Search

Previous studies have shown that the manufacturer's default preoperative plans for total knee arthroplasty with patient-specific guides require frequent, time-consuming changes by the surgeon. Currently, no research has been done on predicting preoperative plans for orthopedic surgery using machine learning. Therefore, this study aims to evaluate whether artificial intelligence (AI) driven planning tools can create surgeon and patient-specific preoperative plans that require fewer changes by the surgeon. A dataset of 5409 preoperative plans, including the manufacturer's default and the plans corrected by 39 surgeons, was collected. Features were extracted from the preoperative plans that describe the implant sizes, position, and orientation in a surgeon- and patient-specific manner. Based on these features, non-linear regression models were employed to predict the surgeon's corrected preoperative plan. The average number of corrections a surgeon has to make to the preoperative plan generated using AI was reduced by 39.7% compared to the manufacturer's default plan. The femoral and tibial implant size in the manufacturer's plan was correct in 68.4% and 73.1% of the cases, respectively, while the AI-based plan was correct in 82.2% and 85.0% of the cases, respectively, compared to the surgeon approved plan. Our method successfully demonstrated the use of machine learning to create preoperative plans in a surgeon- and patient-specific manner for total knee arthroplasty. ispartof: FRONTIERS IN ROBOTICS AND AI vol:9 ispartof: location:Switzerland status: published

A preoperative plan is a virtual plan that defines the implant position and orientation allowing a surgeon to prepare for surgery. However, the default preoperative plan for total knee arthroplasty proposed by the manufacturer requires changes to be made by the surgeon in more than 90% of the cases. Previous studies have shown that artificial intelligence can be used to create better preoperative plans compared to manufacturer’s default plans. However, the quality of artificial intelligence based preoperative plans has not yet been compared to surgeon approved preoperative plans. The purpose of this study is to compare default, artificial intelligence and surgeon approved preoperative plans, by having them scored on a range from 1 (totally unacceptable plan) to 5 (no corrections needed) by an experienced surgeon, while being blinded to the plan type. Through a Wilcoxon signed rank test with α=0.05, AI based preoperative plans were found to be a significant improvement upon the default plans (p-val=0.000136), while the differences in score between AI and surgeon approved preoperative plans were insignificant (p-val= 0.083). Consequently these results indicate that AI generated preoperative plans for TKA are an improvement upon current default plans, which could increase the surgeon’s planning efficiency when applied in clinical practice.

An experimental investigation was conducted into the heat transfer characteristics during in-tube condensation of horizontal smooth, micro-fin, and herringbone tubes. The study focused on the heat transfer coefficients of refrigerants R-22, R-134a, and R-407C inside a series of typical horizontal smooth, micro-fin, and herringbone tubes at a representative average saturation temperature of 40°C. Mass fluxes ranged from 300 to 800kg∕m2s, and vapor qualities ranged from 0.85 to 0.95 at condenser inlet, to 0.05 to 0.15 at condenser outlet. The herringbone tube results were compared with the smooth and micro-fin tube results. The average increase in the heat transfer coefficient of the herringbone tube, when compared with the smooth tube at comparable conditions, was found to be 322%, with maximum values reaching 336%. When compared with the micro-fin tube, the average increase in heat transfer coefficient was found to be 196%, with maximum values reaching 215%. Moreover, a new correlation was developed to predict the heat transfer coefficients in a herringbone and micro-fin tube. Semi-local heat transfer coefficients were calculated from the modified Wilson plot technique, using measurements of condenser subsection inlets and outlets, and from knowledge of the temperature gradient on the annulus side. The correlation predicted the semi-local heat transfer coefficients accurately, with 96% and 89% of the data points falling in the ±20% region for the herringbone tube and the micro-fin tube, respectively. The average heat transfer coefficients were accurately predicted, too, with all the data points for the herringbone tube and 83% of the data points for the micro-fin tube falling in the ±20% region. The derived heat transfer correlations can be used for design, especially for reversible heat pumps. This research proves that predicting the flow pattern during intermittent and annular flow is not a prerequisite for predicting the heat transfer accurately to within 20% of the measurements.