Imagine a smartphone app that can help diagnose skin cancer from a simple photograph, before you even see the doctor. Technology like this may become a reality in the near future with the rising prominence of artificial intelligence (AI) tools used in healthcare applications.
A recent study conducted by researchers at NYU School of Medicine found that an AI tool can identify cancer types and genetic changes in lung tumors with impressive accuracy. The computer program uses machine learning - a type of AI - to analyze images of patients' lung tumors and distinguish between adenocarcinoma and squamous cell carcinoma with 97 percent accuracy. These types of lung cancers can be difficult at times for even experienced pathologists to decipher, making the tool a valuable addition in diagnosing and treating cancer patients. The tool was also able to identify the presence of abnormal versions of six genes linked to lung cancer in cells with 73 to 86 percent accuracy, again using automated analysis of visual clues.
The use of AI is gaining traction in the changing realm of healthcare, as improvements in technology lead to tools that can enhance the abilities of physicians, lead to improved outcomes of patients, and even bring medicalservices to underserved rural areas. In 2018, the FDA issued their first approval of marketing of a medical device that uses AI to detect diabetes-related eye problems. Further, the American Medical Association, an ANSI member, adopted a new policy on the use of AI in healthcare at their 2019 annual meeting. The guidelines indicate that AI should "enhance the patient experience of care and outcomes, improve population health, reduce overall costs for the health care system while increasing value, and support the professional satisfaction of physicians and the health care team."
With these new technologies comes the need for standards - both for the development of the tools themselves, and for the use of them in an ethical and productive manner. Standards for AI in general, though not specific to healthcare, have been in use for many years. Members of the American National Standards Institute and international standards developers have contributed these guidelines, among others, to AI's development and use:
INCITS/ISO/IEC 2382-28, Information Processing Systems - Vocabulary - Part 28: Artificial Intelligence - Basic Concepts and Expert Systems, prepared by Joint Technical Committee ISO/IEC JTC 1, Information technology, Subcommittee SC 1, Vocabulary.
The International Organization for Standardization (ISO) and the IEC have a joint technical committee (JTC) subcommittee that provides guidance to JTC, IEC, and ISO committees developing AI applications. ANSI serves as the secretariat to ISO/IEC JTC 1, Information technology, SC 42, Artificial intelligence. ANSI is the U.S. member body to ISO.
Development of AI standards specific to healthcare is emerging. Earlier this year, ANSI member and accredited standards developer the Consumer Technology Association (CTA) formed a group of more than 30 organizations to develop standards and best practices for healthcare AI. The group includes ANSI members the BlackBerry, Google, IBM, Philips, and Verizon, to name a few. The initiative aims to drive industry consensus and standardization on definitions and characteristics of AI, and to address issues surrounding the use and application of AI in healthcare such as trustworthiness, ethics, and bias.
ANSI member and accredited standards developer AAMI has been active in this space as well, recently releasing a position paper on the topic. "The emergence of artificial and machine learning in healthcare: Recommendations to support governance and regulation" analyzes what standardization can offer to AI use in healthcare, and provides recommendations for future standardization activities that will support the deployment and use of these technologies. The report is freely available for all interested stakeholders.
As AI helps improve the way doctors diagnose patients and continues to play a major role in behind the scenes research, the need for standardization will growto support both ethical and practical AI applications. The standardization community depends on experts that can drive all aspects of the standardization to empower AI in the nearfuture, and for years to come.