AI application in the Medical Device Field

 Introduction 

The term “Artificial Intelligence” (AI) is notoriously hard to define. Frequently people use it to mean things that are hard for computers to do (like understanding natural language) as opposed to things that we know computers handle quite well (like accounting). 

But, what is Artificial Intelligence? 

According to ISO/IEC 2382:2015, Artificial Intelligence is defined as the capability of a system to perform tasks or develop data processing systems that perform functions normally associated with human intelligence. 

In other words, Artificial Intelligence, or AI, is a particular type of technology that enables computers and machines to simulate human intelligence and problem-solving capabilities.  

Thanks to two disciplines, machine learning and deep learning, it is possible to develop AI algorithms, modelled after the decision-making processes of the human brain that can learn and at the same time, become more accurate during predictions over time.  

History of AI 

So, where does the term Artificial Intelligence come from? 

In the summer of 1956, at an academic conference held at Dartmouth University, several scientists discussed how to make machines simulate intelligence. McCarthy, first, proposed the term Artificial Intelligence. 

However, the history of Artificial Intelligence began in antiquity, with myths, stories and rumors of artificial beings endowed with intelligence or consciousness by master craftsmen. The seeds of modern AI were planted by philosophers who attempted to describe the process of human thinking as the mechanical manipulation of symbols. This work culminated in the invention of the programmable digital computer in the 1940s, a machine based on the abstract essence of mathematical reasoning. This device and the ideas behind it inspired a handful of scientists to begin seriously discussing the possibility of building an electronic brain. In fact, in 1943, the first artificial neuron was designed by McCulloch and Pitts.  

Since the conference in 1956, Artificial Intelligence (AI) in varying forms and degrees has been used to develop and advance a wide spectrum of fields, such as healthcare, manufacturing, banking and financial markets, education, supply chains, retail and e-commerce. 

 AI in the Healthcare Industry 

In the healthcare industry, AI-based medical devices could automate tasks, synthesize data from multiple sources and pinpoint trends. Also, it is possible to process and analyze information from wearable sensors and identify disease or the onset of medical conditions; predict which patients are at an increased risk for a disease, complications or negative outcomes based on their medical records and then, support research by evaluating large amounts of data and monitoring treatment efficacy.  

The machine learning technique evaluates structured data, like imaging, genetic and electrophysiological data and is useful to assemble patient's characteristics or predict the disease. Deep learning techniques, instead, are used for more complex data obtained from the medical data set.  

Medical sensors, which can convert biomedical parameters into easily measured signals, play an important role in the diagnostic field since, with their utilization, the diagnostic instrument can be more effective and safer. Various medical sensors have been developed for monitoring and diagnosing diseases, like biomedical markers that can be used on the body or inside the body. These kinds of sensors are actively used, for example, to detect cancer. 

Meanwhile, software that incorporates Artificial Intelligence (AI), specifically the subset of AI known as machine learning (ML), has become an essential component of an increasing number of medical devices as technology advances.  

AI/ML in software's greatest advantage is its ability to learn from real-world use and experience and improve its performance. The ability of AI/ML software to learn from real-world feedback (training) and improve its performance (adaptation) makes these technologies uniquely situated among software as a medical device (SaMD) and a rapidly expanding area of research and development. 

Practical Cases 

AI/ML has the potential to generate new and important insights from the vast amount of data generated during the delivery of healthcare every day. Digital health technologies are playing an increasingly significant role in many facets of our health and daily lives, and AI/ML is powering important advancements in this field. 

An AI/ML application designed for ICU patients, which receives electrocardiogram, blood pressure, and pulse-oximetry signals from a primary patient monitor, can be used as an example of AI/ML software. In this instance, the processing and analysis of physiological signals is done to identify patterns that occur at the start of physiologic instability. An audible alarm signal is generated when physiologic instability is detected to signal the need for immediate clinical action to prevent potential harm to the patient.  

Another example can be a mobile app intended to provide the risk assessment of skin lesions using the mobile device’s camera and a flashlight. The AI algorithm is the one which performs the risk assessment. Based on the outcome of the assessment, the user will receive a recommendation to see a dermatologist for further examination to obtain an accurate medical diagnosis. To avoid possible misunderstandings on the result screen, the algorithm plots a box around the risk-assessed lesions, colored according to the assigned risk. This indicates exactly which lesion the risk assessment was calculated for. If there are different lesions, it creates multiple boxes with the corresponding color. If the algorithm fails to detect the lesion, there will be no box on the result screen, which means the user won't be able to assess it. 

Continuous glucose monitoring (CGM) and self-management mobile apps have been used more recently, leading to a digital transformation in diabetes care. Insulin bolus calculators have been created to help with insulin dose adjustment and are now included in the majority of the most recent commercially available insulin pumps and some glucose meters. A device has been created to overcome this barrier by utilizing continuous glucose monitoring (CGM), Run-to-Run control, and Artificial Intelligence. It seems like the algorithm, implemented in a mobile app, communicates in real-time with a continuous glucose sensor and requires the user to manually input different information to calculate a recommended insulin dose that is tailored to the individual and the current circumstances. The algorithm also adopts its recommendation based on the outcomes of past recommended insulin doses and user behaviour. 

In the hearing aid field, the first hearing aid that uses real-time machine learning to empower end users to make adjustments based on their preferences and intentions in different environments has been introduced on the market. By utilizing a distributed computing approach and a smartphone connected to a hearing aid, a live machine-learning application can be incorporated. Via a simple interface, the algorithm can automatically learn and meet the end-user’s preferences and intentions. So, with the use of machine learning, it is possible to adjust the equalizer settings without altering the programming that the hearing healthcare professional has put into the fitting. While the permanent programming of the hearing aid is not altered, the end-user still has the power, in real-time, to easily refine their acoustic settings to meet their specific real-time listening intention.

Conclusion 

In the ever-evolving realm of healthcare technology, Artificial Intelligence stands as a beacon of innovation, offering once-unimaginable solutions. From real-time monitoring to personalized treatment plans, AI is not only enhancing the efficiency of medical interventions but also empowering patients to take control of their health like never before. As we look towards the future, the marriage of AI and healthcare holds the promise of a world where diseases are detected earlier, treatments are more effective, and individuals are empowered to lead healthier lives.