Release date: 2017-08-16
IBM and the New York Genomics Research Center (NYGC) have conducted research on the genomics of dozens of brain cancer patients. The report shows that IBM Watson completed the analysis of brain cancer patients' genomes in only 10 minutes, but human experts need to use 160 Hours to complete the same job. But the treatment plan proposed by IBM Watson is not as precise as the human expert's plan.
This paper analyzes two issues addressed in the study: 1) whether the entire genome is scanned more comprehensively than the information obtained by the "gene panel" test alone; 2) the genomic analysis performed by the IBM Waston and NYGC expert teams is compared.
This article is from spectrum.ieee.org, author Eliza Strickland; compiled by billion euros.
Time is of the essence when treating brain cancer.
IBM and the New York Genomics Research Center (NYGC) have conducted genomics studies on dozens of brain cancer patients. The study showed that IBM Watson completed the analysis of the genome of brain cancer patients in just 10 minutes and proposed a The treatment plan, which demonstrates the potential of artificial intelligence to improve patient care. However, human experts need to spend 160 hours to make similar plans. Although Waston is leading in terms of time, its treatment opinions and effects may not be optimal, so this does not mean that the machine will completely defeat humans.
The study is based on the following case. A 76-year-old man with symptoms of headache and difficulty in movement went to the hospital for examination. The brain scan showed malignant glioblastoma, so the doctor quickly implemented the treatment plan. Both IBM Waston and the doctor analyzed the man's genome and proposed a treatment plan, but unfortunately, the man's condition deteriorated too quickly and soon passed away.
IBM is constantly striving to optimize its cognitive technology platform, Waston, to address multiple medical challenges, including accelerating drug development and improving patient care, characterized by its natural language processing capabilities. Therefore, in genomics, Waston can self-review 23 million journal articles from medical literature and other sources, and conduct digestion understanding. Laxmi Parida, head of the Watson genomics research team, said that most cancer patients do not scan the entire genome (including 3 billion DNA units) when examining genetic mutations, but only perform "gene panel" tests, ie only Examine those subsets of genes that play a role in known cancers. The “gene panel†test is fast compared to genome-wide testing, but the results may not be comprehensive.
The new study, published in the journal Neurogenetics, cites the case of the 76-year-old patient described above to address the following two questions.
First, the researchers wanted to know if scanning the entire genome was more comprehensive than just getting the "gene panel" test. Parida said: "We are trying to answer this question, will the whole genome test get more useful information?" The answer to this question is yes. NYGC clinicians and Waston found that genetic mutations that are not within the “gene panel†test help advise on finding potential drugs and clinical trials.
Second, the researchers wanted to compare the genome analysis done by IBM Waston and the NYGC expert team. Both the Waston and NYGC team of experts obtained the patient's genomic information and identified the mutated genes, and then examined the mutations in the medical literature to see if they were present in other cancers, and then looked for reports of successful treatments and ultimately examined whether the patients were Have the conditions for clinical trials. The team of experts spent 160 hours to complete the series, while Waston was completed in 10 minutes.
Although Waston is the first to propose a solution, its solution may not be the best. NYGC clinicians have discovered two genetic mutations, and the doctors recommend that patients participate in clinical trials. If a patient has the conditions to participate in a clinical trial, he will be included in the experiment as the best chance of survival. However, Waston did not consider this information comprehensively, so Waston did not recommend patients to conduct clinical trials.
Although it is easy to see this research as a competition between humans and artificial intelligence, Robert Darnell, head of the NYGC and principal researcher, does not think so. He said: "NYGC provides oncologists and biologists. Clinical input, but the annotations provided by Waston make the analysis faster."
Robert Darnell said doctors need artificial intelligence tools like Waston to keep up with the rapid growth of medical data. IBM's Parida pointed out that in recent years, the cost of whole-genome sequencing has plummeted, making genome-wide sequencing soon a regular part of cancer treatment. If IBM Watson or a similar artificial intelligence system can quickly access this data, they have the opportunity to provide timely treatment recommendations to save the lives of patients with brain cancer.
Darnell said he hopes IBM Waston will become a regular part of cancer treatment because the data that clinicians need to process is simply too large. He said: "I think it is not a good choice for doctors to put more effort into avalanche-like data. Time is a very important factor for patients, and machine learning and natural language processing tools may provide more. Something of value."
Original search
"IBM Watson Makes a Treatment Plan for Brain-Cancer Patient in 10 Minutes; Doctors Take 160 Hours"
Source: billion euro network
Soft Comfortable Dressing,Breathable Dressing,Silicone Dressing,Custom Hydrocolloid Dressing
Zhende Medical Co.,Ltd , https://www.zdmedicalproducts.com