ML-Based Disease Detection Using Trained Canines
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current disease detection methods, such as biopsies, suffer from poor sensitivity and specificity for early detection, and often require invasive procedures and point-of-care visits, which are costly, inconvenient, and have low adherence rates.
Innovation Solution
The use of machine learning-based disease-detection models in conjunction with novel sample collection devices and trained detection animals, such as canines, to analyze biological samples and identify disease types and monitor disease progression.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional diagnostic tests are used, then disease detection can be performed, but the tests are expensive, invasive, and require point-of-care visits
Solution Approach 1:
The patent replaces traditional mechanical/invasive diagnostic procedures with a biological detection system. Trained detection animals (such as canines with highly sensitive olfactory receptors) detect disease-related odorants in patient breath or other samples, eliminating the need for invasive biopsies, blood draws, or complex medical equipment while maintaining high detection accuracy
Solution Approach 2:
The patent introduces trained detection animals as intermediary agents between the patient sample and the diagnostic conclusion. These animals serve as biological mediators that translate chemical signals (odorants) into detectable behavioral responses, which are then interpreted by handlers or sensors to determine disease presence
2Measurement precision
If traditional cancer screening tests are used, then cancer detection is possible, but the tests have low sensitivity and high false positive rates
Solution Approach 1:
The patent changes the detection parameter from visual/chemical analysis to olfactory detection. Detection animals possess extremely sensitive olfactory receptors that can detect volatile organic compounds (VOCs) at concentrations far below what traditional instruments can measure, enabling detection of early-stage cancer with high sensitivity and specificity
Solution Approach 2:
The patent employs multiple detection animals working in parallel, analogous to using multiple sensors or detection methods. By aggregating the detection results from several trained animals, the system achieves higher overall accuracy and reduces false positives compared to using a single detection method
3Measurement precision
If invasive biopsy procedures are used, then definitive diagnosis can be obtained, but patient adherence and willingness to undergo testing are low
Solution Approach 1:
The patent replaces invasive mechanical procedures (biopsies, surgical interventions) with non-invasive olfactory detection. Patients simply provide breath samples or wear sensors that collect odorant information, eliminating pain, bleeding, and recovery time while maintaining diagnostic capability through detection of disease-specific chemical signatures
Solution Approach 2:
The patent enables patients to self-collect samples (such as breath into a collection device) at home without requiring medical professional intervention. This self-service approach dramatically increases accessibility and adherence while the trained detection animals perform the diagnostic evaluation
4Adaptability or versatility
If multiple separate screening procedures are used for different cancer types, then comprehensive cancer detection is achieved, but the testing process becomes time-consuming and complex
Solution Approach 1:
The patent creates a universal detection system where trained detection animals can identify multiple types of cancer through a single testing procedure. Different cancer types produce distinct odorant profiles that trained animals can differentiate, allowing one animal or pack of animals to screen for various cancers simultaneously rather than requiring separate specialized tests for each cancer type
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach provides high sensitivity and specificity, reduces the need for invasive procedures, and enables efficient, non-invasive, and cost-effective disease detection, improving patient adherence and outcomes.
Implementation Method 1
Canines have extremely sensitive olfactory receptors and are able to detect many scents that a human cannot. Canines can pick out specific scent molecules in the air, even at low concentrations.
Implementation Method 2
cancers produce volatile organic compounds (VOCs) which are excreted into the blood, sweat, saliva, urine, and breath of people with cancer. VOCs are a crucial, early indication of cancer.
Data Source
AI summary
Described herein are systems for disease detection from a biological sample using a machine learning-based (ML-based) disease-detection model trained on a dataset of detection events. Also described are methods for detecting a disease category from a biological sample received from a subject, and further detecting the specific disease type within the disease category using the systems and ML-based disease detection models. Also described are methods for monitoring progression of a disease in a subject by analyzing a biological sample using the systems and ML-based disease-detection model disclosed herein.