Recognizing lung cancer and stages using a self-developed electronic nose system

Area of Science:

• Biomarkers
• Respiratory Medicine
• Analytical Chemistry

Background:

• Exhaled breath contains volatile organic compounds (VOCs) with potential as non-invasive lung cancer biomarkers.
• Breath-based screening offers a convenient, low-cost, and easily popularized method for lung cancer detection.

Purpose of the Study:

• To develop and evaluate a self-developed electronic nose (e-nose) system for lung cancer detection and staging.
• To assess the performance of the e-nose in distinguishing lung cancer patients from healthy individuals.

Main Methods:

• Utilized an electronic nose (e-nose) system analyzing 235 breath samples.
• Employed kernel principal component analysis (KPCA) for feature extraction and extreme gradient boosting (XGBoost) for classification.
• Investigated the system's ability to classify lung cancer stages (III and IV) in 90 patients.

Main Results:

• The KPCA-XGBoost model achieved 93.59% accuracy, 95.60% sensitivity, and 91.09% specificity in distinguishing lung cancer patients.
• The e-nose system demonstrated over 80% accuracy in classifying lung cancer stages.
• The system successfully differentiated lung cancer, smoking, and other respiratory diseases based on breath characteristics.

Conclusions:

• The developed e-nose system, combined with advanced pattern recognition algorithms, shows significant potential for non-invasive lung cancer diagnosis and staging.
• This approach can identify unique expiratory VOC profiles associated with lung cancer, smoking, and other respiratory conditions.