The intrapapillary capillary loop (IPCL) characteristics, visualized using magnifying endoscopy, are commonly assessed for preoperative evaluation of the infiltration depth of esophageal squamous cell carcinoma (ESCC). Japan Esophageal Society (JES) classification is the most widely used classification. Microvascular structural changes are evaluated by magnifying endoscopy for the presence or absence of each morphological factor: tortuosity, dilatation, irregular caliber, and different shapes. However, the pathological characteristics of IPCLs have not been thoroughly investigated, especially the microvascular structures corresponding to the deepest parts of the lesions' infiltration.

To investigate differences in pathological microvascular structures of ESCC, which correspond to the deepest parts of the lesions' infiltration.

Patients with ESCC and precancerous lesions diagnosed at Peking University Third Hospital were enrolled between January 2019 and April 2023. Patients first underwent magnified endoscopic examination, followed by endoscopic submucosal dissection or surgical treatment. Pathological images were scanned using a three-dimensional slice scanner, and the pathological structural differences in different types, according to the JES classification, were analyzed using nonparametric tests and t-tests.

The 35 lesions were divided into four groups according to the JES classification: A, B1, B2, and B3. Statistical analyses revealed significant differences (a P < 0.05) in the short and long calibers, area, location, and density between types A and B. Notably, there were no significant differences in these parameters between types B1 and B2 and between types B2 and B3 (P > 0.05). However, significant differences in the short calibers, long calibers, and area of IPCL were observed between types B1 and B3 (a P < 0.05); no significant differences were found in the density or location (P > 0.05).

Pathological structures of IPCLs in the deepest infiltrating regions differ among various IPCL types classified by the JES classification under magnifying endoscopy, especially between the types A and B.

Eosinophils in the esophageal epithelium are unevenly distributed in eosinophilic esophagitis (EoE). Esophageal eosinophilia (EE) may be observable by endocytoscopy (EC). This study aimed to evaluate the diagnostic performance of EC for the diagnosis of EE.

A total of 33 EoE patients underwent EC with methylene blue staining from March 2020 to April 2021. A total of 194 EC images with corresponding biopsies were obtained. Three findings of EC, increased squamous cells (item I), increased inflammatory cells (item II), and cells with bilobed nuclei (item III), were established. These findings were reviewed by two endoscopists to diagnose EE. Another four endoscopists reviewed the images for interobserver agreement.

When all three items were met by EC, the sensitivity and the accuracy for the diagnosis of EE were 88% and 76%, respectively. The integrated diagnostic odds ratios (ORs) for the diagnosis of EE of the four endoscopists were significant (OR: 3.98, 95% CI: 2.94-5.40, p < 0.001). The results were similar when only item III was met. Interobserver agreement was good for item III to diagnose EE (kappa value = 0.653).

The diagnostic performance of EC for EE is acceptable and has good interobserver agreement. It may be useful for targeted biopsy in EoE patients.

This review outlines the process of the development of the endocytoscope (EC) with reference to previously reported studies including our own. The EC is an ultra-high-magnification endoscope capable of imaging at the cellular level. The esophagus is the most suitable site for EC observation because it is amenable to vital staining. The diagnosis of esophageal lesions using EC is based on nuclear density and nuclear abnormality, allowing biopsy histology to be omitted. The observation of nuclear abnormality requires a magnification of ×600 or higher using digital technology. Several staining methods have been proposed, but single staining with toluidine blue or methylene blue is most suitable because the contrast at the border of a cancerous area can be easily identified. A three-tier classification of esophageal lesions visualized by EC is proposed: Type 1 (non-cancerous), Type 2 (endocytoscopic borderline), and Type 3 (cancerous). Since characteristic EC images reflecting pathology can be obtained from non-cancerous esophageal lesions, a modified form of classification with four additional characteristic non-cancerous EC features has also been proposed. Recently, deep-learning AI for analysis of esophageal EC images has revealed that its diagnostic accuracy is comparable to that of expert pathologists.