Imagine a future where advanced technology can improve healthcare and medicine by providing faster, easier and more accurate diagnoses [6]. The field of artificial intelligence (AI), which is currently being led by some of the biggest technology companies and startups, enables machines and robots to be intelligent and perform complex tasks, whereas humans and other animals naturally demonstrate intelligence. Medical facilities are constantly looking for innovative approaches to increase treatment efficiency and to improve their quality of care and outcomes. AI, combined with machine learning and predictive analytics, can help speed drug discovery time, provide virtual assistance to patients, and diagnose ailments by processing medical images [3]. The market for AI will become especially popular when healthcare switches to a value-based reimbursement style [6], if the values of the Affordable Care Act can be salvaged from the current political climate, where providers are paid based on the quality of care they provide rather than the time spent or number of patients seen.
So far, AI has produced some promising inventions to aid in the treatment and detection of chronic disease. For example, a tool called Cognitive Cloud, developed by CognitiveScale – a company based in Austin, TX, can help detect chronic disease in patients [6]. There is also potential for AI to be able to look at medical records and predict outcomes a few months into the future, providing an opportunity for better preventative strategies [3]. For example, New York University is developing a software that can accurately predict the onset of diseases such as Type 2 diabetes, heart or kidney failure and stroke [3]. Furthermore, Boston-based startup FDNA uses facial recognition technology to match with a database associated with over 8,000 rare diseases and genetic disorders, and shares the data to medical centers in 129 countries via the Face2Gene application [3]. Lab-based research in California research havehas been able to detect cardiac arrhythmia with 97% accuracy on Apple Watch users with the AI-based Cariogram application, allowing for early treatment options and helping to avoid strokes [3]. However, some possible challenges to the use of AI in the medical field exist around the integration of medical records with data from smart watch technologies, such as Apple Watch or FitBit, largely because of difficulties in access and use due to privacy and regulations [3].
Today, with innumerable innovations occurring daily, it is difficult for doctors to stay updated with the latest medical knowledge, technology and techniques [6]. One of the most challenging problems for doctors is finding the right clinical trial for a patient if a large number of options are available. AI opens up new possibilities for personalized medicine by being adaptive to an individual’s genetics and being able to search through clinical research at a speed impossible by humans [3]. For example, an AI tool called IBM Watson can read through hundreds of clinical trial protocol to find the right one for a particular patient and inform the doctor [6]. Similarly, to treat cancer effectively, doctors must identify molecularly distinct cancer subtypes, potential targets and drug combinations, which requires high quality analysis of very specific data [5]. Having electronic medical records and molecular diagnostics presents AI the opportunity to filter through the data and rapidly determine the best personalized cancer treatment option.
Another example of AI used to detect cancer is the Magnetic Resonance Imaging (MRI) and Ultrasound Robotic Assisted Biopsy (MURAB) project [1]. This developing technology will make it possible to make more precise and effective biopsies to diagnose cancer. The project will create a robot that will scan a patient’s body using a combination of MRI and ultrasound technology. Currently, cancer screening techniques provide a false negative rate of 10-20%, informing patients that they do not have cancer, when in fact, they do. The new MURAB robot will enhance patient comfort and quality of care as it will only take 15-20 minutes instead of the standard 45-60 minute MRI scan time; it will also have the potential to identify early-stage signs of cancer. The major benefit of this project is that it will be able to use highly accurate MRI technology without high costs [1]. Google’s DeepMind division uses a similar AI tool to help doctors analyze tissue samples to determine the likelihood that cancers will spread along with developing the best radiotherapy treatments [3].
Although AI may not necessarily be able to find a cure, it can help make correct diagnoses faster as well as understand people’s behaviors and habits [3]. Its use to analyze an individual’s mental health provides an opportunity for early screening and detection of mental illnesses. Research by Florida State University’s Jessica Ribeiro found that AI can predict with 80 to 90% accuracy whether someone will attempt suicide in the next two years [3]. Facebook also uses AI as part of a test project to prevent suicides by analyzing the content of social network posts [3]. Similarly, scientists from Harvard and the University of Vermont developed a tool that enables computers to learn and identify signs of depression by studying Instagram posts [3].
Taken beyond the doctors’ office, AI technology can be used to help patient recovery as well as be used in therapy. AI can be designed to encourage patients to meet their wellness goals and to adhere to treatment and medications. For example, Welltok’s Cafewell Concierge app based on Watson, offers customized health solutions to patients [6]. Even therapeutic animal robots are being developed using AI to help patients in long-term care or with chronic diseases such as Alzheimer’s to recover [6]. These pet robots can help stimulate brain function in patients to improve cognition, thus improving the quality of their lives.
The use of AI can also be applied to infectious diseases and outbreaks. In the case of vector control, a predictive real-time response using AI would be more effective in preventing outbreaks compared to a passive response [4]. Most countries worldwide have many good measures for vector prevention including fumigation, larvicides and genetically modified mosquitoes; these measures would be effective in prevention only by knowing where and when the next outbreak were to occur.
Using vectors as an example, Asia spends over US$300 million a year on mosquito control [4]. Dengue, a mosquito-borne viral infection that can lead to hospitalization, is particularly common across Asia. Artificial Intelligence in Medical Epidemiology (AIME) program system is reportedly able to detect dengue outbreaks by analyzing medical data, databases and variables influencing the spread of disease in real time [2]. AIME cross-references data, searching through approximately 90 databases and taking into account 276 variables that influence the spread of dengue. The system analyzes data provided by doctors and databases containing all known cases of dengue disease [2]. AIME considers factors and variables such as wind speed, local roof architecture, water accumulation and population density, that can be used to predict the next dengue outbreak [4]. It is capable of detecting dengue outbreaks up to three months in advance and detects the location of the outbreak exactly within a 400-meter radius [2]. It also provides outbreak responders with the most effective intervention for the particular area, such as fogging mosquitoes or removing breeding sites [4]. The program’s designers believe it is over 80% reliable and accurate in predicting dengue outbreaks [2]. Since January 2018, the Malaysian state of Penang has been using the program and it is also being tested in several cities in Asia and South America.
Something to consider, however, is that AI systems such as AIME can work efficiently, but become complicated as people and the environment react to the information provided and start taking preventative measures [4]. When the environment reacts, all the factors and variables used to predict outbreaks change and must be adjusted; the years spent studying intricate relationships at a time when outbreaks were not actively prevented are not representative of the present and must be updated and refined [4].
Sources:
[1] “About MURAB · Murab.” Murab, European Commission, 2018, www.murabproject.eu/about-murab/.
[2] “Artificial Intelligence: Tool with the Potential to Detect Dengue Outbreaks Three Months in Advance.” Newsroom Allianz Worldwide Partners Business Insights, Allianz Partners, 2 June 2018, allianzpartners-bi.com/news/artificial-intelligence-tool-with-the-potential-to-detect-dengue-outbreaks-three-months-in-advance-f7eb-333d4.html.
[3] “Experts See Advances in AI Helping Diagnose, Find Cures for Hard-to-Treat Diseases.” HPN Online, KSR Publishing Inc., 6 June 2017, www.hpnonline.com/experts-see-advances-ai-helping-diagnose-find-cures-hard-treat-diseases/.
[4] Irwin, Aisling. “This AI Tool Could Detect Dengue Outbreaks 3 Months in Advance.” The Next Web, MaxCDN, 16 May 2018, thenextweb.com/syndication/2018/05/16/this-ai-tool-could-detect-dengue-outbreaks-3-months-in-advance/.
[5] Tenenbaum, Jay M., and Jeff Shrager. “Cancer: A Computational Disease That AI Can Cure.”Association for the Advancement of Artificial Intelligence, 2011, www.aaai.org/ojs/index.php/aimagazine/article/view/2345.
[6] Thoggy, Arina. “Can Artificial Intelligence Diagnose and Cure Diseases?” Big Data Made Simple, Crayon Data, 30 Nov. 2017, bigdata-madesimple.com/can-artificial-intelligence-diagnose-and-cure-diseases/.
