First Line Software is a premier provider of software engineering, software enablement, and digital transformation services. Headquartered in Cambridge, Massachusetts, the global staff of 400 technical experts serve clients across North America, Europe, Asia, and Australia.
Many of us think that artificial intelligence (AI) is about connected and driverless cars, smart homes, and robotics. But AI is much more than that. AI experts see it as a revolutionary technology that could help many industries, with healthcare among them. Key clinical health AI applications hold the potential to create $150 billion in annual savings for the US healthcare economy by 2026. The market value of AI in the health care industry is predicted to reach $6.6 billion by 2021.
The impact of AI in the health care sector is truly life-changing. AI is driving innovations in clinical operations, drug development, surgery, and data management. There are many recent examples.
AI as a doctor
In 2017, a robot in China passed the medical licensing exam using its AI brain. During that same year, the first semi-automated surgical robot stitched blood vessels as narrow as 0.03 mm. Surgical robots are being credited with performing operations with a precision that rivals best-skilled surgeons. Finally, a robot dentist now uses its AI skills to autonomously perform complex and delicate dental procedures.
In a related development, KT and Samsung Medical Center are working together to enable 5G medical service. Services include digital diagnostic pathology, access to proton therapy information, teaching surgery, AI-enabled care for in-patients, and an autonomous robot for operating rooms.
The NYU School of Medicine has tested the use of an AI-powered ‘radiology assistant’ to reduce the frequency of more required imaging after the first screening mammography exam. Radiologists will see images the way they currently see them but when there is something unusual they will ask the AI “assistant” for an opinion. Results to date have been impressive. Interestingly, the radiologists and AI use different features of the data, which together has made the average findings more accurate than when viewed separately.
AI for diagnostics
Google’s DeepMind has developed an AI system through a joint research partnership with the Moorfields Eye Hospital in London. The system trained a neural network that can accurately detect over 50 types of eye diseases as accurately as a doctor by only analyzing 3D retinal scans. These results show just how effective AI technology can be at identifying real anomalies.
In other AI health news, Google’s AI platform is now aiding oncologists in breast cancer screenings. The results of a study by Google Health published in Nature indicate its model was able to spot cancer more quickly and accurately in de-identified screening mammograms with fewer false positives and false negatives than experts. While doctors had access to patient histories and prior screenings, the AI model relied solely on X-ray images.
Recently, the Food and Drug Administration (FDA) approved a software product from Caption Health that makes it easier for doctors and other medical professionals to take ultrasound pictures of the heart – also known as echocardiograms. The FDA approved the software developed by this artificial intelligence startup, through its “de novo” pathway, a process used for brand new medical devices.
The opportunities for clinical improvements using AI are broad and diverse. According to experts, the technology is expected to continue making big strides with healthcare in 2020. A case in point is AI technology based on a deep learning model that has demonstrated the potential for helping cardiologists predict irregular heart rhythm – atrial fibrillation before it develops. This conclusion was made from two studies conducted by Geisinger researchers that were presented at the American Heart Association Scientific Sessions 2019.
The American Diabetes Association’s (ADA) new 2020 Standards of Medical Care in Diabetes recognizes the use of autonomous artificial intelligence for the screening of some medical conditions. The ADA’s new set of clinical standards includes language stating that the use of AI systems authorized for use by the FDA with a specific capability to detect more than mild diabetic retinopathy and diabetic macular edema “represents an alternative to traditional screening approaches.” This move will likely help win over those who are skeptical about the value and efficacy of AI in clinical care.
AI for predictive medicine
Precision medicine refers to dispensing accurate treatments that are based on an accurate diagnosis of the patient’s specific symptoms and behavior.
Recently, Mayo Clinic announced the launch of a new Clinical Data Analytics Platform, which is part of its data-driven Mayo Clinic Platform initiative. This platform will apply advanced analytics on de-identified data from across Mayo and other health systems, and search the scientific literature for insights that will help speed drug discovery and development.
AI for drug discovery
The drug research process is complicated, lengthy, and expensive. Only five in 5,000 drugs that make it to the pre-clinical trial stage will ever advance to human testing. And only one of those five drugs will be released for use in pharmacies.
Many pharmaceutical giants like Sanofi and Pfizer are teaming up with tech companies such as IBM and Google, that are already invested in AI technology. Their plan is to build a drug discovery program using deep learning and AI. The results are already paying off.
MIT and Takeda announced their collaboration on new healthcare AI applications. Together with MIT’s School of Engineering, the pharma company will establish an education program focused on developing new machine learning approaches to drug development and more. The goal of the MIT-Takeda Program is to fund as many as ten flagship research projects per year in the areas of machine learning and health. Projects will include diagnosis of disease, prediction of treatment response, development of novel biomarkers, process control, and improvement, drug discovery, and clinical trial optimization.
French artificial intelligence (AI) startup Iktos announced a research collaboration with global skin-health-focused pharmaceutical company Almirall to accelerate its drug discovery process. In this partnership, Almirall will use the Iktos generative modeling technology to optimize advanced compounds.
More solutions on the way
During HIMSS20, the AI-based clinician information systems vendor, Nuance Communications will introduce new AI-powered systems for imaging to radiologists who visit the Nuance AI Marketplace for Diagnostic Imaging. The marketplace provides a one-stop shop for radiologists to review and purchase AI models to improve their workflow.
Similar to the concept of an app store, this digital marketplace enables subscribers to purchase applications, such as the AI models offered by Nuance. It also provides the opportunity for developers to connect with radiologists and hear directly about evolving needs that inform them to identify ways for improving the functionality of algorithms.
Significant investments in AI development for healthcare are being made by many companies. Microsoft recently announced AI for Health, a five-year, $40 million program designed to help healthcare organizations leverage artificial intelligence and machine learning for improved patient care and population health. GE Healthcare aims to accelerate the adoption of AI and ML applications and developer services for health systems with its new Edison Developer Program.
In the UK, a new AI in Health and Care Award has been launched by health secretary Matt Hancock. The goal of the award is to help speed up testing, evaluation, and adoption of the “most promising” AI technologies for healthcare by making £140 million available to innovators over the next three years. Applications focused on screening, diagnosis, clinical decision support, and system efficiency will be the focus during the early stages of this program.
First Line Software is actively involved in developing AI technologies for healthcare for our customers. We’ll be attending HIMSS20 in March and we invite you to stop by our booth (#1070). You can also schedule a brief meeting to tell us about your healthcare technology initiatives and learn more about our capabilities.