We aimed to assess the feasibility of using confocal laser scanning microscopy (CLSM) for the real-time ex vivo examination of histological samples of laryngeal lesions and to evaluate the correlation between CLSM and definitive histological results. This preliminary study included eight consecutive patients with "suspected" laryngeal lesions who were candidates for endoscopic laryngeal surgery. The obtained samples were evaluated using CLSM and classified as "inadequate" or "adequate" (high- and low-grade dysplasia, in situ and invasive carcinoma, positive surgical margin, and inflammatory outbreaks). CLSM showed the macro image in all cases and generated a digital version. All the samples were defined as adequate during CLSM and confirmed at histopathology: low-grade dysplasia (n = 5), low- and high-grade dysplasia (n = 2), and high-grade dysplasia (n = 1). Four samples had an involved resection margin, and three samples revealed the presence of inflammatory outbreaks. CLSM can be applied to larynx pathology with excellent agreement with final histological results.
Keywords: artificial intelligence; confocal laser scanning microscopy; endoscopic laryngeal surgery; laryngeal dysplasia.