diagnosing echocardiography images, but the cost of Vietnam is too expensive, about 3-5 billion. Therefore, Hanoi National University is collecting echocardiography images from hospitals to build AI applications in analysis and explanation. Application that can measure heart condition, partition of ventricular structure, atrial, warning myocardial infarction .

However, this is facing many difficulties, especially image labeling. Because each medical record has about 14 videos, each video has 60 - 90 images. The doctor must study all the images to label and store in the database.

Sharing difficulties with the above projects, PGS Tran Quy Tuong (Director of Information Technology, Ministry of Health) affirmed that this unit is making great efforts to build a smart medical system with many legal documents. breaking the law.

But he acknowledged, awareness from hospital leaders is a major barrier to the application of information technology in medical examination and treatment."For AI applications in hospitals requires managers to be fierce, accept publicity and transparency," Tuong said.

The artificial intelligence festival program (AI4VN) takes place at Hanoi University of Science and Technology, 1 Dai Co Viet, Hai Ba Trung District, Hanoi from August 15-16. The event was organized by the Ministry of Science and Technology, Ministry of Planning and Investment, Bach Khoa University and VnExpress e-newspaper.

Vietnam Artificial Intelligence Festival is accompanied by FPT, Viettel, VietinBank, VNPT, VIB, VinGroup, VNPost, BKAV. In addition, the program also received coordination from Five9, Misa, SmartOCR, Dinogo, FastGo, Topica, Kambria, VietAI, BKHoldings, Netnam.