In a first study of its kind, researchers in Spain and the USA have developed an AI system for reading and quantifying cryptococcal antigen (CrAg), using a smartphone. David Bermejo-Peláez and colleagues developed a smartphone-based digital system, assisted by AI, to automatically interpret CrAg LFA results and measure the antigen concentration in the strip.
The development was done by SpotLab using the Immy CrAg test and a mobile phone app called TiraSpot was trialled at the Mycology Reference Laboratory, Instituto de Salud III in Madrid.
Dr Ana Alastruey-Izquierdo of the Mycology Reference Laboratory and Global Action For Fungal Infections (GAFFI) who led the study said:
“One of GAFFI’s main goals is to improve access to diagnostics. This proof of concept shows that all lateral flow assays could be genuine point of care tests without a lot of staff training to interpret them and with automatic registration of results in a cloud database.”
Cryptococcal antigen circulates in the blood in infected patients (usually with HIV infection) at variable concentrations. The higher the antigen concentration, the more likely is meningitis to be present. Without meningitis, oral fluconazole antifungal alone can be used. If meningitis is present, a lumbar puncture is required, then amphotericin B intravenously, plus oral flucytosine as well as with-holding of antiretroviral therapy initially. TiraSpot app reads the result, uploads it to the web platform for quantification and stores the result for sharing with other healthcare providers. Critical clinical decisions can be made in a few minutes.
TiraSpot mobile phone reading enabled more accurate detection of low antigen concentrations and accurate quantification from 6 to 150ng/mL. The app is small and fast enough to run quickly on a mobile phone. The experimental work was done with 2 different makes of android phones.
Link to paper: https://lnkd.in/dT9UM7xx
Video showing operation: https://www.mdpi.com/article/10.3390/jof9020217/s1