Project Lead(s): Mirko Zimic
Tuberculosis (TB) causes millions of deaths annually worldwide.
Cell culturing, the current gold standard for diagnosis, is slow (4–6 weeks) and may result in further progression of the disease, or further spread. Molecular techniques offer a suitable promise but they are not widely used.
The sputum liquid culture, “MODS” (microscopic-observation drug-susceptibility) assay, is a faster, cheaper and more sensitive test than other culture-based tests. It is a convenient test to diagnose TB, as well as to determine multidrug-resistance (MDR) in approximately seven days.
While developed in Peru in 2000 and now used in several countries, use of MODS is still limited by the need for local expertise to interpret the microscopic images. This prevents MODS from being used widely in remote rural areas (where it is needed most), as they lack the resources and professionals.
This project implemented a MODS-based telediagnostics system to detect tuberculosis and multidrug-resistance (MDR) in two laboratories in Peru.
The system consists of a mathematical algorithm for pattern recognition, to automatically classify and interpret digital microscopic images from MODS cultures.
An ad-hoc, web-based system was also developed to allow the remote laboratories to upload the microscopic digital images, which are processed in real time by the mathematical algorithm in less than one minute.
The project also involved nine experts in interpreting MODS cultures. This team of experts provided an independent diagnosis to the one reached by the algorithm.
In the project, each digital image corresponding to a sample was uploaded to a web-based platform and sent automatically to three different experts for assessment.
The project demonstrated the feasibility of using a MODS-based telediagnostics system to detect tuberculosis and multidrug-resistance in two laboratories in Peru.
Over the course of the study, 99.8% of images transmitted were “interpretable” and resulted in the diagnosis of 847 TB patients, of which 46 were MDR.
Using the traditional approach, the MDR patients would not have been detected one week after their first sputum sample, but instead might have been delayed approximately 10–12 months (after failing the two first-line drug courses of treatment) before being recognized as being MDR patients.
Acceptability of the system was high and it was easy to train healthcare workers to use it.
This research study confirms that it is feasible to implement TB telediagnostics, even in sites with no previous experience using the MODS assay.
The research group has received funds from the Peruvian agencies CONCYTEC (Consejo Nacional de Ciencia, Tecnología e Innovación Tecnológica) and FINCyT (Fondo para la Innovación, Ciencia y Tecnología) to develop telemedicine and telediagnostics in rural areas of the country. This funding has resulted in the development of the early/preliminary different components of the TB telediagnostics system that are being scaled to other diseases, including cervical cancer, melanoma and, more recently, pneumonia.
The team has also received funding from the Tropical Diseases Research division of the World Health Organization, to improve the intervention of TB telediagnostics in Trujillo.
Knowledge of the project was widely distributed in conference presentations.