Detection of bone infections using Vis-NIR and SWIR hyperspectral imaging coupled with machine learning.

Bone infections, such as fracture-related and periprosthetic joint infections, present significant diagnostic and therapeutic challenges in orthopaedic surgery. Current diagnosic standards rely primarily on tissue cultures of intraoperatively obtained samples - a time-consuming approach with limited sensitivity and specificity and delayed clinical decision-making. This study investigates the use of hyperspectral imaging (HSI) in the visible and near-infrared (Vis-NIR), and short-wave infrared (SWIR) spectral ranges for the rapid detection of bone infections. Using ex vivo human bone samples, an in vitro biofilm model was established with Staphylococcus aureus and Staphylococcus epidermidis. Spectral data were analyzed using machine learning algorithms, including k-nearest neighbors (kNN), support vector machine (SVM), partial least squares discriminant analysis (PLS-DA), and soft independent modeling of class analogy (SIMCA). Vis-NIR-HSI models outperformed SWIR-based classification, achieving classification accuries of up to 99.58 % for distinguishing inoculated from uninoculated human bone samples, and enabling accurate bacterial species differentiation. These results highlight the diagnostic potential of Vis-NIR-HSI as real-time, label-free intraoperative tool for bone infection detection, bridging the gap between preoperative imaging and delayed microbiological results, and supporting immediate surgical decision making.