Laryngeal Lesion Classification Based on Vascular Patterns in Contact Endoscopy and Narrow Band Imaging: Manual Versus Automatic Approach

Abstract

Longitudinal and perpendicular changes in the vocal fold’s blood vessels are associated with the development of benign and malignant laryngeal lesions. The combination of Contact Endoscopy (CE) and Narrow Band Imaging (NBI) can provide intraoperative real-time visualization of the vascular changes in the laryngeal mucosa. However, the visual evaluation of vascular patterns in CE-NBI images is challenging and highly depends on the clinicians’ experience. The current study aims to evaluate and compare the performance of a manual and an automatic approach for laryngeal lesion’s classification based on vascular patterns in CE-NBI images. In the manual approach, six observers visually evaluated a series of CE+NBI images that belong to a patient and then classified the patient as benign or malignant. For the automatic classification, an algorithm based on characterizing the level of the vessel’s disorder in combination with four supervised classifiers was used to classify CE-NBI images. The results showed that the manual approach’s subjective evaluation could be reduced by using a computer-based approach. Moreover, the automatic approach showed the potential to work as an assistant system in case of disagreements among clinicians and to reduce the manual approach’s misclassification issue.

Cyclist Effort Features: A Novel Technique for Image Texture Characterization Applied to Larynx Cancer Classification in Contact Endoscopy—Narrow Band Imaging

“…In studies [ 24 , 25 ], a subset of current CE-NBI image dataset were used for classification of CE-NBI images into benign and malignant classes. In comparison to the results in [ 24 ], the CyEfF set with Polykernel and RBF SVM showed a better classification accuracy.…”

“…The GF set describes the level of disorder of vascular patterns in CE-NBI images [ 23 , 24 ]. This set of features intended to take into account geometrical characteristics of vessels including the consistency of gradient direction and the vessels’ curvature and showed high performances on CE-NBI classification in different datasets [ 24 , 25 ].…”

“…Several computer-based diagnosis approaches were applied to laryngeal endoscopic images to overcome this issue and present complementary information about the state of the larynx for clinicians [ 18 ]. Recent studies included a Deep Convolutional Neural Network (DCNN) using laryngoscopic images for larynx cancer detection [ 19 ], a set of texture-based features and Deep Learning-based descriptors extracted from endoscopic NBI images for laryngeal Squamous Cell Carcinoma (SCC) detection [ 20 ], a set of texture-based and first-order statistical features [ 21 ] plus an ensemble of Convolution Neural Networks (CNN) with texture and frequency domain based features [ 22 ] for larynx cancerous tissue classification using endoscopic NBI images, a set of features combined with supervised Machine Learning techniques for vascular patterns’ assessment in CE-NBI images and laryngeal cancer diagnosis [ 23 , 24 , 25 ].…”

Sensors, Signal and Image Processing in Biomedicine and Assisted Living

“…In the context of endoscopy, narrow band imaging (NBI) constitutes a promising technique for the discrimination of different kinds of lesions as it can provide intraoperative real-time visualization of the vascular changes in the mucosa. A relevant work [ 15 ], presents an automatic approach for classification of laryngeal lesions, based on vascular patterns in Contact Endoscopy NBI images. The proposed approach demonstrated its capacity to act as an aid when there are disagreements among otolaryngologists, or when they all misclassify the patients.…”

“…This special issue gathers a broad range of novel contributions on sensors, systems, and signal/image processing methods for biomedicine and assisted living. These include methods for heart, sleep and vital sign measurement [ 1 , 2 , 3 , 4 , 5 ]; human motion-related signal analysis in the context of rehabilitation and tremor assessment [ 6 , 7 , 8 ]; assistive systems for color deficient and visually challenged individuals, as well as for wheelchair control by people with motor disabilities [ 9 , 10 , 11 , 12 ]; and, image and video-based diagnostic systems [ 13 , 14 , 15 , 16 ].…”