In Cancer science
While renal pelvic and ureteral urothelial carcinoma shared similarities in the origin, disparities on a genetic and clinical level make them divergent entities. Clinical information from SEER database were used to validate the characteristics and molecular subtypes using single-center data were compared between two muscle-invasive tumors. Simultaneously, to expand the sample size for further verification, we explored a deep learning algorithm to correctly classify two tumor's molecular subtype from hematoxylin and eosin (H&E) histology slides. We suggested that the renal pelvic group might have a proclivity towards luminal and the ureter towards basal and P53-like. Furthermore, we explore the heterogeneity of matrix and immune tumor microenvironment, and ureteral group had more immune cell infiltration and higher stiffness. Collectively, these results showed that muscle-invasive UTUC exist in distinct properties of clinical characteristics, molecular subtype, and tumor microenvironment.
Qiwei Chen, Jiajun Shi, Cheng Liang, Shengbo Huang, Yue Kuai, Shujing Wang, Wenlong Liu, Xinqing Zhu, Hongyu Wang, Deyong Yang
2022-Nov-03
UTUC, deep learning, immunity, mechanics, molecular subtype
