Search

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.

Collagen plays a key role in normal and malignant tissue homeostasis. While the prognostic significance of collagen fibre remodeling in invasive breast cancer has been studied, its role in ductal carcinoma in situ (DCIS) remains poorly defined. Using image analysis, we aimed to evaluate the prognostic significance of the geometric characteristics of collagen surrounding DCIS. A large well-characterized cohort of DCIS comprising pure DCIS (n=610) and DCIS co-existing with invasive carcinoma (n=180) were histochemically stained for collagen using picrosirius red. ImageJ software was used to assess collagen density, degree of collagen fibre dispersion and directionality in relation to DCIS ducts’ boundary. We developed a collagen prognostic index and evaluated its prognostic significance. A poor index was observed in 24% of the pure DCIS and was associated with determinants of high-risk DCIS including higher grade, comedo necrosis, hormonal receptor negativity, HER2 positivity and high proliferation index. High index was associated with overexpression of the collagen remodeling protein prolyl-4-hydroxlase alpha 2 and the hypoxia inducible factor 1α. DCIS co-existing with invasive carcinoma had a higher collagen prognostic index than pure DCIS (p

The spectacular response observed in clinical trials of immunotherapy in patients with previously uncurable Melanoma, a highly aggressive form of skin cancer, calls for a better understanding of the cancer-immune interface. Computational pathology provides a unique opportunity to spatially dissect such interface on digitised pathological slides. Accurate cellular classification is a key to ensure meaningful results, but is often challenging even with state-of-art machine learning and deep learning methods. We propose a hierarchical framework, which mirrors the way pathologists perceive tumour architecture and define tumour heterogeneity to improve cell classification methods that rely solely on cell nuclei morphology. The SLIC superpixel algorithm was used to segment and classify tumour regions in low resolution H&E-stained histological images of melanoma skin cancer to provide a global context. Classification of superpixels into tumour, stroma, epidermis and lumen/white space, yielded a 97.7% training set accuracy and 95.7% testing set accuracy in 58 whole-tumour images of the TCGA melanoma dataset. The superpixel classification was projected down to high resolution images to enhance the performance of a single cell classifier, based on cell nuclear morphological features, and resulted in increasing its accuracy from 86.4% to 91.6%. Furthermore, a voting scheme was proposed to use global context as biological a priori knowledge, pushing the accuracy further to 92.8%. This study demonstrates how using the global spatial context can accurately characterise the tumour microenvironment and allow us to extend significantly beyond single-cell morphological classification., Accepted by MICCAI COMPAY 2018 workshop

The role of HER4 in breast cancer is controversial and its role in relation to trastuzumab resistance remains unclear. We showed that trastuzumab treatment and its acquired resistance induced HER4 upregulation, cleavage and nuclear translocation. However, knockdown of HER4 by specific siRNAs increased trastuzumab sensitivity and reversed its resistance in HER2 positive breast cancer cells. Preventing HER4 cleavage by a γ-secretase inhibitor and inhibiting HER4 tyrosine kinase activity by neratinib decreased trastuzumab-induced HER4 nuclear translocation and enhanced trastuzumab response. There was also increased nuclear HER4 staining in the tumours from BT474 xenograft mice and human patients treated with trastuzumab. Furthermore, nuclear HER4 predicted poor clinical response to trastuzumab monotherapy in patients undergoing a window study and was shown to be an independent poor prognostic factor in HER2 positive breast cancer. Our data suggest that HER4 plays a key role in relation to trastuzumab resistance in HER2 positive breast cancer. Therefore, our study provides novel findings that HER4 activation, cleavage and nuclear translocation influence trastuzumab sensitivity and resistance in HER2 positive breast cancer. Nuclear HER4 could be a potential prognostic and predictive biomarker and understanding the role of HER4 may provide strategies to overcome trastuzumab resistance in HER2 positive breast cancer.

Late-phase clinical trials investigating metformin as a cancer therapy are underway. However, there remains controversy as to the mode of action of metformin in tumors at clinical doses. We conducted a clinical study integrating measurement of markers of systemic metabolism, dynamic FDG-PET-CT, transcriptomics, and metabolomics at paired time points to profile the bioactivity of metformin in primary breast cancer. We show metformin reduces the levels of mitochondrial metabolites, activates multiple mitochondrial metabolic pathways, and increases 18-FDG flux in tumors. Two tumor groups are identified with distinct metabolic responses, an OXPHOS transcriptional response (OTR) group for which there is an increase in OXPHOS gene transcription and an FDG response group with increased 18-FDG uptake. Increase in proliferation, as measured by a validated proliferation signature, suggested that patients in the OTR group were resistant to metformin treatment. We conclude that mitochondrial response to metformin in primary breast cancer may define anti-tumor effect.