AI-based mobile application to fight antibiotic resistance

Abstract

Antimicrobial resistance is a major global health threat and its development is promoted by antibiotic misuse. While disk diffusion antibiotic susceptibility testing (AST, also called antibiogram) is broadly used to test for antibiotic resistance in bacterial infections, it faces strong criticism because of inter-operator variability and the complexity of interpretative reading. Automatic reading systems address these issues, but are not always adapted or available to resource-limited settings. We present an artificial intelligence (AI)-based, offline smartphone application for antibiogram analysis. The application captures images with the phone’s camera, and the user is guided throughout the analysis on the same device by a user-friendly graphical interface. An embedded expert system validates the coherence of the antibiogram data and provides interpreted results. The fully automatic measurement procedure of our application’s reading system achieves an overall agreement of 90% on susceptibility categorization against a hospital-standard automatic system and 98% against manual measurement (gold standard), with reduced inter-operator variability. The application’s performance showed that the automatic reading of antibiotic resistance testing is entirely feasible on a smartphone. Moreover our application is suited for resource-limited settings, and therefore has the potential to significantly increase patients’ access to AST worldwide.

Microbiology Assessments in Critically Ill Patients

“…As a case in point, a smartphone platform to perform automated antibiogram analysis with the aid of a graphic interface and embedded system that validates results was recently shown to be successfully utilized. 101 As with currently available RDTs, any new technology necessitates well-designed large-scale efficacy, cost-effectiveness, and outcome research prior to incorporation into diagnostic flow processes. Assessments in critically ill patients would certainly benefit from host-response signatures in future as well as integration of microbiology metadata in relation to other determinants in real-time patient level dashboards.…”

Deep learning-enabled mobile application for efficient and robust herb image recognition

“…However, these studies mainly focus on a few herb categories, due to the fact that low-level property cannot handle large variations in many categories. Recently, inspired by the development of image recognition in computer vision 12 – 16 and medical areas 17 – 23 , deep learning based methods have shown great improvements in herb recognition 24 – 29 . Sun et al 24 apply convolutional neural network in herb image classification, and they also release a herb image dataset with 95 categories.…”

“…In conventional approaches differentiating herbs according to color, shape and texture, the accuracy cannot be satisfied due to their low-level appearance property 9 – 11 . Recently, inspired by the development of image recognition in computer vision 12 – 16 and medical areas 17 – 23 , deep learning based methods have shown great improvements 24 – 29 , which makes it feasible to provide candidate decisions with satisfied accuracy. Most studies have used the Deep Neural Network (DNN), which combines with powerful and expensive computation hardware to support the computation.…”

Lab-on-a-Disc for Point-of-Care Infection Diagnostics

“…Recent advancement in rapid AST with a combination of machine learning and image processing was used for analysis without microbiology expertise. − These methods are highly suitable for fast analysis with limited resources, especially in low-to-middle income countries, where they can provide access to patients using a smartphone to fill the gap in medical facilities and data collection.…”