Machine learning to identify clinically relevantCandidayeast species

Abstract

BackgroundFungal infections, especially due toCandidaspecies, are on the rise. Multi-drug resistant organism such asCandida aurisare difficult and time consuming to identify accurately.Machine learning is increasingly being used in health care, especially in medical imaging. In this study, we evaluated the effectiveness of six convolutional neural networks (CNNs) to identify four clinically importantCandidaspecies.Materials and MethodsWet-mounted images were captured using bright field live-cell microscopy followed by separating single cells, budding cells, and cell group images which were then subjected to different machine learning algorithms (custom CNN, VGG16, ResNet50, InceptionV3, EfficientNetB0, and EfficientNetB7) to learn and predictCandidaspecies.ResultsAmong the six algorithms tested, the InceptionV3 model performed best in predictingCandidaspecies from microscopy images. All models performed poorly on raw images obtained directly from the microscope. The performance of all models increased when trained on single and budding cell images. The InceptionV3 model identified budding cells ofC. albicans, C. auris,C. glabrata(Nakaseomyces glabrata), andC. haemuloniiin 97.0%, 74.0%, 68.0%, and 66.0% cases, respectively. For single cells ofC. albicans, C. auris,C. glabrata, andC. haemuloniiInceptionV3 identified 97.0%, 73.0%, 69.0%, and 73.0% cases, respectively. The sensitivity and specificity of InceptionV3 were respectively 77.1% and 92.4%.ConclusionThis study provides proof of concept that microscopy images from wet mounted slides can be used to identifyCandidayeast species using machine learning quickly and accurately.