Nuclear features of infiltrating urothelial carcinoma are distinguished from low-grade noninvasive papillary urothelial carcinoma by image analysis

Recent advances in computer technology have been made and image analysis (IA) has been introduced into pathological fields. The present study aimed to investigate the utility of IA for the evaluation of nuclear features and staining of immunohistochemistry (IHC) for Ki-67, p53 and GATA-binding protein 3 (GATA-3) in urothelial carcinoma tissue samples. A total of 49 cases of urothelial carcinoma tissue samples were obtained by transurethral resection of bladder tumors, which included 11 low-grade papillary urothelial carcinomas (LGPUCs), 1 non-invasive high-grade urothelial carcinoma and 37 infiltrating urothelial carcinomas (IUCs). Whole slide imaging (WSI) and IA were performed in Feulgen reaction and IHC-stained tissue samples. There was a significant difference in the average nuclear density, standard deviation (SD) of nuclear size and SD of nuclear minimum and maximum diameter between LGPUC and IUC, which is equivalent to the diagnostic features of IUC in nuclear variability, and hyperchromatic nuclei. In addition, the present study revealed that the SD of nuclear density was significantly different between the two groups. Regarding IA in IHC-stained tissue samples, Ki-67 was significantly overexpressed in IUC. Furthermore, the GATA-3 expression level in IUC samples with muscle invasion was significantly downregulated compared with that in non-muscle invasive tumors. The results of the present study suggest that IA in combination with WSI may be a beneficial tool for evaluating morphometric characteristics and performing semi-quantitative analysis of IHC.