Classification of lung cancer subtypes based on autofluorescence bronchoscopic pattern recognition: A preliminary...

Area of Science:

• Pulmonology
• Medical Imaging
• Artificial Intelligence

Background:

• Lung cancer is a leading global cause of cancer mortality.
• Autofluorescent bronchoscopic imaging aids early lung cancer detection.
• Pathologic examination has limitations in diagnosing lung cancer types.

Purpose of the Study:

• To develop a computer-aided diagnosis (CAD) system for distinguishing lung cancer types.
• To enhance diagnostic objectivity and consistency in lung cancer pathology.
• To utilize autofluorescent bronchoscopy imaging for improved cancer classification.

Main Methods:

• A database of 12 adenocarcinomas and 11 squamous cell carcinomas was used.
• Autofluorescent bronchoscopic images were converted to Hue (H), Saturation (S), and Value (V) color spaces.
• Color textural features from H, S, and V channels were extracted and classified using logistic regression and machine learning.

Main Results:

• The CAD system achieved 83% accuracy in distinguishing lung cancer types.
• Sensitivity was 73% and specificity was 92% for cancer type classification.
• The area under the receiver operating characteristic curve was 0.81.

Conclusions:

• The developed CAD system provides a viable diagnostic method for malignant lung cancer types.
• Color texture analysis of autofluorescent bronchoscopic images is effective for cancer classification.
• This system supports clinical decision-making in lung cancer diagnosis.