Cleo-Aron Weis

Multi-class texture analysis in colorectal cancer histology.

Automatic recognition of different tissue types in histological images is an essential part in the digital pathology toolbox. Texture analysis is commonly used to address this problem; mainly in the context of estimating the tumour/stroma ratio on histological samples. However, although histological images typically contain more than two tissue types, only few studies have addressed the multi-class problem. For colorectal cancer, one of the most prevalent tumour types, there are in fact no published results on multiclass texture separation. In this paper we present a new dataset of 5,000 histological images of human colorectal cancer including eight different types of tissue. We used this set to assess the classification performance of a wide range of texture descriptors and classifiers. As a result, we found an optimal classification strategy that markedly outperformed traditional methods, improving the state of the art for tumour-stroma separation from 96.9% to 98.6% accuracy and setting a new standard for multiclass tissue separation (87.4% accuracy for eight classes). We make our dataset of histological images publicly available under a Creative Commons license and encourage other researchers to use it as a benchmark for their studies.

Thymoma related myasthenia gravis in humans and potential animal models

Thymoma-associated Myasthenia gravis (TAMG) is one of the anti-acetylcholine receptor MG (AChR-MG) subtypes. The clinico-pathological features of TAMG and its pathogenesis are described here in comparison with pathogenetic models suggested for the more common non-thymoma AChR-MG subtypes, early onset MG and late onset MG. Emphasis is put on the role of abnormal intratumorous T cell selection and activation, lack of intratumorous myoid cells and regulatory T cells as well as deficient expression of the autoimmune regulator (AIRE) by neoplastic thymic epithelial cells. We review spontaneous and genetically engineered thymoma models in a spectrum of animals and the extensive clinical and immunological overlap between canine, feline and human TAMG. Finally, limitations and perspectives of the transplantation of human and murine thymoma tissue into nude mice, as potential models for TAMG, are addressed.

In silico modeling of immunotherapy and stroma-targeting therapies in human colorectal cancer

Despite the fact that the local immunological microenvironment shapes the prognosis of colorectal cancer, immunotherapy has shown no benefit for the vast majority of colorectal cancer patients. A better understanding of the complex immunological interplay within the microenvironment is required. In this study, we utilized wet lab migration experiments and quantitative histological data of human colorectal cancer tissue samples (n = 20) including tumor cells, lymphocytes, stroma, and necrosis to generate a multiagent spatial model. The resulting data accurately reflected a wide range of situations of successful and failed immune surveillance. Validation of simulated tissue outcomes on an independent set of human colorectal cancer specimens (n = 37) revealed the model recapitulated the spatial layout typically found in human tumors. Stroma slowed down tumor growth in a lymphocyte-deprived environment but promoted immune escape in a lymphocyte-enriched environment. A subgroup of tumors with less stroma and high numbers of immune cells showed high rates of tumor control. These findings were validated using data from colorectal cancer patients (n = 261). Low-density stroma and high lymphocyte levels showed increased overall survival (hazard ratio 0.322, P = 0.0219) as compared with high stroma and high lymphocyte levels. To guide immunotherapy in colorectal cancer, simulation of immunotherapy in preestablished tumors showed that a complex landscape with optimal stroma permeabilization and immune cell activation is able to markedly increase therapy response in silico These results can help guide the rational design of complex therapeutic interventions, which target the colorectal cancer microenvironment. Cancer Res; 77(22); 6442-52. ©2017 AACR.

New Colors for Histology: Optimized Bivariate Color Maps Increase Perceptual Contrast in Histological Images.

Background Accurate evaluation of immunostained histological images is required for reproducible research in many different areas and forms the basis of many clinical decisions. The quality and efficiency of histopathological evaluation is limited by the information content of a histological image, which is primarily encoded as perceivable contrast differences between objects in the image. However, the colors of chromogen and counterstain used for histological samples are not always optimally distinguishable, even under optimal conditions. Methods and Results In this study, we present a method to extract the bivariate color map inherent in a given histological image and to retrospectively optimize this color map. We use a novel, unsupervised approach based on color deconvolution and principal component analysis to show that the commonly used blue and brown color hues in Hematoxylin—3,3’-Diaminobenzidine (DAB) images are poorly suited for human observers. We then demonstrate that it is possible to construct improved color maps according to objective criteria and that these color maps can be used to digitally re-stain histological images. Validation To validate whether this procedure improves distinguishability of objects and background in histological images, we re-stain phantom images and N = 596 large histological images of immunostained samples of human solid tumors. We show that perceptual contrast is improved by a factor of 2.56 in phantom images and up to a factor of 2.17 in sets of histological tumor images. Context Thus, we provide an objective and reliable approach to measure object distinguishability in a given histological image and to maximize visual information available to a human observer. This method could easily be incorporated in digital pathology image viewing systems to improve accuracy and efficiency in research and diagnostics.

Identification of a characteristic vascular belt zone in human colorectal cancer

Blood vessels in cancer Intra-tumoral blood vessels are of supreme importance for tumor growth, metastasis and therapy. Yet, little is known about spatial distribution patterns of these vessels. Most experimental or theoretical tumor models implicitly assume that blood vessels are equally abundant in different parts of the tumor, which has far-reaching implications for chemotherapy and tumor metabolism. In contrast, based on histological observations, we hypothesized that blood vessels follow specific spatial distribution patterns in colorectal cancer tissue. We developed and applied a novel computational approach to identify spatial patterns of angiogenesis in histological whole-slide images of human colorectal cancer. A characteristic spatial pattern of blood vessels in colorectal cancer In 33 of 34 (97%) colorectal cancer primary tumors blood vessels were significantly aggregated in a sharply limited belt-like zone at the interface of tumor tissue to the intestinal lumen. In contrast, in 11 of 11 (100%) colorectal cancer liver metastases, a similar hypervascularized zone could be found at the boundary to surrounding liver tissue. Also, in an independent validation cohort, we found this vascular belt zone: 22 of 23 (96%) samples of primary tumors and 15 of 16 (94%) samples of liver metastases exhibited the above-mentioned spatial distribution. Summary and implications We report consistent spatial patterns of tumor vascularization that may have far-reaching implications for models of drug distribution, tumor metabolism and tumor growth: luminal hypervascularization in colorectal cancer primary tumors is a previously overlooked feature of cancer tissue. In colorectal cancer liver metastases, we describe a corresponding pattern at the invasive margin. These findings add another puzzle piece to the complex concept of tumor heterogeneity.

Diagnostic Performance of Contrast Enhanced Pulmonary Computed Tomography Angiography for the Detection of Angioinvasive Pulmonary Aspergillosis in Immunocompromised Patients

Invasive pulmonary aspergillosis (IPA) is one of the major complications in immunocompromised patients. The mainstay of diagnostic imaging is non-enhanced chest-computed-tomography (CT), for which various non-specific signs for IPA have been described. However, contrast-enhanced CT pulmonary angiography (CTPA) has shown promising results, as the vessel occlusion sign (VOS) seems to be more sensitive and specific for IPA in hematologic patients. The aim of this study was to evaluate the diagnostic accuracy of CTPA in a larger cohort including non-hematologic immunocompromised patients. CTPA studies of 78 consecutive immunocompromised patients with proven/probable IPA were analyzed. 45 immunocompromised patients without IPA served as a control group. Diagnostic performance of CTPA-detected VOS and of radiological signs that do not require contrast-media were analyzed. Of 12 evaluable radiological signs, five were found to be significantly associated with IPA. The VOS showed the highest diagnostic performance with a sensitivity of 0.94, specificity of 0.71 and a diagnostic odds-ratio of 36.8. Regression analysis revealed the two strongest independent radiological predictors for IPA to be the VOS and the halo sign. The VOS is highly suggestive for IPA in immunocompromised patients in general. Thus, contrast-enhanced CTPA superior over non-contrast_enhanced chest-CT in patients with suspected IPA.

On the representation of cells in bone marrow pathology by a scalar field: propagation through serial sections, co-localization and spatial interaction analysis

BackgroundImmunohistochemical analysis of cellular interactions in the bone marrow in situ is demanding, due to its heterogeneous cellular composition, the poor delineation and overlap of functional compartments and highly complex immunophenotypes of several cell populations (e.g. regulatory T-cells) that require immunohistochemical marker sets for unambiguous characterization. To overcome these difficulties, we herein present an approach to describe objects (e.g. cells, bone trabeculae) by a scalar field that can be propagated through registered images of serial histological sections.MethodsThe transformation of objects within images (e.g. cells) to a scalar field was performed by convolution of the object’s centroids with differently formed radial basis function (e.g. for direct or indirect spatial interaction). On the basis of such a scalar field, a summation field described distributed objects within an image.ResultsAfter image registration i) colocalization analysis could be performed on basis scalar field, which is propagated through registered images, and - due to the shape of the field – were barely prone to matching errors and morphological changes by different cutting levels; ii) furthermore, depending on the field shape the colocalization measurements could also quantify spatial interaction (e.g. direct or paracrine cellular contact); ii) the field-overlap, which represents the spatial distance, of different objects (e.g. two cells) could be calculated by the histogram intersection.ConclusionsThe description of objects (e.g. cells, cell clusters, bone trabeculae etc.) as a field offers several possibilities: First, co-localization of different markers (e.g. by immunohistochemical staining) in serial sections can be performed in an automatic, objective and quantifiable way. In contrast to multicolour staining (e.g. 10-colour immunofluorescence) the financial and technical requirements are fairly minor. Second, the approach allows searching for different types of spatial interactions (e.g. direct and indirect cellular interaction) between objects by taking field shape into account (e.g. thin vs. broad). Third, by describing spatially distributed groups of objects as summation field, it gives cluster definition that relies rather on the bare object distance than on the modelled spatial cellular interaction.

Prognostic Value of Molecular Breast Cancer Subtypes based on Her2, ESR1, PGR and Ki67 mRNA-Expression in Muscle Invasive Bladder Cancer

INTRODUCTION: Gene expression analyses have identified similarities between bladder and breast cancer, where clinical risk stratification is based on Her2, ESR1, PGR and Ki67 expression. The aim of the study was to assess the respective marker gene expression in patients treated with radical cystectomy for muscle-invasive bladder cancer (MIBC) and to evaluate the applicability of breast cancer subtypes for MIBC risk stratification. MATERIALS & METHODS: 102 patients treated with radical cystectomy for MIBC were assessed. Using routine FFPE tissue and an IVD validated kit, mRNA expression was measured by single step RT-qPCR. Partition test were employed to define cut-off values for high or low marker gene expression. Association of expression with outcome was assessed using Kaplan-Meier analysis and multivariate cox regression analysis. Finally, we performed validation of our results in the MD-Anderson cohort (n = 57). RESULTS: Cancer specific survival (CSS) was impaired in patients with high gene expression of Her2 (P = 0.0009) and ESR1 (P = 0.04). In the multivariate regression model Her2 expression remained significant for the prediction of CSS (HR = 2.11, CI 1.11-4.21, P = 0.024). Furthermore, molecular stratification by breast cancer subgroups was significant (P = 0.023) for CSS prediction. Especially the differentiation between Her2-positive and Luminal A (HR = 4.41, CI 1.53-18.71, P = 0.004) and Luminal B (HR = 1.96, CI 0.99-4.08, P = 0.053) respectively was an independent prognostic parameter for CSS. External validation resulted in comparable risk stratification with differences in fractional subgroups distribution. CONCLUSION: Gene expression of Her2, ESR1, PGR, Ki67 and corresponding breast cancer subtypes allow a risk-stratification in MIBC, whereby Her2 overexpressing tumors reveal a particularly poor prognosis.

Histopathology of thymectomy specimens from the MGTX-trial: Entropy analysis as strategy to quantify spatial heterogeneity of lymphoid follicle and fat distribution

The thymectomy specimens from the “thymectomy trial in non-thymomatous myasthenia gravis patients receiving prednisone therapy” (MGTX) underwent rigid and comprehensive work-up, which permits analysis of the spatial distribution of histological and immunohistological features. This analysis revealed strong intra- and inter-case variability. While many histological features (e.g. median percent fat content among different specimens) can easily be correlated with clinical parameters, intra-case spatial variability of histological features has yet defied quantification and statistical evaluation. To overcome this gap in digital pathology, we here propose intra-case entropy of measured histological features in all available slides of a given thymectomy specimen as a quantitative marker of spatial histological heterogeneity. Calculation of entropy led to one value per specimen and histological feature. Through these ‘entropy values’ the so far neglected degree of spatial histological heterogeneity could be fed into statistical analyses, extending the scope of clinico-pathological correlations.

Challenging the current model of early‐onset myasthenia gravis pathogenesis in the light of the MGTX trial and histological heterogeneity of thymectomy specimens

The MGTX trial provided evidence that, in general, thymectomy is beneficial in adult patients up to 60 years of age with anti‐acetylcholine receptor–positive, nonthymomatous myasthenia gravis (MG). This finding supports the long‐held view that the pathogenesis of this type of MG (early‐onset MG (EOMG)) starts inside the thymus, results in the long‐term intrathymic recruitment of autoantibody‐producing B cells and plasma cells, and eventually spreads to the peripheral immune system. However, observed clinical responses to treatment in the MGTX trial were diverse. This might be due to heterogeneous epidemiological and genetic features of EOMG patients and variable durations of corticosteroid treatment before surgery, including a paucity of patients that were corticosteroid naive. Furthermore, the observed histological heterogeneity suggests that a single pathogenetic model may not fully reflect the spectrum of events that modify the course of EOMG. Here, we describe the morphology of the normal and MG‐associated thymus, how to evaluate morphological changes, and the current pathogenetic model of EOMG and discuss how it could be refined by integrating MGTX‐derived histological findings in thymectomy specimens and associated clinical observations.