Artificially intelligent nasal perception for rapid sepsis diagnostics

Area of Science:

• Biomedical Engineering
• Clinical Diagnostics
• Artificial Intelligence

Background:

• Sepsis is a life-threatening condition with high mortality, necessitating rapid diagnosis.
• Conventional methods like bacterial cultures are slow, delaying critical treatment.
• Detecting bacterial volatile organic compounds (VOCs) offers a promising alternative diagnostic approach.

Purpose of the Study:

• To develop a rapid and accurate diagnostic method for sepsis.
• To investigate the use of colorimetric gas sensor arrays for VOC detection.
• To enhance sepsis diagnostic accuracy using an AI-based algorithm.

Main Methods:

• Design of colorimetric gas sensor arrays for visual detection of sepsis-related VOCs.
• Application of an artificial intelligence algorithm, Rapid Sepsis Boosting (RSBoost), for data analysis.
• Validation of the diagnostic approach using blood samples.

Main Results:

• The developed sensor array demonstrated high sensitivity and specificity for detecting sepsis biomarkers.
• The RSBoost algorithm achieved a diagnostic accuracy of 96.2%.
• The combined approach enables sepsis diagnosis within 24 hours.

Conclusions:

• Colorimetric gas sensor arrays coupled with AI offer a significant advancement in rapid sepsis diagnostics.
• This method improves speed and accuracy, potentially reducing healthcare costs and improving patient outcomes.
• The technology holds potential for transforming clinical practice in sepsis management.