Fig. 6: Histopathologic validation of the universal immunohistochemistry (UIHC) model. | npj Precision Oncology

Fig. 6: Histopathologic validation of the universal immunohistochemistry (UIHC) model.

From: A universal immunohistochemistry analyzer for generalizing AI-driven assessment of immunohistochemistry across immunostains and cancer types

Fig. 6: Histopathologic validation of the universal immunohistochemistry (UIHC) model.The alternative text for this image may have been generated using AI.

a Example of the prediction by P-L and UIHC for positively stained tumor cells (TC+) and negatively stained tumor cells (TC-) on a lung cancer whole slide image (WSI) stained with mesenchymal-epithelial transition factor (MET). The UIHC prediction yields values closer to the pathologists’ ground truth TPS value, as shown in the upper. b Example from a gastric cancer WSI stained with fibroblast growth factor receptor 2 (FGFR2). GT TPS, ground truth tumor proportion score; P-L, Programmed Death-Ligand 1 (PD-L1) 22C3 of lung; TC tumor cell.

Back to article page