[박문일의 생명여행] (34) Artificial Intelligence and Medical
It is no exaggeration to say that the beginning of obstetrics and gynecology treatment begins with ultrasound imaging. When I was an obstetrician-gynecologist in the 1980s, I put a stethoscope on the pregnant woman’s abdomen to hear her fetal heartbeat. Pelvic examination was essential to measure the cervix of pregnant women nearing delivery.
These days, medical doctors can use a simple portable ultrasound device that they carry in their pocket to check not only the heartbeat of the fetus but also the organs of the fetus. An ultrasound device that can measure not only the length of the cervix but also the strength of cervical tissue has been developed, contributing greatly to the diagnosis and prevention of premature birth. By applying this ultrasound diagnostic device to patients with uterine asthenia, which is the main cause of premature birth, I find it very rewarding to be able to effectively treat many high-risk patients who may be at risk of premature birth. Thanks to advances in science and advanced medical artificial intelligence (AI).
Artificial intelligence (AI) is a term used to simulate intelligent behavior and critical thinking comparable to humans using computers and technology. The first person to use the term was Dr. John McCarthy, an emeritus professor at Stanford University, who first described the term AI as ‘the science and engineering of making intelligent machines’ in 1956.
Recently, artificial intelligence has surpassed human intelligence in some ways. No one can beat AI anymore in Go or Chess games. In addition, various types of robots that help our lives are increasing noticeably around us. If you look at the principle of operation of common electronic products around you, it is actually the beginning of artificial intelligence. Delivery robots that deliver food according to customer orders in restaurants these days are also simple artificial intelligence applications that make electronic products in the shape of people. The use of artificial intelligence at work and at home continues to increase. AI in its various forms is being used to develop and advance a wide range of sectors such as banking and financial markets, education, supply chain, manufacturing, retail and e-commerce, and healthcare.
AI technology that exists in business and daily life of modern people is now steadily being applied to modern people’s health care. AI is not only changing the medical device industry, but also in the medical field, supporting medical personnel in various aspects of the diagnosis and treatment process management process for diseases, allowing them to quickly decide on existing diagnostic mechanisms, as well as provide solutions to overcome problems when problems arise. You can easily suggest. For example, in the field of cancer, AI algorithms can diagnose according to a patient’s genes and suggest customized medicines for complex treatment of the cancer.
Artificial intelligence is widely used not only in the field of cancer, but also in other fields of medicine. Typical applications include diagnostics as well as discovery and development of drugs most appropriate for a given patient, improving communication between doctors and patients, copying medical documents such as prescriptions, and remote patient care. In general, computer systems often perform tasks more efficiently than humans, but the recently developed diagnostic field of artificial intelligence is approaching expert level accuracy.
Not only in the field of cancer but also in many medical fields, artificial intelligence has great potential beyond our imagination. In particular, its usefulness in imaging is remarkable. Artificial intelligence has numerous practical applications in the field of imaging. Most of the patient visits to the hospital begin with diagnostic images. Simple chest X-rays, X-rays of bones and muscles, as well as complex imaging data such as CT and MRI play an important role in making important decisions during treatment. This includes, of course, the special ultrasound diagnosis of the obstetrics and gynecology area mentioned at the outset. In analyzing these image data, doctors can now get help from artificial intelligence.
In terms of clinical indicators, image data contains various valuable information. Since AI has an algorithm that can easily identify parts invisible to the human eye, its accuracy can exceed the existing diagnosis rate. If important clues are captured in the imaging data at the initial stage of diagnosis, the diagnosis process can be stopped and treatment can be started immediately, contributing to a reduction in medical costs.
The above facts are the potential benefits of artificial intelligence in medicine. Experts summarize the benefits into four categories: The first is to “exceed the limits of human performance” and the second is to democratize (information sharing) medical knowledge and excellence. The third is to “automate laborious tasks in the medical field” and the fourth is to “manage patients and medical resources efficiently”.
However, while AI offers many of the possible benefits above, we must not forget that it also comes with some risks. The most obvious risk is that AI systems can sometimes go wrong, which can exacerbate a patient’s illness or cause other health care issues. AI systems may recommend the wrong drug to patients, fail to detect tumors in imaging scans, or incorrectly predict which patients will benefit more.
Another problem is data availability. Training AI systems requires large amounts of data from sources such as electronic health records (EMRs), pharmacy records, and consumer-generated information such as the contents of insurance claims records. In other words, the issue of privacy is raised. Another risk may arise with respect to patient privacy.
In particular, there is a risk presented by experts: data bias and inequality. In the United States, for example, African-American patients, on average, receive less pain treatment than white patients. Thus, an AI system that learns from medical records can learn to suggest lower doses of pain relievers to African-American patients. AI systems may also exacerbate inequality by allocating fewer resources to patients who, for a variety of reasons, are considered less desirable or less profitable in the health care system.
Nevertheless, experts now predict that AI will perform essential functions in the medical field. AI predicts that it will change healthcare practices in hitherto unknown ways. Of course, practical applications in the clinical field should be better explored and developed. Therefore, it is true that there are still many medical professionals who are negative about AI.
