Search

High computational cost associated with digital pathology image analysis approaches is a challenge towards their translation in routine pathology clinic. Here, we propose a computationally efficient framework (SuperHistopath), designed to map global context features reflecting the rich tumor morphological heterogeneity. SuperHistopath efficiently combines i) a segmentation approach using the linear iterative clustering (SLIC) superpixels algorithm applied directly on the whole-slide images at low resolution (5x magnification) to adhere to region boundaries and form homogeneous spatial units at tissue-level, followed by ii) classification of superpixels using a convolution neural network (CNN). To demonstrate how versatile SuperHistopath was in accomplishing histopathology tasks, we classified tumor tissue, stroma, necrosis, lymphocytes clusters, differentiating regions, fat, hemorrhage and normal tissue, in 127 melanomas, 23 triple-negative breast cancers, and 73 samples from transgenic mouse models of high-risk childhood neuroblastoma with high accuracy (98.8%, 93.1% and 98.3% respectively). Furthermore, SuperHistopath enabled discovery of significant differences in tumor phenotype of neuroblastoma mouse models emulating genomic variants of high-risk disease, and stratification of melanoma patients (high ratio of lymphocyte-to-tumor superpixels (p = 0.015) and low stroma-to-tumor ratio (p = 0.028) were associated with a favorable prognosis). Finally, SuperHistopath is efficient for annotation of ground-truth datasets (as there is no need of boundary delineation), training and application (~5 min for classifying a whole-slide image and as low as ~30 min for network training). These attributes make SuperHistopath particularly attractive for research in rich datasets and could also facilitate its adoption in the clinic to accelerate pathologist workflow with the quantification of phenotypes, predictive/prognosis markers. [ABSTRACT FROM AUTHOR]

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. [ABSTRACT FROM AUTHOR]

Simple Summary: Invasive lobular breast cancers (ILCs) are histologically classified by their discohesive growth pattern, due to loss of the cell adhesion glycoprotein E-cadherin (CDH1), which arises via mutation in CDH1 in around half of these tumours. A subset of these tumours, however, show mixed levels of E-cadherin expression. Here, we sought to address whether the distinct parts of individual tumours showing heterogeneous E-cadherin expression harbour distinct driver alterations. Using whole genome sequencing and methylation profiling of nine such cases, we identified that these tumours are clonally related, suggesting that they are part of the spectrum of ILC tumours. CDH1 mutant tumours showed a higher mutational burden indicative of APOBEC-mediated mutagenesis. In some cases, known clinically actionable driver mutations, such as PIK3CA, were exclusive to one component. Together, these results highlight the heterogeneity underpinning this special histological breast cancer. Mutations and loss of E-cadherin protein expression define the vast majority of invasive lobular carcinomas. In a subset of these cases, the heterogeneous expression of E-cadherin is observed either as wild-type (strong membranous) expression or aberrant expression (cytoplasmic expression). However, it is unclear as to whether the two components would be driven by distinct genetic or epigenetic alterations. Here, we used whole genome DNA sequencing and methylation array profiling of two separately dissected components of nine invasive lobular carcinomas with heterogeneous E-cadherin expression. E-cadherin negative and aberrant/positive components of E-cadherin heterogeneous tumours showed a similar mutational, copy number and promoter methylation repertoire, suggesting they arise from a common ancestor, as opposed to the collision of two independent tumours. We found that the majority of E-cadherin heterogeneous tumours harboured CDH1 mutations in both the E-cadherin negative and aberrant/positive components together with somatic mutations in additional driver genes known to be enriched in both pure invasive carcinomas of no special type and invasive lobular breast cancers, whereas these were less commonly observed in CDH1 wild-type tumours. CDH1 mutant tumours also exhibited a higher mutation burden as well as increased presence of APOBEC-dependent mutational signatures 2 and 13 compared to CDH1 wild-type tumours. Together, our results suggest that regardless of E-cadherin protein expression, tumours showing heterogeneous expression of E-cadherin should be considered as part of the spectrum of invasive lobular breast cancers. [ABSTRACT FROM AUTHOR]

Simple Summary: Ovarian clear cell carcinoma (OCCC) is a rare subtype of epithelial ovarian cancer that has a poor response to chemotherapy. Here, we assessed the immunological features of a series of 33 OCCCs and identified an immune-related gene expression signature that correlated with a patient's risk of recurrence. Additionally, using multiplex immunofluorescence, we assessed the spatial distribution and abundance of immune cell populations at the protein level and identified that tumour-associated macrophages (TAM) and regulatory T cells are excluded from the vicinity of tumour cells in low-risk patients, suggesting that high-risk patients have a more immunosuppressive microenvironment. We also found that TAMs and cytotoxic T cells were also excluded from the vicinity of tumour cells in ARID1A mutated OCCCs, suggesting that the exclusion of these immune effectors could determine the host response in ARID1A mutant OCCCs. Ovarian clear cell carcinoma (OCCC) is a rare subtype of epithelial ovarian cancer characterised by a high frequency of loss-of-function ARID1A mutations and a poor response to chemotherapy. Despite their generally low mutational burden, an intratumoural T cell response has been reported in a subset of OCCC, with ARID1A purported to be a biomarker for the response to the immune checkpoint blockade independent of micro-satellite instability (MSI). However, assessment of the different immune cell types and spatial distribution specifically within OCCC patients has not been described to date. Here, we characterised the immune landscape of OCCC by profiling a cohort of 33 microsatellite stable OCCCs at the genomic, gene expression and histological level using targeted sequencing, gene expression profiling using the NanoString targeted immune panel, and multiplex immunofluorescence to assess the spatial distribution and abundance of immune cell populations at the protein level. Analysis of these tumours and subsequent independent validation identified an immune-related gene expression signature associated with risk of recurrence of OCCC. Whilst histological quantification of tumour-infiltrating lymphocytes (TIL, Salgado scoring) showed no association with the risk of recurrence or ARID1A mutational status, the characterisation of TILs via multiplexed immunofluorescence identified spatial differences in immunosuppressive cell populations in OCCC. Tumour-associated macrophages (TAM) and regulatory T cells were excluded from the vicinity of tumour cells in low-risk patients, suggesting that high-risk patients have a more immunosuppressive microenvironment. We also found that TAMs and cytotoxic T cells were also excluded from the vicinity of tumour cells in ARID1A-mutated OCCCs compared to ARID1A wild-type tumours, suggesting that the exclusion of these immune effectors could determine the host response of ARID1A-mutant OCCCs to therapy. Overall, our study has provided new insights into the immune landscape and prognostic associations in OCCC and suggest that tailored immunotherapeutic approaches may be warranted for different subgroups of OCCC patients. [ABSTRACT FROM AUTHOR]

Tumor-infiltrating lymphocytes (TILs) have prognostic significance in several cancers, including breast cancer. Despite interest in combining radiation therapy with immunotherapy, little is known about the effect of radiation therapy itself on the tumor-immune microenvironment, including TILs. Here, we interrogated longitudinal dynamics of TILs and systemic lymphocytes in patient samples taken before, during, and after neoadjuvant radiation therapy (NART) from PRADA and Neo-RT breast clinical trials. We manually scored stromal TILs (sTILs) from longitudinal tumor samples using standardized guidelines as well as deep learning–based scores at cell-level (cTIL) and cell- and tissue-level combination analyses (SuperTIL). In parallel, we interrogated absolute lymphocyte counts from routine blood tests at corresponding time points during treatment. Exploratory analyses studied the relationship between TILs and pathologic complete response (pCR) and long-term outcomes. Patients receiving NART experienced a significant and uniform decrease in sTILs that did not recover at the time of surgery (P <.0001). This lymphodepletive effect was also mirrored in peripheral blood. Our SuperTIL deep learning score showed good concordance with manual sTILs and importantly performed comparably to manual scores in predicting pCR from diagnostic biopsies. The analysis suggested an association between baseline sTILs and pCR, as well as sTILs at surgery and relapse, in patients receiving NART. This study provides novel insights into TIL dynamics in the context of NART in breast cancer and demonstrates the potential for artificial intelligence to assist routine pathology. We have identified trends that warrant further interrogation and have a bearing on future radioimmunotherapy trials. [ABSTRACT FROM AUTHOR]

BACKGROUND. Phase 1 study of ATR inhibition alone or with radiation therapy (PATRIOT) was a first-in-human phase I study of the oral ATR (ataxia telangiectasia and Rad3-related) inhibitor ceralasertib (AZD6738) in advanced solid tumors. METHODS. The primary objective was safety. Secondary objectives included assessment of antitumor responses and pharmacokinetic (PK) and pharmacodynamic (PD) studies. Sixty-seven patients received 20--240 mg ceralasertib BD continuously or intermittently (14 of a 28-day cycle). RESULTS. Intermittent dosing was better tolerated than continuous, which was associated with dose-limiting hematological toxicity. The recommended phase 2 dose of ceralasertib was 160 mg twice daily for 2 weeks in a 4-weekly cycle. Modulation of target and increased DNA damage were identified in tumor and surrogate PD. There were 5 (8%) confirmed partial responses (PRs) (40--240 mg BD), 34 (52%) stable disease (SD), including 1 unconfirmed PR, and 27 (41%) progressive disease. Durable responses were seen in tumors with loss of AT-rich interactive domain-containing protein 1A (ARID1A) and DNA damage-- response defects. Treatment-modulated tumor and systemic immune markers and responding tumors were more immune inflamed than nonresponding. CONCLUSION. Ceralasertib monotherapy was tolerated at 160 mg BD intermittently and associated with antitumor activity. TRIAL REGISTRATION. Clinicaltrials.gov: NCT02223923, EudraCT: 2013-003994-84. FUNDING. Cancer Research UK, AstraZeneca, UK Department of Health (National Institute for Health Research), Rosetrees Trust, Experimental Cancer Medicine Centre. [ABSTRACT FROM AUTHOR]