A Novel Classification Model Based on Radiomics for Narrow Band Imaging in Laryngeal Cancer

Effective prevention and diagnosis of laryngeal cancer are crucial for improving its prognosis, particularly among populations at elevated risk. This highlights the significance of this oncologic condition. An early diagnosis in these group people is crucial. It could allow a patient to take a timely measure and have a higher survival rate. Here a novel classifying model based on radiomics for narrow band imaging in contact endoscopy for laryngeal cancer is proposed in tackling this issue. The results are analyzed by a series of performance metrics: accuracy, precision, recall, F1 score and area under cure. The analysis shows the model based on radiomics could distinguish benign and malignant with reliable accuracy. It has a good potential for preliminary diagnosis assistant in the future.