Abstract: Described herein are methods and computer systems for classification of CD8 T-cell topology using a patient response-based linear cutoff model. A plurality of histology images of tissue samples in a plurality of patients are received by a computer system. An image analysis of the plurality of histology images is performed to obtain a CD8+ T-cell abundance in the tumor parenchyma and stroma in each of the plurality of histology images. Real inflammation scores and tumor infiltration scores are determined based on a polar coordinate transformation of the CD8+ T-cell abundance in the tumor parenchyma and stroma. Based on the real inflammation scores and tumor infiltration scores, a feature space is generated, and linear boundaries or linear cutoffs between a plurality of classifications in the feature space are identified based on the real inflammation scores, the tumor infiltration scores, and patient response data.

Abstract: The present disclosure provides methods of identifying a subject suitable for an anti-PD-?PD-L1 antagonist therapy comprising measuring assay CD8 localization and PD-L1 expression in a tumor sample obtained from the subject. In some aspects, method further comprises administering (i) an anti-PD-?PD-L1 antagonist therapy or (ii) an anti-PD-?PD-L1 antagonist and anti-CT-LA-4 antagonist combination therapy to a subject identified as having a tumor exhibiting an excluded CD8 localization phenotype, wherein the tumor is PD-L1 negative.

Abstract: Described herein are methods and computer systems for classification of CD8 T-cell topology using artificial intelligence and machine learning. A plurality of histology images of tissue samples in a plurality of patients are received by a computer system. An image analysis of the plurality of histology images is performed to obtain a CD8+ T-cell abundance in the tumor parenchyma and stroma in each of the plurality of histology images. A machine learning algorithm is then trained using results of the image analysis and the CD8+ T-cell abundance in the tumor parenchyma and stroma. Based on the training, a machine learning feature space comprising a plurality of classifications is generated, and boundaries between the plurality of classifications in the machine learning feature space are identified.