Abstract
Computational pathology-based cell classification algorithms are revolutionizing the study of the tumor microenvironment and can provide novel predictive/prognosis biomarkers crucial for the delivery of precision oncology. Current algorithms used on hematoxylin and eosin slides are based on individual cell nuclei morphology with limited local context features. Here, we propose a novel multi-resolution hierarchical framework (SuperCRF) inspired by the way pathologists perceive regional tissue architecture to improve cell classification and demonstrate its clinical applications. We develop SuperCRF by training a state-of-art deep learning spatially constrained- convolution neural network (SC-CNN) to detect and classify cells from 105 high-resolution (20×) H&E-stained slides of The Cancer Genome Atlas melanoma dataset and subsequently, a conditional random field (CRF) by combining cellular neighborhood with tumor regional classification from lower resolution images (5, 1.25×) given by a superpixel-based machine learning framework. SuperCRF led to an 11.85% overall improvement in the accuracy of the state-of-art deep learning SC-CNN cell classifier. Consistent with a stroma-mediated immune suppressive microenvironment, SuperCRF demonstrated that (i) a high ratio of lymphocytes to all lymphocytes within the stromal compartment (p = 0.026) and (ii) a high ratio of stromal cells to all cells (p < 0.0001 compared to p = 0.039 for SC-CNN only) are associated with poor survival in patients with melanoma. SuperCRF improves cell classification by introducing global and local context-based information and can be implemented in combination with any single-cell classifier. SuperCRF provides valuable tools to study the tumor microenvironment and identify predictors of survival and response to therapy.
Highlights
Cancer is a highly complex, non-autonomous disease
We demonstrated that our new system is computationally efficient and significantly improves single cell classification
The spatially constrained- convolution neural network (SC-convolution neural networks (CNNs)) network yielded an accuracy of 84.63% over 4,059 cells in the independent test set (Table 1, Supplementary Table 5)
Summary
Cancer is a highly complex, non-autonomous disease. The interactions between microenvironmental selective pressures and cancer cells dictate how cancer progresses and evolves. Accurate and spatially explicit characterization of the tumor microenvironmental landscape including how cancer cells interact with the extra-cellular matrix and other cellular players such as stromal cells and immune cells within the tumoral niche, is needed to understand the context in which cancer evolves, and may provide robust predictor of cancer behavior for riskstratification [1]. The recent success of cancer immunotherapy including the spectacular response observed in patients with previously incurable melanoma, a highly aggressive form of skin cancer, calls for a better understanding of the cancer-immune interface. More recently deep learning algorithms, both exploiting the phenotypic differences in nuclear morphology between each cell type, revolutionized the field yielding significantly better cell detection, segmentation, and classification results [3,4,5,6,7,8,9]
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
