The aim of this study was to evaluate the ability of a machine-learning algorithm to diagnose prosthetic loosening from preoperative radiographs and to investigate the inputs that might improve its performance. A group of 697 patients underwent a first-time revision of a total hip (THA) or total knee arthroplasty (TKA) at our institution between 2012 and 2018. Preoperative anteroposterior (AP) and lateral radiographs, and historical and comorbidity information were collected from their electronic records. Each patient was defined as having loose or fixed components based on the operation notes. We trained a series of convolutional neural network (CNN) models to predict a diagnosis of loosening at the time of surgery from the preoperative radiographs. We then added historical data about the patients to the best performing model to create a final model and tested it on an independent dataset.Aims
Methods
Radiological assessment of total and unicompartmental
knee replacement remains an essential part of routine care and follow-up.
Appreciation of the various measurements that can be identified
radiologically is important. It is likely that routine plain radiographs
will continue to be used, although there has been a trend towards
using newer technologies such as CT, especially in a failing knee,
where it provides more detailed information, albeit with a higher
radiation exposure. The purpose of this paper is to outline the radiological parameters
used to evaluate knee replacements, describe how these are measured
or classified, and review the current literature to determine their
efficacy where possible.
