Diagnostic Performance of Artificial Intelligence for Detection of Anterior Cruciate Ligament and Meniscus Tears: A Systematic Review

Artificial Intelligence, Machine Learning, and Medicine: A Little Background Goes a Long Way Toward Understanding

“…Kunze et al 6 note that AI is often criticized as employing a "black box" that obscures our ability to understand what factors are used to determine a recommendation or result. This is known as a lack of "explainability."…”

“…In an editorial commentary on Kunze et al, 6 by journal Associate Editor Nikolaos Paschos, "Artificial Intelligence in Sports Medicine Diagnosis Needs to Improve," 11 he acknowledged the capabilities of AI but cautioned against adopting technologies without high-level supporting evidence. Paschos' guarded stance prompted a letter to the editor by Ramkumar et al, 8 and subsequent author replies by both Paschos 12 and Kunze et al 13 Taken in sum, the exchange was spirited, yet from our point of view, the commentary, letter, and 2 replies reflect greater consensus than controversy.…”

“…While we have a few additional background points to share, if readers take home a single point it is this: AI and machine learning is made possible by computers programmed by humans. If we keep this in mind, we may have an easier time understanding when we read about AI and machine learning in medicine, including some recent publications in Arthroscopy including: 6 Taken in sum, for example: "Hey computer, take a look at all of my past patient medical records including related imaging studies, and then please review how I treated each patient. Next, analyze what factors and treatments resulted in the best outcomes for each individual.…”

Editorial Commentary: Artificial Intelligence in Sports Medicine Diagnosis Needs to Improve

“…3,4 It appears that in highly specialized fields, such as sports medicine, with diagnostic tools such as magnetic resonance imaging scans that are more sophisticated than simple radiographs, experts overperform AI systems at present. 2,4 The key question is whether AI is ready to take over a part of the diagnostic process. The data so far have shown promise, but clearly, additional studies are necessary to show better accuracy with AI models.…”

Response to “Regarding ‘Editorial Commentary: Artificial Intelligence in Sports Medicine Diagnosis Needs to Improve’”

“…The recent letter to the editor written by Ramkumar et al 1 concerned several points raised by Dr. Nikolaos Paschos in a recent editorial commentary 2 on our systematic review entitled "Diagnostic Performance of Artificial Intelligence for Detection of Anterior Cruciate Ligament and Meniscus Tears: A Systematic Review." 3 In particular, Ramkumar et al 1 sought to clarify the distinction between the appropriate use and current interpretation of artificial intelligence (AI) on the basis of several findings in our systematic review; or, in other words, between what AI is and what it is not. Although we are grateful for their well-intended defense of AI given their expertise in this growing area of research and concur with the majority of their primary disputes, we believe that the purpose and conclusion of our systematic review was misconstrued to align with their intended message.…”