Deep-learning-based diagnosis for nasal morphological abnormalities

In recent years, diagnostic support system using medical images has significantly advanced, particularly in the automatic detection of highly prevalent diseases such as colorectal cancer using endoscopic imaging. However, the application of such technologies to nasal cavity disorders in the field of otolaryngology remains limited, indicating a need for further development. This study proposes a deep learning–based system for the automatic detection of nasal morphological abnormalities. In the proposed method, nasal endoscopic videos are labeled by physicians as severe, mild, or normal. These categories are converted into continuous values and modeled using a regression-based approach, allowing the network to capture gradual morphological variations that cannot be clearly represented by discrete classification. Experimental results have shown that the proposed model achieved a sensitivity of 91% and a specificity of 83% for severe cases, demonstrating its effectiveness in detecting clinically significant nasal narrowing. This approach could serve as a foundation for supporting diagnostic support systems during nasal endoscopy, potentially assisting clinicians in objective severity assessment.