Clear cell renal cell carcinoma is the most common type of kidney cancer, comprising 80% of all malignant tumors found within the kidney. Although some clear cell renal cell carcinoma tumors are sensitive to immune checkpoint inhibitors, currently, there are no measures to predict whether a tumor will respond to immune checkpoint inhibition. Researchers investigating a potential new way to assess clinically valuable features of clear cell renal cell carcinoma using an artificial intelligence (AI)-based assessment tool that evaluates two-dimensional pictures of a tumor sample on a pathology slide have identified previously underappreciated features that may help predict whether a tumor will respond to immunotherapy. The study by Nyman et al was published in Cell Reports Medicine.
Study Methodology
The researchers developed prediction models that provide high-resolution, quantitative, and human-understandable representations of clear cell renal cell carcinoma hematoxylin and eosin using diagnostic whole-slide images to identify established pathology features, such as tumor tissue and nuclear grade at scale. The study included tumor samples from 1,102 patients with clear cell renal cell carcinoma. Some of the samples were taken from 439 patients enrolled in the CheckMate 025 randomized phase III clinical trial, which tested monotherapy with an immune checkpoint inhibitor or an mTOR inhibitor in patients who had been previously treated with standard therapy.
Results
The researchers discovered patterns of nuclear grade heterogeneity in the diagnostic whole-slide images that were not achievable through human pathologist analysis. These graph-based “microheterogeneity” structures were associated with PBRM1 loss of function, adverse clinical factors, and selective patient response to immune checkpoint inhibitors.
Joint computer vision analysis of tumor phenotypes with inferred tumor-infiltrating lymphocyte density identified a further subpopulation of highly infiltrated, microheterogeneous tumors responsive to immune checkpoint inhibitors. In paired multiplex immunofluorescence images of clear cell renal cell carcinoma tumors, microheterogeneity was associated with greater PD-1 activation, CD8-positive lymphocytes, and increased tumor-immune interactions.
“Our work reveals novel spatially interacting tumor-immune structures underlying clear cell renal cell carcinoma biology that can also inform selective response to immune checkpoint inhibitors,” concluded the study authors.
Ongoing Investigation
Although this AI assessment tool is not ready for clinical use, the researchers are testing it in an ongoing clinical trial involving combination immunotherapy as first-line treatment in patients with clear cell renal cell carcinoma. They also plan to investigate whether the visual clues they found in pathology slides are related to molecular features of the tumor, such as alterations in genes.
“The use of deep-learning strategies to identify tumor and microenvironmental features from histopathology slides and determine their relationship to molecular and clinical states may have value across tumor types and therapeutic modalities,” said Eliezer Van Allen, MD, Chief, Division of Population Sciences at Dana-Farber Cancer Institute, Boston, and principal investigator of the study.
Disclosure: For full disclosures of the study authors, visit cell.com.
