A Comparative Study of NBI and Contact Endoscopy for Early Assessment of Malignancy in Head and Neck Lesions at A Tertiary Center in Northwest India
摘要
This study is the embodiment of the challenges made by newer diagnostic tests to the standard ones in Head and Neck oncology. With the advent of NBI and Contact endoscopy it has become possible to assess the malignant potential of early and premalignant lesions today. This study aims at finding whether these newer modalities can serve as reliable indicators or become newer standards for diagnosis in Head and Neck oncology as opposed to HPE which is the current gold standard. The study was designed as a prospective cohort study. Patients included in the study(leukoplakia, erythroplakia, erythroleukoplakia, DLE, DKC, Glossitis, Laryngeal Dysplasia, Polyps) undergo NBI and CE and then proceeded to HPE of the lesion post excision and the observations made in all three tests were compared statistically. A total of 155 patients were enrolled and went on for final analysis based on the availability of complete Narrow Band Imaging (NBI), Contact Endoscopy (CE), and histopathology data. NBI demonstrated superior diagnostic performance, with sensitivity, specificity, PPV, NPV, and overall accuracy of 88.7%, 92.9%, 91.3%, 90.7%, and 91%, respectively. CE showed slightly lower values: 85.9%, 86.9%, 84.7%, 87.9%, and 86.5%. NBI performed particularly well in dysplasia, proliferative verrucous leukoplakia (PVL), and erythroleukoplakia (ELP), achieving near-perfect sensitivity and specificity in several categories. Lesion types such as leukoplakia, denture keratosis, and polyps showed no true positives, resulting in zero sensitivity for both modalities in these subgroups. NI classification analysis further supported NBI’s advantage, with Type V lesions showing 100% specificity and PPV and 88.9% sensitivity. Although CE followed similar patterns, its specificity and predictive values were consistently lower. Both modalities showed strong diagnostic association with histopathology (p < 0.00001), though direct statistical comparison revealed no significant difference (t-test p = 0.210; McNemar’s p = 0.296). ROC analysis favored NBI (AUC = 0.91) over CE (AUC = 0.86), indicating better discriminative capability, particularly in high-risk lesions. By comparing lesion-specific diagnostic performance and integrating insights from prior literature, this study confirms that NBI outperforms CE in specificity across both malignant and benign lesions, without compromising sensitivity. These findings support the integration of NBI into standard diagnostic algorithms for mucosal lesion evaluation in the head and neck region, especially when the clinical goal is to minimize unnecessary biopsies while ensuring early cancer detection.