Monday , January 18 2021

The AI ​​system tracks childhood disease as a doctor

An artificial intelligence (AI) program developed in China, struggling through test results, health records, and even handwritten notes diagnosed childhood diseases as precisely as doctors, researchers said Monday.

From influenza and asthma to life-threatening pneumonia and meningitis, the system consistently matched or performed primary care practitioners, they reported in natural medicine.

Dozens of studies in recent months have detailed how AI is revolutionary detection of diseases, including cancers, genetic diseases and Alzheimer's.

AI-based technology learns and improves in the same way as humans, but has virtually unlimited data processing and storage capabilities.

"I think it will be able to do most of the jobs a doctor does," said senior researcher Kang Zhang, a researcher at the University of California, San Diego, at AFP.

"But I will never replace a doctor," he added comparing the relationship to an autonomous car that is still under the supervision of a human driver.

"It will only allow doctors to do better work in less time and at lower cost."

The new technology, Zhang said, is the first in which AI absorbs unstructured data and "natural language" to mimic the process by which a doctor describes what is wrong with a patient.

"It can mimic a human pediatrician to interpret and integrate all types of medical data – patient complaints, medical history, blood and imaging tests – to make a diagnosis," he said.

The system can easily be transferred to other languages ​​and settings, he added.

By comparing hundreds of information on a single patient with a large store of acquired knowledge, technology ignores links, like past statistical methods – and sometimes meat-and-blood doctors – ignore.

– Just in time –

To educate the proof-of-concept system, Zhang and a team of 70 researchers injected more than 100 million data points from 1.3 million pediatric patient visits to a major reference center in Guangzhou, China.

The AI ​​program diagnosed respiratory infections and sinusitis – a common sinus infection – with 95 percent accuracy.

More surprisingly, Kang said, it did as well with less common diseases: acute asthma (97 percent), bacterial meningitis and varicella (93 percent) and mononucleosis (90 percent).

Such technologies can come in time.

"Various diseases, diagnostic tests and treatment options have increased exponentially in recent years, making the decision-making process for doctors more complicated," says nature in a press release.

Experts who were not involved in the study said the study is further evidence of AI's growing role in medicine.

"Work has the potential to improve healthcare by helping the clinician make quick and accurate diagnoses," said Duc Pham, a professor of engineering at the University of Birmingham.

"The results show that the system performed on average better than younger doctors."

"But it won't replace clinicians," he added.

Machine learning – which generates general rules from specific training examples – "cannot guarantee 100 percent accurate results, no matter how many training examples they use."

AI-based diagnostic tools abound, especially for interpreting machine-generated images such as MRI and CAT scans.

One method unveiled last month in the United States to detect cervical cancer lesions found cancer cells at 91 percent accuracy, as compared to 69 percent for physical examinations performed by doctors and 71 percent for conventional laboratory tests.

Similarly, a cellphone app based on AI technology performed experienced dermatologists to distinguish potentially malignant skin lesions from benign, according to a study in the annals of oncology.

Source link