Interobserver agreement of confocal laser endomicroscopy for detection of head and neck neoplasia. (15th September 2015)

• Record Type: Journal Article
• Authors: Moore, Charles; Mehta, Vikas; Ma, Xiaohui; Chaudhery, Shabnum; Shi, Runhua; Moore‐Medlin, Tara; Lian, Timothy; Nathan, Cherie‐Ann O.
• Is Part Of: Laryngoscope. Volume 126:Number 3(2016:Mar.)
• Page Start: 632
• Page End: 637
• Publication Date: 2015-09-15
• Subjects: Head and neck -- oral cavity -- larynx; Otolaryngology -- Periodicals; 617.51005
• DOI: 10.1002/lary.25646
• Languages: English
• ISSNs: 0023-852X

Abstract

Abstract : Objectives/Hypothesis: We have described the feasibility of using the probe‐based confocal laser endomicroscopy (pCLE) in differentiating benign from malignant lesions of the head and neck. Therefore, we wanted to determine the interobserver agreement of pCLE offline images of noncancerous, precancerous, and cancerous lesions of the head and neck. Study Design: Single tertiary referral center. Methods: In the feasibility study, image criteria for nondysplasia, dysplasia, and cancer were defined. The pCLE was performed before lesions were biopsied. Fifty offline images and 10 videos of good quality were selected. Seven surgeons and one pathologist were asked to review and categorize the images into the three categories above. The overall accuracy of 29 offline pCLE images and six videos were compared with histopathology. Interobserver agreement and accuracy kappa (κ) scores were measured with 95% confidence intervals (CI). Results: There were six nondysplasia, seven dysplasia, and 11 squamous cell carcinoma (SCCA) cases, each with multiple images. There was substantial agreement between the eight reviewers on the pCLE images and videos (κ = 0.66; 95% CI 0.51–0.82 and κ = 0.71; 95% CI 0.42–0.97, respectively). The overall agreement with the final histopathology was also substantial for both the images and video sequences (κ = 0.70; 95% CI 0.50–0.88 and κ = 0.73; 95% CI 0.39–1.00, respectively). Conclusion: The ability to differentiate normal mucosa, dysplasia, and invasive SCCA using pCLE with high accuracy and reliability was demonstrated. This technology has the potential to decrease sampling error of lesions in the head and neck. This is the first study to test the reliability of this technology in mucosal lesions of the head and neck. Level of Evidence: NA. Laryngoscope, 126:632–637, 2016