Digitized canine olfaction and multimodal biosensing for breath-based multi-cancer detection: a hypothesis-driven...

Area of Science:

• Biotechnology
• Computational Biology
• Veterinary Medicine

Background:

• Current Multi-Cancer Early Detection (MCED) methods face challenges in cost, accessibility, and sensitivity.
• Breath analysis of volatile organic compounds (VOCs) presents a promising non-invasive screening approach.
• Trained dogs show high olfactory sensitivity for disease detection, but lack standardization.

Purpose of the Study:

• To propose a novel canine olfaction-based MCED platform integrating multiple biomarkers.
• To leverage machine learning for analyzing dog's EEG, behavioral, and physiological data.
• To establish a pathway for reproducible, portable, and scalable cancer detection.

Main Methods:

• Utilizing electroencephalography (EEG) signals from trained dogs to analyze olfactory recognition.
• Combining EEG data with behavioral (vision tracking) and physiological (respiration, heart rate) biomarkers.
• Developing a machine learning model to integrate multimodal data and generate a cancer risk score.

Main Results:

• The proposed Dognosis platform integrates bio-digital data for cancer risk assessment.
• Preliminary validation strategies and a phased experimental roadmap are outlined.
• The framework aims for reproducible, portable, scalable, and low-cost cancer detection.

Conclusions:

• The bio-digital canine olfaction platform offers a novel approach to MCED.
• This system has potential for deployment in low-resource settings and broader disease detection.
• Standardization and reproducibility are key advancements for canine scent detection in diagnostics.