Artificial Intelligence – Revolutionizing the Healthcare Industry

Artificial Intelligence – Revolutionizing the Healthcare Industry

Published on October 27, 2023

What is Artificial Intelligence?

Artificial intelligence (AI) is a broad term for a category of algorithms and models that perform tasks and exhibit behaviors such as learning and making decisions and predictions. “Machine learning (ML)” is the subset of AI that allows ML training algorithms to establish ML models when applied to data, rather than models that are explicitly programmed (1).

In simpler terms, AI refers to machines (including software) that perform functions that normally require human cognition without direct human aid (2). In other words, it is a branch of computer science with development of systems that can perform tasks that would usually require human intelligence, such as problem-solving, reasoning and recognition. AI is anticipated to impact multiple areas of health care including, but not limited to, process optimization, drug discovery and development, preclinical research, drug manufacturing, clinical pathways, and patient-facing and population-level applications (3). It can support health professionals improve patient outcomes through optimizing areas such as, emergency medicine, diagnostics, treatments, preventative medicine and care delivery (4,5).

Types of AI in healthcare

AI is an umbrella term encompassing several subfields and approaches. In health care, AI systems often include, but are not limited to, one or more of the following (3):

Machine Learning

The health care sector has a historic interest in prediction, making a subset of AI, machine learning (ML), of particular relevance, notably for the detection of disease and for personalized treatment planning. ML is one of the most common forms of AI and involves training an algorithm to perform tasks by learning from patterns in data rather than performing a task it is explicitly programmed to do. To train a ML program, data are typically divided into training sets (where a human indicates whether an outcome of interest is present or absent) and validation sets (where the system uses what it learns to indicate the presence or absence of outcomes of interest). Its application in healthcare includes precision medicine – predicting what treatment protocols are likely to succeed on a patient based on various patient attributes and the treatment context. ML algorithms can be “locked”, so their function does not change, or “adaptive”, meaning their behavior can change over time (3,6).

Support Vector Machine

A support vector machine is a type of machine learning that is used mainly to classify subjects into two groups, often used for the diagnosis or prediction of disease (3).

Artificial Neural Network

A more complex form of machine learning is the neural network. It is used for categorization of applications like determining whether a patient will acquire a particular disease. It views problems in terms of inputs, outputs and weights of variables or ‘features’ that associate inputs with outputs. The network adapts to the information that it is provided (such as images) and, through a series of layered calculations, learns on its own what features can be used to determine specified outputs such as the presence or absence of a condition (3,6).

Deep Learning

The most complex forms of machine learning involve deep learning, or neural network models with many levels of features or variables that predict outcomes. There may be thousands of hidden features in such models, which are uncovered by the faster processing of today's graphics processing units and cloud architectures. A common application of deep learning in healthcare is recognition of potentially cancerous lesions in radiology images. Deep learning is increasingly being applied to radiomics, or the detection of clinically relevant features in imaging data beyond what can be perceived by the human eye. Both radiomics and deep learning are most found in oncology-oriented image analysis. Their combination appears to promise greater accuracy in diagnosis than the previous generation of automated tools for image analysis, known as computer-aided detection or CAD (6).

Natural Language Processing

This field, NLP, includes applications such as speech recognition, text analysis, translation and other goals related to language. There are two basic approaches to it: statistical and semantic NLP. Statistical NLP is based on machine learning (deep learning neural networks in particular) and has contributed to a recent increase in accuracy of recognition. It requires a large ‘corpus’ or body of language from which to learn (6).

In healthcare, the dominant applications of NLP involve the creation, understanding and analysis of clinical documentation (such as electronic medical records) and published research or as an augmented agent to respond to patient questions. NLP systems can analyze unstructured clinical notes on patients, prepare reports (such as on radiology examinations), transcribe patient interactions and conduct conversational AI (3,6).

Emergence of AI in Canada

The availability of powerful, low-cost computers has led to a burst of innovation in AI systems. In health care, this innovation is supported by the increased availability of data from sources such as electronic health records (EHR), clinical and pathological images, and wearable connected sensors. This data can be used to train algorithms and provide more opportunities for these systems to practice and learn. The current health care landscape in Canada has been noted to be conducive to the adoption of AI. For example, the Canadian Association of Radiologists (CAR) stated that, “The integrated nature of the Canadian health care system makes it ideal for pooling anonymized medical data from several institutions or provinces, which is required to improve and validate AI tools for patient management”(7).

The rapid evolution of AI technologies is expected to improve healthcare and change the way it is delivered (8). For example, AI is being explored, along with other tools, as a means of increasing diagnostic accuracy, improving treatment planning, and forecasting outcomes of care (9). The primary type of AI research pursued in Canada over the years is known as neural networks, or deep learning.  According to the Government of Canada Report on Integrating Robotics, Artificial Intelligence and 3D Printing Technologies into Canada’s Healthcare Systems  AI has, and will have, applications in: direct patient care to improve medical decision-making in diagnostics, prognosis, selecting treatment methods and in providing robotic surgeries and examinations; indirect patient care such as optimized hospital workflows and improved inventory management; and, in homecare where wearable devices and sensors will be used to assess and predict patient needs (8). AI has shown particular promise for clinical application in image-intensive fields, including radiology, pathology, ophthalmology, dermatology, and image-guided surgery, as well as broader public health purposes, such as disease surveillance (3).

Health Canada Regulations and Involvement

Health Canada is seeing the emergence of machine learning predominantly in image-based healthcare applications (e.g. diagnostic imaging/radiology). Several licenses have already issued that employ machine learning. Health Canada recognizes that AI has the potential to revolutionize the healthcare sector, including advancements in diagnosis, disease onset prediction, prognosis and more. However, the current regulatory approach used by Health Canada to approve AI-based medical devices is not suited for the iterative and data-driven nature of AI development. Health Canada has faced a variety of challenges in developing an appropriate framework for AI/ML medical devices (10).

In October 2021, Health Canada, the US Food and Drug Administration (FDA), and the United Kingdom’s Medicines and Healthcare Products Regulatory Agency (MHRA) jointly published the Good Machine Learning Practice for Medical Device Development: Guiding Principles. The document consists of 10 guiding principles to help promote safe, effective, and high-quality use of artificial intelligence and machine learning (AI/ML) in medical devices and identify areas where the International Medical Device Regulators Forum (IMDRF), international standards organizations and other collaborative bodies could work to advance GMLP. Areas of collaboration include research, creating educational tools and resources, international harmonization, and consensus standards, which may help inform regulatory policies and regulatory guidelines (11).

On September 18, 2023, Health Canada published draft pre-market guidance for machine learning-enabled medical devices. This Guidance helps manufacturers that submit an application for a machine ‎learning-enabled medical device ("MLMD"), outlines Health Canada's expectations for demonstrating the safety and ‎effectiveness of an MLMD throughout its lifecycle, and introduces a mechanism to pre-authorize planned changes to an MLMD to address risks through a pre-determined change ‎control plan ("PCCP") (1).

AI/ML medical devices have the potential to revolutionize the healthcare sector, as they continuously learn from and improve on their performance using real-world data after deployment in the market (11). A number of AI applications have already been approved by Health Canada. Some of these are described below (12):

  1. Critical Care Suite (GE Healthcare): This is a collection of AI algorithms embedded on a mobile X-ray device. Built in collaboration with Humber River Hospital in Toronto, using GE Healthcare’s Edison platform, this AI algorithm helps to reduce the turn-around time it can take for radiologists to review a suspected pneumothorax, a type of collapsed lung. It integrates seamlessly with an existing X-ray workflow and assists in delivering the highest quality care to patients.
  2. AI-RAD COMPANION (SIEMENS HEALTHCARE GMBH): This is an AI-powered, augmented, image-based clinical decision-making procedure, which creates solutions to help radiologists cope with growing demands and challenges when reading images and planning treatment. Once the computed tomography (CT), magnetic resonance (MR), or X-ray images are processed by the software solutions, the deep learning algorithms automatically support analysis of the data by providing valuable annotated clinical images, quantifications, structured findings, and comprehensive reports. All results can be reviewed and checked either on the confirmation screen or directly in the Picture Archiving System (PACS), and then can be determined to be accepted or declined. The confirmed results are then sent to the PACS, the Treatment Planning System (TPS), or to another selected target node, with one click. All the while, every step remains under the clinicians’ control. It is available as AI-RAD Companion (Cardiovascular), AI-RAD Companion (Musculoskeletal), AI-RAD companion (Pulmonary), AI-RAD Companion (Brain MR), AI-RAD Companion (Chest X-Ray), AI-RAD Companion (Organs RT) and AI-RAD Companion (Prostate MR) in Canada.
  3. Advanced intelligent Clear-IQ Engine (AiCE) for MR (CANON MEDICAL SYSTEMS CORPORATION): AiCE is the world's first magnetic resonance (MR) Deep Learning reconstruction technology, producing stunning MR images that are exceptionally detailed and with the low-noise properties which results in higher signal-to-noise ration (SNR) and enables increased resolution. By improving SNR, sharp, clear and distinct images can be achieved utilizing the power of Deep Learning to see through the noise. AiCE is applicable to all body regions and almost all sequences. AiCE has expanded body imaging solutions to provide fast, accurate, and essential information to help you provide better care for your patients, now covering up to 96% of all MR procedures.
  4. EnsoSleep (EnsoData): EnsoSpleep is an AI-powered with software with one cloud-based platform that helps clinicians for viewing, scoring, editing and reporting of polysomnography (PSG) and home sleep apnea testing (HSAT). Compatible with the leading PSG and HSAT devices, EnsoSleep simplifies and accelerates the sleep testing, diagnosis, and treatment workflow, providing sleep teams with more opportunities to expand patient care, improve outcomes, reach more patients, and impact sleep center growth.

AI in Drug Development and Clinical Trials

Clinical Trials remain the foundation of safe and effective drug development. Given the evolving data-driven and personalized medicine approach in healthcare, companies and regulators are beginning to utilize tailored AI solutions that enable expeditious and streamlined clinical research (13). AI has many potential applications in clinical trials both near- and long-term including but not limited to, automating routine study data entry functions, analyzing EHR data to find suitable candidates and sites for clinical studies, patient recruitment, monitoring and encouraging patient compliance with study protocols, predicting patient outcomes in clinical trials, predicting probability of clinical trial success, adaptive dose-finding, discovering and modelling potential new molecules and therapies (13,14). 

AI has many near- and long-term applications for improving clinical research returns (13), including:

  • Patient identification - AI capabilities, including natural language processing and association rule mining, help extract data from unstructured medical records to find patients suitable for clinical studies, and can help identify those most likely to complete a trial.
  • Site selection - Helps identify sites with the right patients and capabilities to successfully recruit and retain patients.
  • Patient monitoring and support - AI enabled mobile devices such as patient wearables that monitor specific aspects of patients’ health relevant to the clinical trial, help identify when patients deviate from protocols and send reminders.
  • Cohort composition - AI helps identify biomarkers to find patients most likely to show benefit from a particular dose or combination therapy.

Early drug discovery is one of the areas with significant interest and activity in utilizing AI/ML. As a starting point, the process of identifying biological targets and elucidating disease relationships can utilize AI/ML to analyze and synthesize significant amounts of information from existing scientific research, publications, and other data sources. The growth of available genomic, transcriptomic, proteomic, and other data sources from healthy persons and those with a specific disease of interest provide a significant opportunity to inform biological target selection. These datasets are often complex and originate from disparate sources, which can be well-suited for the utilization of AI/ML approaches. Building from existing validated data, AI/ML can be applied to mine and analyze these large multi-omics and other datasets to provide information on the potential structure and function of biological targets to predict their role in a disease pathway (15).

The United States Food and Drug Administration (US FDA) had recently released a discussion paper on Using Artificial Intelligence and Machine Learning in the Development of Drug and Biological products in areas such as early drug discovery, compound screening and design, nonclinical and clinical research (recruitment, selection and stratification of trial participants, dosing regimen optimization, participant adherence, retention, site selection etc.), collection, management and analysis of clinical trial data, clinical endpoints assessment, post-market safety surveillance. Considering the rapid technological innovations in data collection and generation tools, combined with robust information management and exchange systems and advanced computing abilities, the use of AI/ML has the potential to accelerate the drug development process and make clinical trials safer and more efficient (15).

A range of opportunities are already identified in literature, starting with AI’s contribution to discovery in areas where return on investment might not support profitability (rare diseases, targeted therapies) (13). Still, AI has its limits and must be handled with care to ensure it is producing valid, reliable results. It is important to assess whether the use of AI/ML introduces specific risks and harms. For example, AI/ML algorithms have the potential to amplify errors and preexisting biases present in underlying data sources and, when the findings are extrapolated outside of the testing environment, raise concerns related to generalizability and ethical considerations. Additionally, an AI/ML system may exhibit limited explainability due to its underlying complexity or may not be fully transparent for proprietary reasons. Given the conservative nature of clinical research, in addition to the cost and complexity of developing AI solutions, it is likely to be a long time before their full potential can be realized (14,15).

AI in Drug Manufacturing

A critical aspect of drug development includes the methods, facilities, and controls used in manufacturing, processing, packing, and holding of a drug to help ensure that the drug meets the requirements of safety and effectiveness, has the identity and strength it is represented to possess, and meets quality and purity characteristics. Advanced analytics leveraging AI/ML in the pharmaceutical manufacturing industry offers many possibilities, including, but not limited to, enhancing process control, increasing equipment reliability and throughput, monitoring early warnings or signals that the manufacturing process is not in a state of control, detecting recurring problem clusters, and preventing batch losses. The use of AI/ML to support pharmaceutical manufacturing can be deployed together with other advanced manufacturing technologies (e.g., process analytical technology, continuous manufacturing) to achieve the desired benefits. AI/ML is an enabler for the implementation of the rapidly evolving technologies, and could result in a well-controlled, hyper-connected, digitized ecosystem and pharmaceutical value chain for the manufacturer (15).

AI/ML could also be used to improve the reliability of the manufacturing supply chain through forecasting product demand, analyzing production schedules, estimating and mitigating the impact of potential disruptions, and optimizing inventory. Use of AI/ML based approaches in pharmaceutical manufacturing cover the entire drug manufacturing life cycle, from design to commercial manufacturing. Some of the examples of the potential applications of AI in pharmaceutical manufacturing include but not limited to, process design and scale up, advanced process control (APC), process monitoring and fault detection, trend monitoring. The US FDA has released a discussion paper on Artificial Intelligence in Drug Manufacturing and has reopened it for industry’s comments until 23 November, 2023 (16).

Ethical, Legal and Social Dimensions of Artificial Intelligence

The ethical, legal, and social dimensions of AI in health and health care is a flourishing area of inquiry. The World Health Organization (WHO) has published a report on Ethics and Governance of AI for Health, recognizing that AI holds great promise for the practice of public health and medicine. At the same time, for AI to have a beneficial impact on public health and medicine, ethical considerations and human rights must be placed at the center of the design, development, and deployment of AI technologies for health. For AI to be used effectively for health, existing biases in healthcare services and systems based on race, ethnicity, age, and gender, that are encoded in data used to train algorithms, must be overcome. Governments will need to eliminate a pre-existing digital divide (or the uneven distribution of access) to the use of information and communication technologies. Such a digital divide not only limits use of AI in low- and middle-income countries but can also lead to the exclusion of populations in rich countries, whether based on gender, geography, culture, religion, language, or age. The proliferation of AI could lead to the delivery of healthcare services in unregulated contexts and by unregulated providers, which might create challenges for government oversight of health care. Therefore, appropriate regulatory oversight mechanisms must be developed to make the private sector accountable and responsive to those who can benefit from AI products and services and can ensure that private sector decision-making and operations are transparent (2).

Conclusion

AI has enormous potential for strengthening the delivery and accessibility of health care and medicine. This includes improved diagnosis and clinical care, enhancing health research and drug development and assisting with the deployment of different public health interventions, such as disease surveillance, outbreak response, and health systems management. Adopting AI into health care will come with challenges, but some promising work has already occurred. Canada is uniquely positioned on a number of fronts to be a global leader in the application of AI to healthcare with medical imaging already having tangible advances. There is tremendous potential for Canada to be a leader in healthcare innovation by leveraging recent advances in AI – including key contributions by Canadian researchers (2, 7).

Author: Pratibha Duggal, ICON Plc.

References:

  1. Health Canada Draft Guidance Document. Pre-market guidance for machine learning-enabled medical devices. Accessed September 25, 2023. https://www.canada.ca/en/health-canada/services/drugs-health-products/medical-devices/application-information/guidance-documents/pre-market-guidance-machine-learning-enabled-medical-devices.html.
  2. World Health Organization (WHO). 2021, June 28. Ethics and Governance of Artificial Intelligence for Health: WHO Guidance. Accessed September 25, 2023. https://www.who.int/publications/i/item/9789240029200 .
  3. Canada’s Drug and Health Technology Agency (CADTH). An Overview of Clinical Applications of Artificial Intelligence (November 2022). Accessed September 26, 2023. https://www.cadth.ca/overview-clinical-applications-artificial-intelligence 
  4. Canadian Foundation for Healthcare Improvement. Artificial Intelligence (AI) in Healthcare. Accessed 26 September 2023. https://www.cfhi-fcass.ca/opportunities/webinars/artificial-intelligence-in-healthcare#:~:text=Emerging%20technologies%20continue%20to%20transform,preventative%20medicine%20and%20care%20delivery\ 
  5. Eastwood, K.W., May, R., Andreou, P. et al. Needs and expectations for artificial intelligence in emergency medicine according to Canadian physicians. BMC Health Serv Res 23, 798 (2023). https://doi.org/10.1186/s12913-023-09740-w
  6. Davenport T, Kalakota R. The potential for artificial intelligence in healthcare. Future Healthc J. 2019 Jun;6(2):94-98. doi: 10.7861/futurehosp.6-2-94. PMID: 31363513; PMCID: PMC6616181. 
  7. Canadian Association of Radiologists. Artificial intelligence. [2018]; Accessed 27 September 2023. https://car.ca/innovation/artificial-intelligence/
  8. Government of Canada, Standing Senate Committee on Social Affairs, Science and Technology. Challenge Ahead: Integrating Robotics, Artificial Intelligence and 3D Printing Technologies into Canada’s Healthcare Systems. 2017 Oct. 
  9. Canadian Medical Protective Association (CPMA). The emergence of AI in healthcare: Risks, regulation, and a measured approach to adoption. Published September 2019. Revised: May 2023. 
  10. Health Canada Presentation on Regulatory challenges of AI products: A pre-market perspective. 15 April, 2019. Accessed 27 September 2023. https://www.cadth.ca/sites/default/files/symp-2019/presentations/april15-2019/A3-presentation-tdumouchel.pdf 
  11. Health Canada Guidance Document: Good Machine Learning Practice for Medical Device Development: Guiding Principles. Published Oct 2021. Accessed 27 September 2023. https://www.canada.ca/en/health-canada/services/drugs-health-products/medical-devices/good-machine-learning-practice-medical-device-development.html 
  12. Health Canada. Medical Devices Active Licence Listing (MDALL). 
  13. Askin S, Burkhalter D, Calado G, El Dakrouni S. Artificial Intelligence Applied to clinical trials: opportunities and challenges. Health Technol (Berl). 2023;13(2):203-213. doi: 10.1007/s12553-023-00738-2. Epub 2023 Feb 28. PMID: 36923325; PMCID: PMC9974218. 
  14. ICON plc. Digital Disruption Technologies that accelerate and improve clinical trials. Accessed 28 September 2023. Digital disruption and clinical trials | AI | Wearables | RWD (iconplc.com) 
  15. United States Food and Drug Administration (US FDA) Discussion Paper: Using Artificial Intelligence and Machine Learning in the Development of Drug and Biological products. Accessed 29 September 2023. Using Artificial Intelligence & Machine Learning in the Development of Drug and Biological Products (fda.gov
  16. United States Food and Drug Administration (US FDA) Discussion Paper: Artificial Intelligence in Drug Manufacturing. Accessed 29 September 2023. Artificial Intelligence Discussion Paper (fda.gov)

Related Articles

VIRTUAL HEALTHCARE IN CANADA

VIRTUAL HEALTHCARE IN CANADA

“Virtual care” is defined as any “interaction between patients and/or members of their circle of care, occurring remotely, using any forms of communication or...

Technology Continues to Transform Healthcare and the Regulatory World

Technology Continues to Transform Healthcare and the Regulatory World

Introduction: Digital health solutions are advancing rapidly in the field of diagnostics, emergency response, hospital management. As the goal of individuals shifts more...