The Use of Artificial Neural Networks for the Prediction of Surgical Site Infection Following TKA

Abstract

This is a retrospective study. Surgical site infection (SSI) is associated with adverse postoperative outcomes following total knee arthroplasty (TKA). However, accurately predicting SSI remains a clinical challenge due to the multitude of patient and surgical factors associated with SSI. This study aimed to develop and validate machine learning models for the prediction of SSI following primary TKA. This is a retrospective study for patients who underwent primary TKA. Chart review was performed to identify patients with superficial or deep SSIs, defined in concordance with the criteria of the Musculoskeletal Infection Society. All patients had a minimum follow-up of 2 years (range: 2.1–4.7 years). Five machine learning algorithms were developed to predict this outcome, and model assessment was performed by discrimination, calibration, and decision curve analysis. A total of 10,021 consecutive primary TKA patients was included in this study. At an average follow-up of 2.8 ± 1.1 years, SSIs were reported in 404 (4.0%) TKA patients, including 223 superficial SSIs and 181 deep SSIs. The neural network model achieved the best performance across discrimination (area under the receiver operating characteristic curve = 0.84), calibration, and decision curve analysis. The strongest predictors of the occurrence of SSI following primary TKA, in order, were Charlson comorbidity index, obesity (BMI >30 kg/m2), and smoking. The neural network model presented in this study represents an accurate method to predict patient-specific superficial and deep SSIs following primary TKA, which may be employed to assist in clinical decision-making to optimize outcomes in at-risk patients.

Predicting Factors for Blood Transfusion in Primary Total Knee Arthroplasty Using a Machine Learning Method

“…Machine learning (ML) is a form of statistical artificial intelligence that has been used to predict risk factors for transfusion rates in THA, prolonged opioid usage in THA, and surgical site infection in TKA. 6,15,16 ML uses computational algorithms to predict outcomes. 17,18 These algorithms use large data sets to recognize notable patterns and learn from a given data set.…”

Prediction of Lumbar Drainage-Related Meningitis Based on Supervised Machine Learning Algorithms

“…At present, scholars have an increasing enthusiasm for utilizing supervised ML to predict the occurrence of infection, including iatrogenic and non-iatrogenic (21,(25)(26)(27)(28)(29)(30)(31)(32)(33). Nonetheless, the study that forecasts the risk of LDRM in an early stage before the clinical diagnosis is limited, although we have proposed a prediction model using the traditional logistic regression algorithm (19).…”

Enabling Personalized Medicine in Orthopaedic Surgery Through Artificial Intelligence

“…Several AI models have been developed on data sets used to predict orthopaedic-specific postoperative complications 66 . In the future, these models 66,102–127 (Table IV) may serve an essential role in determining patient eligibility for orthopaedic surgical procedures. Although studies have reported excellent model performance, their level of evidence remains limited 116,117 .…”

“…Overall, the models identified young patient age and greater patient BMI, along with the surgeon, surgical team, and hospital infrastructure as some of the most important factors affecting operation time 154,155 . Utilization of this technology in the hospital setting could lead to 104 THA Identification of patients at risk of prolonged LOS after THA Klemt et al 114 rTHA Prediction of re-revision surgery after rTHA (prediction of failure of rTHA) Ramkumar et al 125 THA Prediction of LOS and cost of inpatient stay after THA Klemt et al 115 THA Prediction of risk for revision surgery after THA Khosravi et al 113 THA Prediction of patient-specific risk of dislocation following THA Lazic et al 120 THA Prediction of THA surgery duration and intraoperative and postoperative complications Haeberle et al 106 Hip arthroscopy Prediction of the need for subsequent surgery after primary hip arthroscopy for FAIS Polus et al 124 THA Prediction of fall risk in postoperative patients with THA Knee Anis et al 66 TKA, PKA Accurate estimation of the likelihood of improved pain, function, QOL, LOS, and 90day readmission after knee arthroplasty Klemt et al 118 rTKA Prediction of periprosthetic joint infection after rTKA Klemt et al 116 rTKA Prediction of prolonged LOS after rTKA Klemt et al 117 rTKA Prediction of nonhome discharge disposition after rTKA Hinterwimmer et al 108 TKA Prediction of surgery duration and complications after primary TKA Yeo et al 127 TKA Prediction of surgical site infection after primary TKA Shoulder Oeding et al 123 RSA Prediction of risk for postoperative dislocation resulting in readmission within 90 days of RSA Lopez et al 121 TSA Prediction of nonhome discharge after TSA Arvind et al 103 TSA Prediction of postoperative readmission for patients undergoing TSA Spine Khazanchi et al 112 ACDF Application and prediction of postoperative health care utilization after ACDF Arvind et al 102 ACDF Prediction of postoperative complications after ACDF Hopkins et al 109 PSF Prediction of patients who would not require readmission after PSF surgery Kuris et al 119 Lumbar arthrodesis Prediction of 30-day readmission after anterior, posterior, and posterior interbody lumbar spinal fusion Harada et al 107 Lumbar microdiscectomy Prediction of recurrent herniated nucleus pulposus after lumbar microdiscectomy Scheer et al…”