We examined whether endocytoscopic observation of esophageal squamous cell carcinoma can replace the histologic examination of biopsy specimens. In a basic investigation, we examined 57 iodine-unstained areas in the resected specimens of the esophagus from 28 individuals. The endocytoscopic findings were graded from 0 to 3 in tandem with observations of the iodine staining. For endocytoscopic observation, we sprayed 1% methylene blue or toluidine blue as a vital dye on the surface of the esophageal mucosa, allowing 15-20 s for sufficient staining. One endoscopist observed the target lesions and decided their endocytoscopic type classification. Histological diagnoses were made by two pathologists who were unaware of the endoscopic findings. We then compared the endocytoscopic diagnosis and conventional histological diagnosis. In an in vivo investigation, we examined 71 lesions of esophageal squamous cell carcinoma. Two endoscopists diagnosed the type classification in consultation with a pathologist with regard to 'nuclear density,''nuclear abnormality,' and 'whether biopsy histology could have been omitted on the basis of endocytoscopic findings.' For the in vivo observation, we utilized XEC120U (higher magnification type [x1100]), XEC300F (lower magnification type [x450]), and XGIF-Q260EC1 (lower magnification type [x450]) instruments. In the basic investigation, among the 11 areas classified as Type 1, 10 (91%) were category 1 by the Vienna classification. Among the 39 lesions classified as Type 3, 36 (92%) were category 4 or 5. The sensitivity of endocytoscopy for malignant lesions (Vienna classification categories 4 and 5) was 94.7%, if Type 3 was considered malignant. The specificity was 84.2% according to the same criteria. In the in vivo observation, two endoscopists diagnosed more than 90% of esophageal squamous cell carcinomas as neoplasms using each type of endocytoscope. With regard to nuclear density, the pathologist considered it to be increased in 98% of cases with the XEC120U, in 94% with the XEC300F, and in 93% with the XGIF-Q260EC1. With regard to nuclear abnormality, the positivity rate was 90% with the XEC120U, 78% with the XEC300F, and 80% with the XGIF-Q260EC1. As to whether or not biopsy histology examination was considered necessary, the pathologist made a 'Yes' judgment for 84% of cases observed with the XEC120U, 66% with the XEC300F, and 67% with the XGIF-Q260EC1. Cancerous lesions diagnosed as Type 3 by both endoscopists using the XEC120U accounted for 46 (90.2%) of the 51 cases. Among these 46 cases, biopsy histology was considered unnecessary by the pathologist in 43 (93.5%). We believe that endocytoscopic observation has the potential to reduce the extent of histologic examination of biopsy specimens in cases corresponding to Types 1 and 3 of our classification.