Volume 38 Number 3 | June 2024

Christopher Swartz, PhD, MLS(ASCP), ASCLS Today Volunteer Contributor

Christopher SwartzMy “official” introduction to artificial intelligence (AI) in laboratory medicine was at a departmental lunch in January 2023. As we prepared to leave the restaurant, my colleague typed a prompt into ChatGPT on his smartphone, asking for a summary of relevant clinical laboratory testing related to the treatment of diabetes patients. We reviewed the AI response to his prompt and saw that the information was correct, including the types of laboratory testing, specimen types, and appropriate reference ranges. That was an enlightening experience and foreshadowed the rise of AI-related applications in laboratory medicine, which are becoming increasingly ubiquitous.

Artificial intelligence is a transformational technology currently at the vanguard of high-profile research topics and will undoubtedly play a major role in our careers as medical laboratory science professionals in the years ahead. We collectively utilize various forms of AI (perhaps unwittingly) in our personal lives already, including the use of virtual assistants such as Siri or Alexa, or through interactions with AI-powered chatbots online. The release of ChatGPT in November 2022 assisted with fueling the explosion of interest in the integration of AI into many disciplines, including laboratory medicine, and numerous generative AI platforms are now available for use by the general public.1 Examples of these tools include chat-based platforms (Google Gemini, ChatGPT, Microsoft CoPilot), image generation (DALL-E, Canva, Photoshop), and other educational tools (PhotoMath, Grammarly, Socratic by Google).2

Artificial intelligence is a field of computer science that is designed to produce systems that can mimic human behavior, including learning, thinking, and making decisions.3 Machine learning (ML) is a subfield of AI that allows systems to learn from new data without the use of specific programming, through the analysis of data and patterns that may exist in the data. Machine learning algorithms are capable of improving their predictive performance for a given task, through the use of more training data (i.e., ML can learn from prior examples).4 Different machine learning models are currently used, including supervised learning, unsupervised learning, and reinforcement learning.4 In order to train these models, large amounts of structured data are required, and the models typically improve their predictive capabilities with time.4 In supervised learning, labeled data is provided for training purposes, while unsupervised learning does not require the use of labeled data.5 One key requirement for AI is that accurate and detailed data is critically important for both training and validation purposes.3 Since predictive analytics represents one of the principle applications of AI in laboratory medicine, higher quality data can yield better predictions than low quality data.3,6,7

“Numerous applications of AI in laboratory medicine have been described in the literature. These include usage in instrument automation, error detection, predicting laboratory test values, result interpretation, assistance with streamlining laboratory test utilization, improving laboratory information systems, and genomic and image analysis.”

Numerous applications for artificial intelligence and machine learning in educational programs in the clinical laboratory sciences were recently disclosed at the Clinical Laboratory Educators Conference (CLEC) in February 2024. Potential opportunities for AI adoption by course instructors include collaboration, curricular design, and dynamic content creation.8 However, possible challenges are posed by the adoption of AI in educational programs, including issues with accessibility or bias, academic integrity, and unclear privacy standards.8 ChatGPT specifically has utility for both faculty and students; these applications include content generation, assistance with various administrative tasks, student assessment, and opportunities for student engagement and feedback.9 Additional examples for the use of AI in the classroom include AI-generated rubrics, case studies, discussion questions, and image generation.10

Numerous applications of AI in laboratory medicine have been described in the literature. These include usage in instrument automation, error detection, predicting laboratory test values, result interpretation, assistance with streamlining laboratory test utilization, improving laboratory information systems, and genomic and image analysis.6,7 Examples of FDA-approved automated analyzers that utilize AI or machine-learning are currently in use, including one analyzer that performs automated image analysis, cell classification, and enumeration of erythrocytes, leukocytes, and platelets in blood and body fluids, and another automated system utilizes neural network approaches for urine sediment analysis.6 Additional applications of AI and machine learning in laboratory medicine include diagnosis of rheumatic diseases, patient blood management, and point-of-care testing.6,11,12,13 The use of AI-assisted image analysis for infectious disease testing in clinical microbiology (mycobacteria detection, ova and parasite examinations, etc.) is also starting to become more important in clinical laboratories.14

Despite these applications, there are consequential roadblocks for widespread adoption of AI in laboratory medicine. Data quality is one of the biggest limiting factors facing AI and machine learning, along with access to high quality data.3,7 Data sets that are incorrectly labeled or that are missing key data or other information will ultimately limit the performance of an AI-based application.7 Financial costs associated with widespread integration of AI in the clinical laboratory include the cost of trained personnel and computational infrastructure.7 Ethical concerns related to patient safety and privacy are also of paramount importance.3,15 Bias in AI algorithms is also problematic, and can lead to the propagation of existing inequities in the healthcare system.16

The pace of AI deployment will be driven in part by workforce needs and perceptions towards the technology, along with recognition of the growing clinical utility of AI. Survey results reported from employees at a reference laboratory indicate that a large percentage of respondents are supportive of the development of AI projects in the laboratory, while acknowledging potential concerns related to fear of job loss or replacement caused by AI adoption.17 Medical laboratory science professionals have also described the need for increased education about AI-related topics.18 While certain challenges to adoption in AI in the clinical laboratory have been identified, these are not insurmountable. AI will unquestionably play a larger role in clinical laboratories of the future.

References
  1. Yang, HS, Wang, F, Greenblatt, MB, Huang, SX, Zhang, Y. AI Chatbots in Clinical Laboratory Medicine: Foundations and Trends. Clinical Chemistry. 2023; 69(11): 1238-1246.
  2. Stein, M, Brunson, E. Embracing the Future: How Students Harness AI for Enhanced Learning in Clinical Laboratory Sciences. Presented at the Clinical Laboratory Educators Conference, Las Vegas, NV (February 2024).
  3. Oduoye, MO, Fatima, E, Muzammil, MA, Dave, T, Irfan, H, Fariha, FNU, Marbell, A, Ubechu, SC, Scott, GY, Elebesunu, EE. Impacts of the advancement in artificial intelligence on laboratory medicine in low- and middle-income countries: Challenges and recommendations-A literature review. Health Sci. Rep. 2024; 7(1): e1794.
  4. Tran, NK, Albahra, S, May, L, Waldman, S, Crabtree, S, Bainbridge, S, Rashidi, H. Evolving Applications of Artificial Intelligence and Machine Learning in Infectious Diseases Testing. Clinical Chemistry. 2022; 68(1): 125-133.
  5. Çubukçu HC, Topcu D, Yenice, S. Machine learning-based clinical decision support using laboratory data. Clin Chem Lab Med. 2023 Nov. 29. doi: 10.1515/cclm-2023-1037. Epub ahead of print. PMID: 38015744.
  6. Haymond, S, McCudden, C. Rise of the Machines: Artificial Intelligence and the Clinical Laboratory. Journal of Applied Laboratory Medicine. 2021; 6(6): 1640-1654.
  7. Rabbani, N, Kim. GYE, Suarez, CJ, Chen, JH. Applications of Machine Learning in Routine Laboratory Medicine: Current State and Future Directions. Clin Biochem. 2022; 103: 1-7.
  8. Riding, K. Integrating Generative AI Into the Classroom. Presented at the Clinical Laboratory Educators Conference, Las Vegas, NV (February 2024).
  9. Lerret, N. Elevating Education: Harnessing ChatGPT & AI for MLS Educators and Program Directors. Presented at the Clinical Laboratory Educators Conference, Las Vegas, NV (February 2024).
  10. Erickson, K, Nelsen, A. When Virtual Reality Becomes Reality: Navigating the Past, Present, and Future of Virtual Reality and Artificial Intelligence. Presented at the Clinical Laboratory Educators Conference, Las Vegas, NV (February 2024).
  11. Galozzi, P, Basso, D, Plebani, M, Padoan, A. Artificial intelligence and laboratory data in rheumatic diseases. Clinica Chimica Acta. 2023; 546: 117388.
  12. Meier, JM, Tschoellitsch, T. Artificial Intelligence and Machine Learning in Patient Blood management: A Scoping Review. Anesthesia and Analgesia. 2022; 135(3): 524-531.
  13. Khan, AI, Khan, M, Khan, R. Artificial Intelligence in Point-of-Care Testing. Ann Lab Med. 2023; 43: 401-407.
  14. Burns, BL, Rhoads, DD, Misra, A. The Use of Machine Learning for Image Analysis Artificial Intelligence in Clinical Microbiology. Journal of Clinical Microbiology. 2023; 61(9), e0233621.
  15. Pennestri, F, Banfi, G. Artificial intelligence in laboratory medicine: fundamental ethical issues and normative key-points. Clin Chem Lab Med. 2022; 60(12): 1867-1874.
  16. Mittermaier, M, Raza, MM, Kvedar, JC. Bias in AI-based models for medical applications: challenges and mitigation strategies. npj Digital Medicine. 2023;6(1): article number 113.
  17. Ardon, O, Schmidt, RL. Clinical Laboratory Employees’ Attitudes Toward Artificial Intelligence. Lab Medicine. 2020; 51: 649-654.
  18. Paranjape, K, Schinkel, M, Hammer, RD, Schouten, B, Nannan Panday, RS, Elbers, PWG, Kramer, MHH, Nanayakkara, P. Am J Clin Pathol. 2021; 155: 823-831.

Christopher Swartz is Assistant Professor at the University of Kentucky College of Health Sciences in the Department of Health and Clinical Sciences in Lexington, Kentucky.