Breaking barriers in Candida spp. detection with Electronic Noses and artificial intelligence.

Area of Science:

• Medical Diagnostics
• Microbiology
• Biosensor Technology

Background:

• Candidemia, a serious Candida bloodstream infection, is difficult to diagnose promptly.
• Current blood culture methods are slow and lack sensitivity, delaying critical treatment decisions.

Purpose of the Study:

• To develop a novel, rapid diagnostic method for Candida species identification using an Electronic Nose (E-nose).
• To improve diagnostic accuracy and reduce turnaround time for candidemia detection.

Main Methods:

• Utilized an Electronic Nose (E-nose) coupled with Time Series-based classification (Inception Time classifier).
• Analyzed volatile organic compounds from culture samples of various Candida species (C. albicans, C. kodamaea ohmeri, C. glabrata, C. haemulonii, C. parapsilosis, C. krusei).

Main Results:

• The E-nose system achieved an average accuracy of 97.46% in identifying Candida species during the test phase.
• Demonstrated promising proof-of-concept results for rapid and accurate candidemia diagnosis.

Conclusions:

• The E-nose combined with time series classification offers an efficient and reliable tool for early candidemia identification.
• This approach has the potential for cost-effective, widespread clinical application, improving patient outcomes.