Interrater variation of vascular classifications used in enhanced laryngeal contact endoscopy

Clinical Efficacy of Enhanced Contact Endoscopy in Early Detection of Laryngeal Carcinoma: A Single Center Experience

INTRODUCTION

Enhanced contact endoscopy (ECE) is a non-invasive technique used for the assessment of superficial vascular changes of mucosal lesions in high magnification. The aim of our study was to evaluate the clinical efficacy of ECE in an intraoperative settlement.

METHODS

Structured assessment of laryngeal mucosal lesions using enhanced endoscopy (narrow band imaging (NBI) and ECE) was performed in a prospective clinical trial. Lesions were classified according to the European Laryngological Society Classification into non-suspicious and suspicious. Evaluations of endoscopic methods (NBI and ECE) were correlated with histopathology, histopathology being the gold standard. Sensitivity, specificity, accuracy, positive predictive value (PPV), negative predictive value (NPV), area under curve, diagnostic odds ratio (DOR), Kappa, incremental yield, and Youden´s index for NBI and ECE were calculated.

RESULTS

A total of 110 patients with 136 lesions were enrolled, 50 benign non-neoplastic lesions, eight squamous cell papillomas, 45 dysplasias, and 33 squamous cell invasive cancers. Compared to NBI, ECE demonstrated higher sensitivity (91.0% vs 83.1%) and accuracy (90.4% vs 86.8%). NBI achieved higher specificity (91.8% vs 89.7%). PPV and NPV for ECE were 92.2% and 88.1%, whereas for NBI 93.1% and 80.4%. ECE showed greater overall diagnostic performance, with a DOR of 88.3 vs 55.2 and Kappa index of 0.805 vs 0.736.

CONCLUSIONS

ECE enhances diagnostic sensitivity and accuracy and represents a valuable addition to laryngeal cancer diagnostics.

Clinical significance of type IV vascularization of laryngeal lesions according to the Ni classification

Background

Scattered, small, dot-like intraepithelial papillary capillary loops (IPCLs) represent type IV epithelial vascularization according to "Ni classification" and are considered to be nonmalignant. According to the European Laryngological Society classification, these loops are malignant vascular changes. This contradiction has high clinical importance; therefore, clarification of the clinical significance of type IV vascularization according to the Ni classification is needed.

Methods

The study was performed between June 2015 and December 2022. All recruited patients (n = 434) were symptomatic, with macroscopic laryngeal lesions (n = 674). Patients were investigated using the enhanced endoscopic methods of narrow band imaging (NBI) and the Storz Professional Image Enhancement System (IMAGE1 S). The microvascular patterns in the lesions were categorized according to Ni classification from 2011 and all lesions were examined histologically.

Results

A total of 674 lesions (434 patients) were investigated using flexible NBI endoscopy and IMAGE1 S endoscopy. Type IV vascularization was recognized in 293/674 (43.5%) lesions. Among these 293 lesions, 178 (60.7%) were benign (chronic laryngitis, hyperplasia, hyperkeratosis, polyps, cysts, granulomas, Reinkeho oedema and recurrent respiratory papillomatosis); 9 (3.1%) were squamous cell carcinoma; 61 (20.8%) were mildly dysplastic, 29 (9.9%) were moderately dysplastic, 14 (4.8%) were severe dysplastic and 2 (0.7%) were carcinoma in situ. The ability to recognize histologically benign lesions in group of nonmalignant vascular pattern according to Ni (vascularization type I-IV) and distinguish them from precancers and malignancies was with accuracy 75.5%, sensitivity 54.4%, specificity 94.4%, positive predictive value 89.6% and negative predictive value 69.9%.

Conclusion

Laryngeal lesions with type IV vascularization as defined by Ni present various histological findings, including precancerous and malignant lesions. Patients with type IV vascularization must be followed carefully and, in case of progression mucosal lesion microlaryngoscopy and excision are indicated.

Contact Endoscopy - Narrow Band Imaging (CE-NBI) data set for laryngeal lesion assessment

The endoscopic examination of subepithelial vascular patterns within the vocal fold is crucial for clinicians seeking to distinguish between benign lesions and laryngeal cancer. Among innovative techniques, Contact Endoscopy combined with Narrow Band Imaging (CE-NBI) offers real-time visualization of these vascular structures. Despite the advent of CE-NBI, concerns have arisen regarding the subjective interpretation of its images. As a result, several computer-based solutions have been developed to address this issue. This study introduces the CE-NBI data set, the first publicly accessible data set that features enhanced and magnified visualizations of subepithelial blood vessels within the vocal fold. This data set encompasses 11144 images from 210 adult patients with pathological vocal fold conditions, where CE-NBI images are annotated using three distinct label categories. The data set has proven invaluable for numerous clinical assessments geared toward diagnosing laryngeal cancer using Optical Biopsy. Furthermore, given its versatility for various image analysis tasks, we have devised and implemented diverse image classification scenarios using Machine Learning (ML) approaches to address critical clinical challenges in assessing laryngeal lesions.

Narrative Review of Classification Systems Describing Laryngeal Vascularity Using Advanced Endoscopic Imaging

Endoscopic methods are critical in the early diagnosis of mucosal lesions of the head and neck. In recent years, new examination methods and classification systems have been developed and introduced into clinical practice. All of these new techniques target the notion of optical biopsy, which tries to assess the nature of the lesion before histology examination. Many methods suffer from interpretation issues due to subjective interpretation of the findings. Therefore, multiple classification systems have been developed to assist the proper interpretation of mucosal findings and reduce the error rate. They provide various perspectives on the assessment and interpretation of mucosa changes. This article provides a comprehensive and critical view of the available classification systems as well as their advantages and disadvantages.

Leukoplakia: An Invasive Cancer Hidden within the Vocal Folds. A Multivariate Analysis of Risk Factors

Introduction

Discerning the preoperative nature of vocal fold leukoplakia (VFL) with a substantial degree of certainty is fundamental, seeing that the histological diagnosis of VFL includes a wide spectrum of pathology and there is no consensus on an appropriate treatment strategy or frequency of surveillance. The goal of our study was to establish a clear schedule of the diagnostics and decision-making in which the timing and necessity of surgical intervention are crucial to not miss this cancer hidden underneath the white plaque.

Material and Methods

We define a schedule as a combination of procedures (white light and Narrow Band Imaging diagnostic tools), methods of evaluating the results (a combination of multiple image classifications in white light and Narrow Band Imaging), and taking into account patient-related risk factors, precise lesion location, and morphology. A total number of 259 patients with 296 vocal folds affected by leukoplakia were enrolled in the study. All patients were assessed for three classifications, in detail according to Ni 2019 and ELS 2015 for Narrow Band Imaging and according to Chen 2019 for white light. In 41 of the 296 folds (13.9%), the VFL specimens in the final histology revealed invasive cancer. We compared the results from the classifications to the final histology results.

Results

The results showed that the classifications and evaluations of the involvement of anterior commissure improve the clinical utility of these classifications and showed improved diagnostic performance. The AUC of this model was the highest (0.973) with the highest sensitivity, specificity, PPV, and NPV (90.2%, 89%, 56.9%, and 98.3%, respectively).

Conclusion

The schedule that combines white light and Narrow Band Imaging, with a combination of the two classifications, improves the specificity and predictive value, especially of anterior commissure involvement.

Deep Convolution Neural Network for Laryngeal Cancer Classification on Contact Endoscopy-Narrow Band Imaging

(1) Background: Contact Endoscopy (CE) and Narrow Band Imaging (NBI) are optical imaging modalities that can provide enhanced and magnified visualization of the superficial vascular networks in the laryngeal mucosa. The similarity of vascular structures between benign and malignant lesions causes a challenge in the visual assessment of CE-NBI images. The main objective of this study is to use Deep Convolutional Neural Networks (DCNN) for the automatic classification of CE-NBI images into benign and malignant groups with minimal human intervention. (2) Methods: A pretrained Res-Net50 model combined with the cut-off-layer technique was selected as the DCNN architecture. A dataset of 8181 CE-NBI images was used during the fine-tuning process in three experiments where several models were generated and validated. The accuracy, sensitivity, and specificity were calculated as the performance metrics in each validation and testing scenario. (3) Results: Out of a total of 72 trained and tested models in all experiments, Model 5 showed high performance. This model is considerably smaller than the full ResNet50 architecture and achieved the testing accuracy of 0.835 on the unseen data during the last experiment. (4) Conclusion: The proposed fine-tuned ResNet50 model showed a high performance to classify CE-NBI images into the benign and malignant groups and has the potential to be part of an assisted system for automatic laryngeal cancer detection.

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

BACKGROUND

Feature extraction is an essential part of a Computer-Aided Diagnosis (CAD) system. It is usually preceded by a pre-processing step and followed by image classification. Usually, a large number of features is needed to end up with the desired classification results. In this work, we propose a novel approach for texture feature extraction. This method was tested on larynx Contact Endoscopy (CE)-Narrow Band Imaging (NBI) image classification to provide more objective information for otolaryngologists regarding the stage of the laryngeal cancer.

METHODS

The main idea of the proposed methods is to represent an image as a hilly surface, where different paths can be identified between a starting and an ending point. Each of these paths can be thought of as a Tour de France stage profile where a cyclist needs to perform a specific effort to arrive at the finish line. Several paths can be generated in an image where different cyclists produce an average cyclist effort representing important textural characteristics of the image. Energy and power as two Cyclist Effort Features (CyEfF) were extracted using this concept. The performance of the proposed features was evaluated for the classification of 2701 CE-NBI images into benign and malignant lesions using four supervised classifiers and subsequently compared with the performance of 24 Geometrical Features (GF) and 13 Entropy Features (EF).

RESULTS

The CyEfF features showed maximum classification accuracy of 0.882 and improved the GF classification accuracy by 3 to 12 percent. Moreover, CyEfF features were ranked as the top 10 features along with some features from GF set in two feature ranking methods.

CONCLUSION

The results prove that CyEfF with only two features can describe the textural characterization of CE-NBI images and can be part of the CAD system in combination with GF for laryngeal cancer diagnosis.

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

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.