Methods for Detection and Identification of Bacterial Species in Blood Cultures based on Optical Turbidimetry and AI for use in Limited-Resource Settings

Project Details

Layman's description

In the past years Ghent University (UGent) and the Institute for Tropical Medicine (ITM) have jointly developed a method and prototypes for the detection of bacterial infection in resource-limited settings. The method is based on applying optical turbidimetry to blood cultures. Several generations of turbidimeter prototypes have been built and are now being evaluated in the field in two African countries. The lab results (based on experiments in ITM) are generally satisfying, but there are currently two limitations: 1. Detection of infection is possible on several bacterial species but not for all types; 2. Discrimination between different species has so far not been studied but would bring considerable added value for choice of therapeutic treatment. The hypothesis for the proposed PhD-project is that the spectral turbidimetric data hold the promise to detect a broader range of species and also to discriminate between several species. The objective of the project is to test this hypothesis and to come up with modifications of the current prototypes to optimize the sensitivity and selectivity of the turbidimetry as well as to use more refined data extraction methods, likely based on deep learning. Initial ideas include the use of optimal light paths through the culture bottle and more refined spectral measurements. It may also be considered to complement the turbidimetric approach with entirely different sensing methods that are also suitable for use in a resource-limited setting.
Effective start/end date12/12/23 → …

IWETO expertise domain

  • B780-tropical-medicine