Application of Artificial Intelligence and its Subsets in Various Stages of Knee Arthroplasty from Pre-op to Post-op: An Overview

Jacqueline B Anderson Enni, Devendra K Agrawal*

Department of Translational Research, College of Osteopathic Medicine of the Pacific, Western University of Health Sciences, Pomona, California, USA

^*Corresponding Author: Devendra K Agrawal, Department of Translational Research, College of Osteopathic Medicine of the Pacific, Western University of Health Sciences, Pomona, California, USA.

Received: 06 February 2025; Accepted: 19 February 2025; Published: 11 March 2025

1. Introduction

Osteoarthritis is a degenerative disease that involves destruction of joint articular cartilage, mainly a breakdown of type II collagen and proteoglycans. This degradation, as well as adverse remodeling of subchondral bone, is mediated and propagated by inflammatory cytokines and other mediators [1-4]. Risk factors for the development and diagnosis of osteoarthritis include aging, genetics, obesity, vitamin D deficiency, and trauma [5-9]. In 2020, the estimated worldwide incidence of osteoarthritis of the knee, specifically, was 86.7 million people [10].

There are several therapeutic approaches in the treatment of osteoarthritis [11]. Surgical interventions, namely arthroplasties, are a common and effective treatment for knee osteoarthritis to reduce pain and improve function [12-15]. Heckmann et. al. [16] estimates that 645,852 elective knee arthroplasties were performed in the United States from 2017-2019 and this number is projected to reach 3.48 million annually by 2030 [17]. According to the American Joint Replacement Registry 2023 Annual Report, the mean age of 1.8 million people who underwent knee arthroplasties from 2012-2022 was 67.4 years [18]. Both the diagnosis of knee osteoarthritis and intervention via arthroplasty are more common in females [2,19].

The last 30 years has seen considerable expansion in the use of technology in Orthopedic Surgery and knee arthroplasties, primarily related to robotics and computer-assisted surgeries [20]. Artificial intelligence (AI) is a technology that is commonly used along-side computer-assisted surgery but also has solo applications in the field of joint reconstructive surgery. AI uses pattern recognition and algorithms to provide classifications, predictions, and solutions to tasks with a goal of greater efficiency than humans. It is an umbrella term that encompasses subsets such as machine learning (ML) and deep learning (DL). Though knee arthroplasty is an increasingly common procedure, it is associated with negative outcomes such as patient dissatisfaction, surgical complications, limited functionality, and revision surgery [21]. AI in knee arthroplasty has the potential to streamline operations, limit negative outcomes and improve patient quality of life as the number of patients undergoing this procedure continues to grow.

Much of current literature reviews robotic-assisted knee arthroplasty or the use of AI on specific aspects of the procedure from analyzing diagnostic imaging to predicting outcomes. Moreover, existing reviews discuss AI and ML but tend to not mention the additional subset of DL. The objective of this article is to provide an overview of current research on the application of AI and its subsets in each stage of knee arthroplasties.

2. Methods

An initial search on PubMed revealed ML and DL as common MeSH terms and keywords used alongside AI in knee arthroplasties and were included in this review. The following search terms and Boolean operators were used to search PubMed for studies relevant to the current application of AI in knee arthroplasties: knee arthroplasty, knee replacement, total knee arthroplasty, artificial intelligence, machine learning, and deep learning. We did not use the terms review, systematic review, meta-analysis, and editorial, and accordingly did not gather information from such articles. PubMed filters were used to ensure that only studies published in English from 2019-2024 were included. Research involving AI in all types of knee arthroplasty procedures, such as partial knee arthroplasties, total knee arthroplasties (TKA), and revisions, were included. Articles were manually reviewed and commentaries, bibliometric studies, letters to the editor, and papers irrelevant to the research question were excluded. Articles discussing computer-assisted surgery, robotic-assisted surgery, and the use of statistical analyses such as linear regression without explicit statement of use of ML, DL, or AI in knee arthroplasty were also excluded. Data regarding the article title and year published were extracted and organized via spreadsheet. Additionally, the content of each article was analyzed for the levels of evidence, methods, and key findings. Studies were categorized as AI relating to preoperative, perioperative or postoperative knee arthroplasty. Preoperative articles were further specified into predicting and assessing the need for knee arthroplasty or pre-operative patient education and decision making. Postoperative articles were further specified into predicting and assessing adverse outcomes or other post-operative outcomes. A representative selection of articles with the highest level of evidence in each of the three categories has been summarized and presented in Table 1-3.

3. Results

A search in the last 5 years of the defined terms in PubMed produced 423 articles published in English. 182 were deemed relevant based on the research question of this review and pre-defined inclusion and exclusion criteria. 35/182 (19.23%) studies examined the use of AI prior to knee arthroplasty procedure with 21 of these articles discussing AI in predicting the incidence and risk of knee arthroplasty and 14 analyzing AI as a tool for pre-operative decision making and patient education. 42/182 (23.08%) of articles discussed perioperative uses of AI, including how AI can directly impact the execution of surgical procedures (Figure 1). Lastly, 105/182 (57.69%) of research regarding AI in knee arthroplasties considers the prediction of post-surgical outcomes with 63 specifically evaluating risk assessments and the potential for adverse effects.

Table 1: The purpose and the key findings from pre-operative studies.

Table 2: The purpose and the key findings from the peri-operative studies.

Table 3: The purpose and the key findings from the post-operative studies.

Figure 1: Pre-operative, peri-operative, and post-operative clinical applications of AI and involved procedures and steps in both primary TKA and revision TKA. DVT, deep venous thrombosis.

4. Discussion

This review identified 182 research studies published from 2019-2024 that analyze the use of AI, ML, and DL in the preoperative, perioperative, or postoperative phases of knee arthroplasties. The decision to cover explicit uses of AI and its subsets in knee arthroplasty, especially with regards to computer and robotic assisted surgeries, is a strength of this review. This provides a perspective specific to technology that can gather knowledge from patterns of data to complete tasks and make predictions and can be used for more than just intraoperative robotic assistance.

The use of AI and its subsets in the preoperative stage has potential implications for the accessibility of information and provides the potential to standardize patient education and understanding of the knee arthroplasty procedure. It is of note that of the 12/14 studies included in this review which analyze AI as a tool for pre-operative decision making and patient education were published in 2023-2024. This highlights a more recent direction of research focused on investigating patient interaction with AI, particularly the AI large language model ChatGPT. However, the emergence of AI and its active role in guided decision making raises multiple ethical and privacy concerns [37,38]. Patient and provider mistrust of AI may be further explained by the term ‘AI Black Box,’ which refers to the lack of transparency and explanation for output models that presents a difficulty for humans to interpret and explain [39].

Commonly, radiographic analysis is used to diagnose the severity of osteoarthritis, which is one of the factors used to determine the need for treatment via knee arthroplasty [40,41]. Manual analysis of diagnostic imaging leaves room for potential human error. Therefore, AI can provide a more objective method for diagnosing severe osteoarthritis and using patient factors to identify those at risk for future knee arthroplasty to offer improved quality of life to patients affected by osteoarthritis.

The critical review of multiple studies identified perioperative and intraoperative use of AI as the second greatest research field of AI application in knee arthroplasty with 42/182 (22.08%) of articles. This highlights the direct effect that AI can have on surgical planning and implementation. Moreover, AI has intraoperative uses outside of robotic-assisted surgery, primarily with regards to determining proper implant sizes and soft tissue balancing.

Another strength of this review is the quantification of the scope of AI use in the stages of all types of knee arthroplasty procedures, including partial, total, and revision arthroplasties. With the growing projected prevalence of knee arthroplasties also comes the projected increase in arthroplasty revision [17]. Revision knee arthroplasties are commonly due to factors such as component loosening or periprosthetic joint infection [42,43]. In addition to revisions, primary knee arthroplasties are also subject to adverse outcomes such as cardiovascular and respiratory complications and bleeding [44]. This findings in this article indicate that the majority of AI and knee arthroplasty -related research (105/182 articles) consists of using AI to perform risk assessments and predict overall treatment outcomes. This finding supports the necessity to improve discharge planning, patient reported outcome measures, and additional factors to advance outcomes associated with knee arthroplasty.

5. Limitations and Future Directions

This information in this article is limited in the fact that grouping of articles was performed by a single reviewer, which could present a perceived bias and subjective categorization despite objective evaluation of each published finding. This study also does not discuss an in-depth analysis of the quality of articles included. An overview of the types of studies and level of investigation of AI and knee arthroplasty related research as well as the data used to train AI, ML and DL programs (electronic medical records, diagnostic imaging) would be of interest. Moreover, future investigation into the scope research comparing the efficacy between types of AI and statistical analysis such as logistic regression could be beneficial to the field.

Ultimately, AI is an emerging resource and has various applications through every stage of knee arthroplasties. Research is currently being conducted to assess the use of AI and its subsets ML and DL to improve patient experience, clinical practice, and surgical assistance.

6. Key Points

• • Artificial intelligence (AI), including machine learning (ML) and deep learning (DL), has applications in the preoperative, perioperative, and postoperative stages of knee arthroplasties, improving diagnostics, surgical precision, and patient outcomes.
• • AI is increasingly used for patient education and decision-making before surgery. Recent studies (2023-2024) focus on the role of AI in assisting patients with surgical decisions, including using ChatGPT as an informational tool.
• • The use of AI in perioperative settings extends beyond robotic-assisted surgery, significantly enhancing implant positioning, soft tissue balancing, and intraoperative navigation.
• • AI applications in knee arthroplasty are rapidly expanding, with over 57% of recent studies focusing on postoperative outcome prediction and risk assessment.
• • Despite growing integration of AI in knee arthroplasty, concerns over patient trust, data privacy, and ethical implications remain key barriers to widespread clinical adoption.
• • While AI applications in knee arthroplasty are expanding, further research is needed to compare different AI models, validate their efficacy against traditional statistical approaches, and establish standardized implementation protocols.

Funding: The research work of DKA is supported by the R25AI179582 and R01 HL147662 grants from the National Institutes of Health, USA. The contents of this article are solely the responsibility of the authors and do not necessarily represent the official views of the National Institutes of Health.

Competing interests: Both authors have read the manuscript and declare no conflict of interest. No writing assistance was utilized in the production of this manuscript.

Consent for publication: Both authors have read the manuscript and consented for publication.

References

1. Giorgino R, Albano D, Fusco S, et al. Knee Osteoarthritis: Epidemiology, Pathogenesis, and Mesenchymal Stem Cells: What Else Is New? An Update. Int J Mol Sci 24 (2023).
2. Rosenberg JH, Rai V, Dilisio MF, et al. Damage-associated molecular patterns in the pathogenesis of osteoarthritis: potentially novel therapeutic targets. Mol Cell Biochem 434 (2017): 171-179.
3. Rai V, Dilisio MF, Samadi F, et al. Counteractive Effects of IL-33 and IL-37 on Inflammation in Osteoarthritis. Int J Environ Res Public Health 19 (2022): 5690.
4. Rosenberg JH, Rai V, Dilisio MF, et al. Increased expression of damage-associated molecular patterns (DAMPs) in osteoarthritis of human knee joint compared to hip joint. Mol Cell Biochem 436 (2017): 59-69.
5. Katz JN, Arant KR, Loeser RF. Diagnosis and Treatment of Hip and Knee Osteoarthritis: A Review. JAMA 325 (2021): 568-578.
6. Allen KD, Thoma LM, Golightly YM. Epidemiology of osteoarthritis. Osteoarthritis Cartilage 30 (2022): 184-195.
7. Cui A, Li H, Wang D, et al. Global, regional prevalence, incidence and risk factors of knee osteoarthritis in population-based studies. EClinicalMedicine (2020):100587.
8. Garfinkel RJ, Dilisio MF, Agrawal DK. Vitamin D and Its Effects on Articular Cartilage and Osteoarthritis. Orthop J Sports Med 5 (2017): 2325967117711376.
9. Rai V, Radwan MM, Agrawal DK. IL-33, IL-37, and Vitamin D Interaction Mediate Immunomodulation of Inflammation in Degenerating Cartilage. Antibodies (Basel) 10 (2021): 41.
10. Rai V, Dietz NE, Dilisio MF, et al. Vitamin D attenuates inflammation, fatty infiltration, and cartilage loss in the knee of hyperlipidemic microswine. Arthritis Res Ther 18 (2016): 203.
11. Rajalekshmi R, Agrawal DK. Therapeutic Efficacy of Medicinal Plants with Allopathic Medicine in Musculoskeletal Diseases. Int J Plant Anim Environ Sci 14 (2024): 104-129.
12. Rönn K, Reischl N, Gautier E, et al. Current surgical treatment of knee osteoarthritis. Arthritis 2011 (2011): 454873.
13. Rai V, Dilisio MF, Dietz NE, et al. Recent strategies in cartilage repair: A systemic review of the scaffold development and tissue engineering. J Biomed Mater Res A 105 (2017): 2343-2354.
14. Steinhaus ME, Christ AB, Cross MB. Total Knee Arthroplasty for Knee Osteoarthritis: Support for a Foregone Conclusion? Hss j 13 (2017): 207-210.
15. Madry H. Surgical therapy in osteoarthritis. Osteoarthritis Cartilage 30 (2022): 1019-1034.
16. Heckmann ND, Bouz GJ, Piple AS, et al. Elective Inpatient Total Joint Arthroplasty Case Volume in the United States in 2020: Effects of the COVID-19 Pandemic. J Bone Joint Surg Am 104 (2022): e56.
17. Kurtz S, Ong K, Lau E, et al. Projections of primary and revision hip and knee arthroplasty in the United States from 2005 to 2030. J Bone Joint Surg Am 89 (2007): 780-5.
18. Ryan SP, Stambough JB, Huddleston JI, et al. Highlights of the 2023 American Joint Replacement Registry Annual Report. Arthroplast Today 26 (2024): 101325.
19. Nham FH, Patel I, Zalikha AK, et al. Epidemiology of primary and revision total knee arthroplasty: analysis of demographics, comorbidities and outcomes from the national inpatient sample. Arthroplasty 5 (2023): 18.
20. Rullán PJ, Grits D, Potluri A, et al. Identifying Trends and Quantifying Growth for Technological Innovation in Knee Arthroplasty: An Analysis of a Patent Database (1990 to 2020). J Knee Surg 36 (2023): 1209-1217.
21. DeFrance MJ, Scuderi GR. Are 20% of Patients Actually Dissatisfied Following Total Knee Arthroplasty? A Systematic Review of the Literature. J Arthroplasty 38 (2023): 594-599.
22. Kastrup N, Bjerregaard HH, Laursen M, et al. An AI-based patient-specific clinical decision support system for OA patients choosing surgery or not: study protocol for a single-centre, parallel-group, non-inferiority randomised controlled trial. Trials 24 (2023): 24.
23. Jayakumar P, Moore MG, Furlough KA, et al. Comparison of an Artificial Intelligence-Enabled Patient Decision Aid vs Educational Material on Decision Quality, Shared Decision-Making, Patient Experience, and Functional Outcomes in Adults With Knee Osteoarthritis: A Randomized Clinical Trial. JAMA Netw Open 4 (2021): e2037107.
24. Zhou Y, Weeden C, Patten L, et al. Evaluating willingness for surgery using the SMART Choice (Knee) patient prognostic tool for total knee arthroplasty: study protocol for a pragmatic randomised controlled trial. BMC Musculoskelet Disord 23 (2022): 179.
25. Mahmoud K, Alagha MA, Nowinka Z, et al. Predicting total knee replacement at 2 and 5 years in osteoarthritis patients using machine learning. BMJ Surg Interv Health Technol 5 (2023): e000141.
26. Mohajer B, Dolatshahi M, Moradi K, et al. Role of Thigh Muscle Changes in Knee Osteoarthritis Outcomes: Osteoarthritis Initiative Data. Radiology 305 (2022): 169-178.
27. Al-Nasser S, Noroozi S, Harvey A, et al. Exploring the Performance of an Artificial Intelligence-Based Load Sensor for Total Knee Replacements. Sensors (Basel) 24 (2024).
28. Arunachalam VNK. Deep Learning-based Thigh Muscle Investigation Using MRI For Prosthetic Development for Patients Undergoing Total Knee Replacement (TKR). Curr Med Imaging 20 (2024): e15734056284002.
29. Factor S, Gurel R, Dan D, et al. Validating a Novel 2D to 3D Knee Reconstruction Method on Preoperative Total Knee Arthroplasty Patient Anatomies. J Clin Med 13 (2024).
30. Rodrigues P, Antunes M, Raposo C, et al. Deep segmentation leverages geometric pose estimation in computer-aided total knee arthroplasty. Healthc Technol Lett 6 (2019): 226-230.
31. Fernandes LR, Arce C, Martinho G, et al. Accuracy, Reliability, and Repeatability of a Novel Artificial Intelligence Algorithm Converting Two-Dimensional Radiographs to Three-Dimensional Bone Models for Total Knee Arthroplasty. J Arthroplasty 38 (2023): 2032-2036.
32. Langenberger B, Steinbeck V, Schöner L, et al. Exploring treatment effect heterogeneity of a PROMs alert intervention in knee and hip arthroplasty patients: A causal forest application. Comput Biol Med. Sep 163 (2023): 107118.
33. Harris AHS, Kuo AC, Bowe TR, et al. Can Machine Learning Methods Produce Accurate and Easy-to-Use Preoperative Prediction Models of One-Year Improvements in Pain and Functioning After Knee Arthroplasty? J Arthroplasty 36 (2021): 112-117.e6.
34. Ramkumar PN, Haeberle HS, Ramanathan D, et al. Remote Patient Monitoring Using Mobile Health for Total Knee Arthroplasty: Validation of a Wearable and Machine Learning-Based Surveillance Platform. J Arthroplasty 34 (2019): 2253-2259.
35. Sachau J, Otto JC, Kirchhofer V, et al. Development of a bedside tool-kit for assessing sensitization in patients with chronic osteoarthritis knee pain or chronic knee pain after total knee replacement. Pain 163 (2022): 308-318.
36. Bains SS, Dubin JA, Hameed D, et al. Use and Application of Large Language Models for Patient Questions Following Total Knee Arthroplasty. J Arthroplasty 39 (2024): 2289-2294.
37. Kaya Bicer E, Fangerau H, Sur H. Artifical intelligence use in orthopedics: an ethical point of view. EFORT Open Rev 8 (2023): 592-596.
38. Giorgino R, Alessandri-Bonetti M, Luca A, et al. ChatGPT in orthopedics: a narrative review exploring the potential of artificial intelligence in orthopedic practice. Front Surg 10 (2023): 1284015.
39. Bullock GS, Hughes T, Arundale AH, et al. Black Box Prediction Methods in Sports Medicine Deserve a Red Card for Reckless Practice: A Change of Tactics is Needed to Advance Athlete Care. Sports Med 52 (2022): 1729-1735.
40. Chang CB, Yoo JH, Koh IJ, et al. Key factors in determining surgical timing of total knee arthroplasty in osteoarthritic patients: age, radiographic severity, and symptomatic severity. J Orthop Traumatol 11 (2010): 21-7.
41. Khalifa AA, Mullaji AB, Mostafa AM, et al. A Protocol to Systematic Radiographic Assessment of Primary Total Knee Arthroplasty. Orthop Res Rev 13 (2021): 95-106.
42. Dubin JA, Bains SS, Chen Z, et al. Using a Google Web Search Analysis to Assess the Utility of ChatGPT in Total Joint Arthroplasty. J Arthroplasty 8 (2023): 1195-1202.
43. Na BR, Kwak WK, Lee NH, et al. Trend Shift in the Cause of Revision Total Knee Arthroplasty over 17 Years. Clin Orthop Surg 15 (2023): 219-226.

Healy WL, Della Valle CJ, Iorio R, et al. Complications of total knee arthroplasty: standardized list and definitions of the Knee Society. Clin Orthop Relat Res 471 (2013): 215-20.