How AI works in the medical field
By XinYi Feng
Imaging that you are lying on a white bed, expecting a reliable doctor who will do the cardiac operation for you, but you did not see your doctor in the operating room. Creepily, what is waiting for you is a machine with four arms. Sounds unrealistic, but this is a real scene in hospital.
Leonardo Da Vinci surgical robot is an A.I. robot invented by Intuitive Surgical, IBM, MIT and Heartport.
With the help of this A.I. robot, doctors can know the patient’s condition from the screen, and do the surgery miles away. The invention of this robot represents that as a new technology with great potential in the future, artificial intelligence started test the water in medicine area.
The field of medicine stands to benefit greatly from A.I. technologies. For example, image recognition technology can be used in detecting a tumor on the skin of human body and a chatting robot may help with the treatment of autism.
Generally, the future application of A.I. technology can be divided into four parts: medical robots, medical imaging, the management of treatments and the invention of medicine .
Currently, many of the medical robots already have been introduced into the hospital, and surprisingly, they did a great job of doing operations. For instance, for the Da Vinci surgical robot, they adapted well in the environment of real operations and have been widely used in many hospitals in America and in other countries.
The reasons of the extensive application of the robot are many: for the patient, the machine can enlarge the view of a wound for ten to fifteen times so that the surgical operations can be more precise, and the time takes to heal the wound will be shortened largely. Also, from the view of doctors, the robot reduces the impact of shaky hands and the tiredness of doctor at the same time. Often times, a surgical robot performs better than a human in doing surgery.
Medical robots also have other applications besides doing operations. The Creative Lab of Columbia University invented a machine which can help with the rehabilitation and treatment of people who has walking problem.
“Our device can help people with stroke to walk normally and help elders to keep balance,” said Sunil Agrawal, professor of Columbia University. “It has not been applied to hospital yet, but I believe it will, in few decades.”
The medical imaging is an important part in diagnosis patient’s disease, so the safety, accuracy and clearness of it are always being focused. Nowadays, CT and MRI are being used in a large scale, but they all have some drawbacks.
CT forms an image by using rays to scan a specific part of the human body. Sometimes it can cause damage to human cells; MRI uses relatively longer time to form image, so for the restless patients, they cannot get a clear image. Also, it cannot “take pictures” for bone and lungs because they lack the atoms needed for MRI imaging.
Not only that, after the images were taken, doctors need to recognize and determine the abnormal parts, and it may can be replaced by A.I. machine. Scientists now do have an intention to replace the doctor in recognizing abnormal parts in images, and some of them already start to examine their product.
Scientists of Verily held a competition of identifying tumors, unexpectedly, the top oncologist failed to their A.I. machine. The accuracy rate of human oncologist is 73.3 percent, but A.I. machine did a more commendable job — — 88.5 percent!
When it comes to choosing the treatment, A.I. technology also shows its strong ability. In the past, considering the best treatment for the patient is a work which needs the doctor to have a profound knowledge and experiences, but right now, A.I. doctors can replace human doctors to do this job.
IBM Watson for oncology is an A.I. system invented by IBM, it can help clinicians by considering individualized cancer treatments for their patients. “Watson for Oncology is a solution that is fueled by information from relevant guidelines, best practices, and medical journals and textbooks,” says IBM company.
This is the most mature technology currently, and it already passed United States medical license examination in 2012, so many hospitals “recruited” it as a assistant for identifying best cancer treatment. From the view of A.I. technology, the system integrates NLP(Natural Language Processing), machine learning, information retrieval and such technologies.
What is more, as an important part of producing new medicine, medicine design is such a complicated procedure which costs scientists so much time to deal with, so the A.I. scientists came up with an idea: Why cannot we use A.I. technology to do the design of medicine? After their unremitting efforts, some of them gained successful research result.
The company of Berg is an example of successfully combined A.I. and medical invention. Berg invented an A.I. platform called InterrogativeBiology, which is able to analyze potential molecules in patient’s body, and try to find a way to make those molecules start to treat the patient’s disease.
Using this method, the cost can be reduced to half in comparison to invent a new drug.
