“Assistant” or “Protagonist”?Medical artificial intelligence repositioning
medical artificial intelligence
OFweekrobotinternet News “Doctors are the core link to revitalize big data. Without professional doctors, all data is worthless, which means that doctors are still the core of treatment, especially in the field of cancer treatment such as tumors. IBM Watson’s development is to give It is impossible for a doctor to be a partner, to provide clinical assistance, and to replace a doctor.” Michael Jordan, a professor at the University of California, Berkeley and an artificial intelligence expert, recently publicly expressed his concerns about the application of AI technology in medical diagnosis at the Tencent “Cloud + Future” summit. Fu Gang, chairman of Baheal Pharmaceutical Group and Baheal Intelligent Technology, made such a judgment.
Michael Jordan said at Tencent’s “Cloud + Future” summit: “Our so-called artificial intelligence looks very intelligent, but it is not. For example, in the medical industry, we let machines do a lot of medical diagnosis, which is unlikely. There are a lot of people who will have dosing issues because of this poor diagnosis, and in some settings, AI’s diagnosis will be affected, and the result of this impact can be the death of the patient.”
IBM’s big bet
In the field of AI health applications, no company in the world is as big a bet as IBM.
IBM’s Watson has been working for the MD Anderson Cancer Centre for quite some time while people are still wondering how artificial intelligence can work. Through the cooperation with doctors in the center, the specialized knowledge and clinical experience are collected in the clinical diagnosis and treatment of rare tumors, and these data are processed and calculated for personalized treatment, patient participation, imaging diagnosis, new drug development, etc. aspect.
For the Chinese market, IBM is also very interested – this is not only because investors are more “generous” in AI artificial intelligence in China’s capital market, for them, China also has the world’s most abundant cancer data and actual need for treatment.
The “2015 China Cancer Statistics” jointly completed by Academician He Jie, deputy director of the National Cancer Center, and Chen Wanqing, director of the National Cancer Registry Center, shows that 4.3 million cases of cancer are diagnosed in China every year, and 2.8 million people die from cancer. Precision decision support, and patients need precision medicine solutions. The average five-year survival rate for cancer patients in China is 37%, far lower than the 67% in the United States and 72% in Japan.
For example, Watson for Genomics (Watson Gene Solutions), which officially entered the Chinese market not long ago, can add more than 10,000 scientific paper data and 100 new clinical data every month on average. The generation of a genomic variant provides annotated files, and cognitive computing can perform workloads at a scale that conventional humans cannot.
“In the field of tumor treatment, there are currently more than 80 targeted drugs for tumors approved by the U.S. Food and Drug Administration. However, there are tens of millions of possibilities between these mutations and drugs, and manual calculations cannot complete such a large workload. , and can not guarantee 100% accuracy. Through cognitive computing, we can analyze all the bulk drugs that have been marketed and seek precision medical solutions based on the patient’s genetic mutation.” Fu Gang said that this month, his work Baheal Intelligent Technology has invested heavily to win the exclusive distribution rights of WatsonforGenomics in China. Three months ago, WatsonforOncology, which Baheal Intelligent Technology exclusively distributed, is expected to cover 150 tertiary general hospitals by the end of the year.
1 2 Next page > page
The core of artificial intelligence is “algorithms”
In fact, in the opinion of most Chinese doctors and experts in the field of AI medical care, artificial intelligence should still play the role of “assistant” for doctors, rather than the protagonist of medical decision-making.
“At present, artificial intelligence is still at a very elementary level in the world, and some simple applications are made in hardware.” Fu Gang further said.
Take IBM Watson as an example. Previously, at the University of North Carolina School of Medicine, doctors prepared 1,000 cases of cancer patients for Watson, hoping to use this to evaluate Watson’s academic performance. As a result, 99% of the treatment options provided by Watson were consistent with those recommended by oncologists. , Not only that, Watson also took the initiative to put forward many auxiliary suggestions, and 30% of these suggestions were unexpected by doctors and related to the latest clinical research.
“Just as modern doctors are inseparable from auxiliary stethoscopes and blood pressure monitors, more and more clinical experts are beginning to realize that AI doctors may exceed doctors with general experience and provide comprehensive medical advice, but this is still the same as stethoscopes and imaging equipment. For doctors, AI is still only a tool for auxiliary diagnosis.” Sun Chen, a doctor in a tertiary hospital in Hefei, revealed to the First Financial reporter, “We use its algorithms and databases for decision support, but those who make decisions should It is still a doctor who can make judgments based on the actual situation.”
In fact, the birth of new things is always full of voices of support and doubt. Doctors who have experience in AI medical use said that because medical artificial intelligence itself is positioned as the role of auxiliary doctors, it sometimes cannot target diseases. Adjust the patient’s real-life situation, “such as the patient’s actual family income, psychological condition, etc. The situation of tumor treatment is very complicated, and it is not the best treatment plan that the patient can accept. In many cases, doctors need to be based on patients. The actual situation to adjust, and comfort and persuade patients, which is currently difficult for medical artificial intelligence.”
Pre: FLIR CMOS Machine Vision Cameras Make... Next: Hard core strength is not limited to ...