Search papers, labs, and topics across Lattice.
This study integrates a Vision Transformer with attention mechanisms to enhance the analysis of narrow band imaging (NBI) for laryngeal cancer screening, addressing the challenges of interobserver variability and the need for well-trained clinicians. The model achieved an F1 score of 82.72% and an accuracy of 82.33%, demonstrating robust classification performance in distinguishing between benign and malignant lesions. Additionally, the approach incorporates explainability through the MedSAM segmentation method, providing clinicians with valuable pathological insights to support their decision-making.
AI-driven laryngeal cancer screening achieves over 82% accuracy while offering explainable insights for clinicians, transforming diagnostic practices.
Early and timely screening of laryngeal cancer is crucial for improving clinical outcomes. In recent years, NBI endoscopy has become a standard diagnostic tool for the detection of laryngeal lesions. However, its effective use requires well-trained clinicians and the procedure is time-consuming and subject to interobserver variability. In this context, the application of artificial intelligence (AI) offers a promising solution to support clinical decision-making. In this work, we proposed applying transformer and attention mechanism for analyzing the narrow band imaging and distinguish benign and malignant lesions. Results show it has good classification performance with F1 (82.72%), accuracy(82.33%). In addition, the result of laryngeal cancer screening is explainable for clinicians. The explainability is utilizing the state of art segmentation method (MedSAM) to provide the useful pathological information area for clinicians. The proposed methodology fusing classification and segmentation provides a translating on laryngeal cancer screening.
