Pan-keratin Immunostaining in Human Tumors: A Tissue Microarray Study of 15,940 Tumors

To evaluate the efficiency of pan-keratin immunostaining, tissue microarrays of 13,501 tumor samples from 121 different tumor types and subtypes as well as 608 samples of 76 different normal tissue types were analyzed by immunohistochemistry. In normal tissues, strong pan-keratin immunostaining was seen in epithelial cells. Staining intensity was lower in hepatocytes, islets of Langerhans, and pneumocytes but markedly reduced in the adrenal cortex. Pan-keratin was positive in >= 98% of samples in 62 (83%) of 75 epithelial tumor entities, including almost all adenocarcinomas, squamous cell and urothelial carcinomas. Only 17 of 121 tumor entities (13%) had a pan-keratin positivity rate between 25% and 98%, including tumors with mixed differentiation, endocrine/neuroendocrine tumors, renal cell carcinomas, adrenocortical tumors, and particularly poorly differentiated carcinoma subtypes. The 15 entities with pan-keratin positivity in 0.9%-25% were mostly of mesenchymal origin. Reduced/absent pan-keratin immunostaining was associated with high UICC stage (p = 0.0001), high Thoenes grade (p = 0.0183), high Fuhrman grade (p = 0.0049), advanced tumor stage (p < 0.0001) and lymph node metastasis (p = 0.0114) in clear cell renal cell carcinoma, advanced pT stage (p = 0.0007) in papillary renal cell carcinoma, and with advanced stage (p = 0.0023), high grade (p = 0.0005) as well as loss of ER and PR expression (each p < 0.0001) in invasive breast carcinoma of no special type (NST). In summary, pan-keratin can consistently be detected in the vast majority of epithelial tumors, although pan-keratin can be negative a fraction of renal cell, adrenocortical and neuroendocrine neoplasms. The data also link reduced pan-keratin immunostaining to unfavorable tumor phenotype in in epithelial neoplasms.