Integrating multiple types of biological data for cancer samples is key to identifying potential biomarkers that can predict patient outcome. That’s the conclusion of a new study that systematically examines the relationships between various molecular data in breast cancers and morphological features of tumors – such as cell types making up the tumor, as well as the size, shape and density of those cells.
The study was published on 8 July 2019 in Molecular and Cellular Proteomics. It was led by Kun Huang, PhD, assistant dean for data sciences at the Indiana University School of Medicine.
Tumors are made up of different types of cells, including cancer cells, fibroblasts, and lymphocytes. “While the morphological features of tumors are critical for cancer diagnosis and prognosis, the underlying molecular events and genes for tumor morphology are far from being clear,” the researchers wrote. However, investigators can now get at this issue by carrying out integrative analyses using new methods in computational pathology and the large amount of cancer samples with matched molecular and histopathology data.
In 2016, nearly 250,000 new cases of breast cancer were diagnosed, according to the Centers for Disease Control and Prevention. It is one of the most studied cancer types, using both imaging and molecular biology.
In the current study, the research team systematically examined the relationships between morphological features and various molecular data in breast cancers. They used data from 73 breast cancer patients from The Cancer Genome Atlas (TCGA) and The National Cancer Institute’s Clinical Proteomic Tumor Analysis Consortium projects matched whole slide images, RNA-seq, and proteomic data.
The researchers calculated 100 different morphological features. They then did a correlation analysis using transcriptomic and proteomic data. They found that four major biological processes were associated with various morphological features. These processes include metabolism, cell cycle, immune response, and extracellular matrix development. The authors noted that these are all hallmarks of cancers. The associated morphological features are related to area, density, and shapes of epithelial cells, fibroblasts, and lymphocytes, the wrote.
The researchers also used proteomic data to infer protein-specific biological processes, “suggesting the importance of proteomic data in obtaining a holistic understanding of the molecular basis for tumor tissue morphology.” Looking at patient survival, they found that specific morphological features predicted patient prognosis – both favorable and unfavorable — in 1,057 patients in the TCGA BRCA project.
The morphological features that were associated with unfavorable were linked to large cell nuclei or large distances to neighboring cells which were highly associated with metabolic or extracellular matrix-related biological processes. The features associated with favorable prognosis were indicative of small distances to neighboring cells, which were highly correlated with metabolic or immune related processes.
“More sophisticated modeling and integration methods will lead to deeper understanding of the regulation of the tissue morphology and importance of protein in this process, contributing to the generation of new insights for cancer biology and outcome prediction,” the researchers wrote. Future work, they added, “includes causal analysis to identify key regulators for cancer tissue development and validating the findings using more independent datasets.”
