News

Virtual doctors: Artificial intelligence detects numerous eye diseases


Artificial intelligence should be able to recognize over 50 eye diseases

Eye diseases are one of the main causes of vision loss; many of them could be effectively treated through early detection and therapy. A system based on deep learning could also help with diagnosis in the future. Artificial intelligence is said to be able to recognize over 50 eye diseases.

Artificial intelligence in medicine

Artificial intelligence (AI) has been a permanent fixture in healthcare for many years, including helping to examine blood, urine and other body fluids. There has been progress in this area in recent years. For example, US researchers developed an AI system that can detect skin cancer as reliably as medical professionals. And Italian scientists reported on software that Alzheimer's should recognize ten years before the onset of the disease. Even the approximate time of death of patients is said to be predictable by AI, according to Australian experts. An AI system has also been developed in the UK that could be used in healthcare: it should be able to detect over 50 eye diseases.

Early diagnosis and treatment can sometimes get vision

According to a statement from University College London (UCL), more than 285 million people worldwide live with some type of vision loss.

Eye disorders are a major cause of vision loss, and many can be prevented through early detection and treatment.

Researchers at UCL, the British eye clinic Moorfields Eye Hospital and Alphabet subsidiary DeepMind have developed a system based on deep learning that could help diagnose such diseases in the future.

According to a study published in the journal Nature Medicine, DeepMind-AI has learned to recognize more than 50 common eye diseases using thousands of depersonalized OCT (Optical Coherence Tomography) scans.

Detection rate as by a specialist

According to a message from Moorfields Eye Hospital, the detection rate is over 94 percent, which is in the same range as an evaluation by a specialist.

According to the scientists, there is hope that the technology can make work much easier.

"The number of eye scans we do is growing much faster than human experts can interpret them," said Dr. Pearse Keane from Moorfields Eye Hospital.

"There is a risk that this can lead to delays in the diagnosis and treatment of diseases that are dangerous to the eyes, which can be devastating for the patients," said the ophthalmologist.

"The AI ​​technology we are developing aims to prioritize patients who need to be seen and treated by a doctor or ophthalmologist," said Dr. Keane.

"If we can diagnose and treat eye diseases at an early stage, we have the best chance of preserving people's eyesight."

Software is supposed to give treatment recommendations

Clinical tests are now to be carried out to examine how this technology can improve patient care in practice.

Although the AI ​​should not be able to decide which treatment a patient will receive in the foreseeable future, it should be used to determine which patients need help most.

Afterwards, real doctors decide on the respective case.

The researchers expressed hope that the AI ​​system could be used in British clinics in five years.

"I am convinced that AI will play an important role in the future of healthcare, especially when it comes to training and supporting medical professionals so that patients can benefit from vital treatment sooner than before," said Professor Sir Peng Tee Khaw from Moorfields Eye Hospital in London.

However, not everyone is convinced that AI systems should play such a major role in the medical field. Among other things, not because it remains unclear how they come to the respective result. (ad)

Author and source information

Video: 2018 WMIF. The Future of AI in Pathology and Laboratory Medicine (August 2020).