How does an artificial intelligence system diagnose human brain diseases?

Doctors will be out of a job, and "artificial intelligence" will be able to diagnose human brain diseases next year.

Soon, "computers" will be able to give precise diagnoses of "human brain" diseases and provide the best treatment options. Yesterday (December 22), the world's first "artificial intelligence research center for neurological diseases" was formally established in Beijing Tiantan Hospital. Wang Congjun, executive vice president of Beijing Tiantan Hospital, revealed that by June next year, in brain tumors, cerebrovascular disease can be replaced by an artificial intelligence system for doctors from the film to clinical diagnosis of prescription, its ability is equivalent to the neurology senior director of the doctor.

Computer to the human brain

Neurological cell damage is irreversible, so early detection, diagnosis, and intervention in neurological disorders is the key to the effectiveness of treatment.

"But there are fewer than 20 doctors in the country who can see neuropathology. With so many patients, doctors can't keep busy if they rely only on remote consultations." Wang Congjun said, "The artificial intelligence system, can replace doctors with machines."

How do AI systems diagnose human brain diseases?

The core of AI application research and development is the leading deep learning technology with the skills and experience of top clinical experts. The reporter learned in the interview that the learning ability of the artificial intelligence system is very powerful, it is through the identification of the patient's clinical symptoms, capture the image color changes and signal strength changes, complete data processing, quickly and expert diagnostic experience to compare ...... through such a massive amount of disease information deep learning, it will continue to improve the disease Through such deep learning of massive disease information, it will continuously improve the efficiency of disease diagnosis, and the diagnostic accuracy rate can reach more than 95%, which is equivalent to the level of a senior chief physician.

Equipment will come with its own reading experts

"In the preliminary PK of brain tumor NMR diagnosis within the hospital, in addition to the director of the radiology department, the AI technology has defeated almost all the doctors in our hospital." Wang Congjun revealed that the biggest problem it solves is that it can quickly popularize the experience of the nation's top experts in neuroimaging and neuropathology, saving a lot of time for doctors to read the films. Previously, patients may need at least 1 day to wait for the NMR report, and later the AI system will be embedded in the NMR equipment, and the patient will be able to produce a report card instantly after the NMR, and the machine will help the doctor to read the film and write the report card. In the future, the AI system will be pre-installed on the factory NMR system, and any hospital that purchases the equipment will be able to use the system. Grassroots hospitals' access to the system is equivalent to the introduction of an expert with leading international technology.

"It needs about 1,000 cases to learn about each disease, and if the disease is very common, one month is usually enough, and if it's not common, it may take five years to save enough." Wang Congjun revealed that the AI system has just finished learning brain tumors, a disease condition, which includes the most common clinical symptoms of six types of brain tumors, the next step, it will soon learn cerebrovascular disease, is expected to be around the spring of this year can be basically finished learning. "In the next step, we also hope that this AI machine can be closely linked with clinical treatment, and become a clinical assisted decision-making system by providing prescriptions and clinical advice to patients."

"Intelligent doctors" to visit clinics in June next year

"To verify the accuracy of AI diagnosis. At the end of June next year, we are going to conduct a 'man-machine war' in Beijing for the diagnosis of neurological diseases, when this AI machine will compete with the world's top neurologists on the same stage." Wang Congjun said rather confidently, "I believe the machine will get the first place at that time."

It is reported that the Artificial Intelligence Research Center for Neurological Diseases has carried out the research and development of head MRI and CT image AI diagnostic products in advance, which is the world's first overall application product of MRI and CT image AI diagnostics for head diseases (covering brain tumors, small-vessel lesions, large-vessel lesions, strokes and so on), and it is expected that it will be formally launched in the Tian Tan International Cerebrovascular Disease Conference to be held in June next year. The product is expected to be officially launched at the Tiantan International Cerebrovascular Disease Conference to be held next June.

As one of the world's three major neurosurgery research centers and China's only National Clinical Medical Research Center for Neurological Diseases, Beijing Tiantan Hospital receives 300,000 patients with difficult neurological conditions from all over the country each year, conducts more than 10,000 surgeries, and owns the country's largest tissue sample bank for encephalopathies as well as a blood sample bank. "More AI application scenarios will be shown in the future, such as artificial intelligence judgment of neuropathological slides, intelligent care for patients with cerebral and neurological diseases, intelligent development of treatment plans based on patients' medical data, and physical and speech rehabilitation training for neurological diseases." Wang Congjun said.

The gospel of brain patients ah!