Development of an Immune-Pathology Informed Radiomics Model for Non-Small Cell Lung Cancer

Abstract

With increasing use of immunotherapy agents, pretreatment strategies for identifying responders and non-responders is useful for appropriate treatment assignment. We hypothesize that the local immune micro-environment of NSCLC is associated with patient outcomes and that these local immune features exhibit distinct radiologic characteristics discernible by quantitative imaging metrics. We assembled two cohorts of NSCLC patients treated with definitive surgical resection and extracted quantitative parameters from pretreatment CT imaging. The excised primary tumors were then quantified for percent tumor PDL1 expression and density of tumor-infiltrating lymphocyte (via CD3 count) utilizing immunohistochemistry and automated cell counting. Associating these pretreatment radiomics parameters with tumor immune parameters, we developed an immune pathology-informed model (IPIM) that separated patients into 4 clusters (designated A-D) utilizing 4 radiomics features. The IPIM designation was significantly associated with overall survival in both training (5 year OS: 61%, 41%, 50%, and 91%, for clusters A-D, respectively, P = 0.04) and validation (5 year OS: 55%, 72%, 75%, and 86%, for clusters A-D, respectively, P = 0.002) cohorts and immune pathology (all P < 0.05). Specifically, we identified a favorable outcome group characterized by low CT intensity and high heterogeneity that exhibited low PDL1 and high CD3 infiltration, suggestive of a favorable immune activated state. We have developed a NSCLC radiomics signature based on the immune micro-environment and patient outcomes. This manuscript demonstrates model creation and validation in independent cohorts.