Arterial specification precedes microvascular restitution in the peri-infarct cortex that is driven by small microvessels.

Evaluation of microvascular networks was impeded until recently by the need of histological tissue sectioning, which precluded 3D analyses. Using light-sheet microscopy, we investigated microvascular network characteristics in the peri-infarct cortex of mice 3-56 days after transient middle cerebral artery occlusion. In animal subgroups, the sphingosine-1-phosphate analog FTY720 (Fingolimod) was administered starting 24 hours post-ischemia. Light-sheet microscopy revealed a striking pattern of microvascular changes in the peri-infarct cortex, that is, a loss of microvessels, which was most prominent after 7 days and followed by the reappearance of microvessels over 56 days which revealed an increased branching point density and shortened branches. Using a novel AI-based image analysis algorithm we found that the length density of microvessels expressing the arterial specification marker α-smooth muscle actin markedly increased in the peri-infarct cortex already at 7 days post-ischemia. The length and branch density of small microvessels, but not of intermediate or large microvessels increased above pre-ischemic levels within 14-56 days. FTY720 increased the length and branch density of small microvessels. This study demonstrates long-term alterations of microvascular architecture post-ischemia indicative of increased collateralization most notably of small microvessels. Light-sheet microscopy will greatly advance the assessment of microvascular responses to restorative stroke therapies.

Microvascular Network Remodeling in the Ischemic Mouse Brain Defined by Light Sheet Microscopy.

BACKGROUND: Until now, the analysis of microvascular networks in the reperfused ischemic brain has been limited due to tissue transparency challenges. METHODS: Using light sheet microscopy, we assessed microvascular network remodeling in the striatum from 3 hours to 56 days post-ischemia in 2 mouse models of transient middle cerebral artery occlusion lasting 20 or 40 minutes, resulting in mild ischemic brain injury or brain infarction, respectively. We also examined the effect of a clinically applicable S1P (sphingosine-1-phosphate) analog, FTY720 (fingolimod), on microvascular network remodeling. RESULTS: Over 56 days, we observed progressive microvascular degeneration in the reperfused striatum, that is, the lesion core, which was followed by robust angiogenesis after mild ischemic injury induced by 20-minute middle cerebral artery occlusion. However, more severe ischemic injury elicited by 40-minute middle cerebral artery occlusion resulted in incomplete microvascular remodeling. In both cases, microvascular networks did not return to their preischemic state but displayed a chronically altered pattern characterized by higher branching point density, shorter branches, higher unconnected branch density, and lower tortuosity, indicating enhanced network connectivity. FTY720 effectively increased microvascular length density, branching point density, and volume density in both models, indicating an angiogenic effect of this drug. CONCLUSIONS: Utilizing light sheet microscopy together with automated image analysis, we characterized microvascular remodeling in the ischemic lesion core in unprecedented detail. This technology will significantly advance our understanding of microvascular restorative processes and pave the way for novel treatment developments in the stroke field.

Dimensionality reduction for deep learning in infrared microscopy: a comparative computational survey.

While infrared microscopy provides molecular information at spatial resolution in a label-free manner, exploiting both spatial and molecular information for classifying the disease status of tissue samples constitutes a major challenge. One strategy to mitigate this problem is to embed high-dimensional pixel spectra in lower dimensions, aiming to preserve molecular information in a more compact manner, which reduces the amount of data and promises to make subsequent disease classification more accessible for machine learning procedures. In this study, we compare several dimensionality reduction approaches and their effect on identifying cancer in the context of a colon carcinoma study. We observe surprisingly small differences between convolutional neural networks trained on dimensionality reduced spectra compared to utilizing full spectra, indicating a clear tendency of the convolutional networks to focus on spatial rather than spectral information for classifying disease status.

The impact of biostatistics on hazard characterization using in vitro developmental neurotoxicity assays.

In chemical safety assessment, benchmark concentrations (BMC) and their associated uncertainty are needed for the toxicological evaluation of in vitro data sets. A BMC estimation is derived from concentration-response modelling and results from various statistical decisions, which depend on factors such as experimental design and assay endpoint features. In current data practice, the experimenter is often responsible for the data analysis and therefore relies on statistical software, often without being aware of the software default settings and how they can impact the outputs of data analysis. To provide more insight into how statistical decision-making can influence the outcomes of data analysis and interpretation, we have developed an automated platform that includes statistical methods for BMC estimation, a novel endpoint-specific hazard classification system, and routines that flag data sets that are outside the applicability domain for an automatic data evaluation. We used case studies on a large dataset produced by a developmental neurotoxicity (DNT) in vitro battery (DNT IVB). Here we focused on the BMC and its confidence interval (CI) estimation as well as on final hazard classification. We identified five crucial statistical decisions the experimenter must make during data analysis: choice of replicate averaging, response data normalization, regression modelling, BMC and CI estimation, and choice of benchmark response levels. The insights gained are intended to raise more awareness among experimenters on the importance of statistical decisions and methods but also to demonstrate how important fit-for-purpose, internationally harmonized and accepted data evaluation and analysis procedures are for objective hazard classification.

Rapid and fully automated blood vasculature analysis in 3D light-sheet image volumes of different organs.

Light-sheet fluorescence microscopy (LSFM) can produce high-resolution tomograms of tissue vasculature with high accuracy. However, data processing and analysis is laborious due to the size of the datasets. Here, we introduce VesselExpress, an automated software that reliably analyzes six characteristic vascular network parameters including vessel diameter in LSFM data on average computing hardware. VesselExpress is ∼100 times faster than other existing vessel analysis tools, requires no user interaction, and integrates batch processing and parallelization. Employing an innovative dual Frangi filter approach, we show that obesity induces a large-scale modulation of brain vasculature in mice and that seven other major organs differ strongly in their 3D vascular makeup. Hence, VesselExpress transforms LSFM from an observational to an analytical working tool.

Data-driven subtyping of Parkinson's disease: comparison of current methodologies and application to the Bochum PNS cohort.

Considerable efforts have been made to better describe and identify Parkinson's disease (PD) subtypes. Cluster analyses have been proposed as an unbiased development approach for PD subtypes that could facilitate their identification, tracking of progression, and evaluation of therapeutic responses. A data-driven clustering analysis was applied to a PD cohort of 114 subjects enrolled at St. Josef-Hospital of the Ruhr University in Bochum (Germany). A wide spectrum of motor and non-motor scores including polyneuropathy-related measures was included into the analysis. K-means and hierarchical agglomerative clustering were performed to identify PD subtypes. Silhouette and Calinski-Harabasz Score Elbow were then employed as supporting evaluation metrics for determining the optimal number of clusters. Principal Component Analysis (PCA), analysis of variance (ANOVA), and analysis of covariance (ANCOVA) were conducted to determine the relevance of each score for the clusters' definition. Three PD cluster subtypes were identified: early onset mild type, intermediate type, and late-onset severe type. The between-cluster analysis consistently showed highly significant differences (P < 0.01), except for one of the scores measuring polyneuropathy (Neuropathy Disability Score; P = 0.609) and Levodopa dosage (P = 0.226). Parkinson's Disease Questionnaire (PDQ-39), Non-motor Symptom Questionnaire (NMSQuest), and the MDS-UPDRS Part II were found to be crucial factors for PD subtype differentiation. The present analysis identifies a specific set of criteria for PD subtyping based on an extensive panel of clinical and paraclinical scores. This analysis provides a foundation for further development of PD subtyping, including k-means and hierarchical agglomerative clustering.Trial registration: DRKS00020752, February 7, 2020, retrospectively registered.

Fast and label-free automated detection of microsatellite status in early colon cancer using artificial intelligence integrated infrared imaging.

PURPOSE: Microsatellite instability (MSI) due to mismatch repair (MMR) defects accounts for 15-20% of colon cancers (CC). MSI testing is currently standard of care in CC with immunohistochemistry of the four MMR proteins representing the gold standard. Instead, label-free quantum cascade laser (QCL) based infrared (IR) imaging combined with artificial intelligence (AI) may classify MSI/microsatellite stability (MSS) in unstained tissue sections user-independently and tissue preserving. METHODS: Paraffin-embedded unstained tissue sections of early CC from patients participating in the multicentre AIO ColoPredict Plus (CPP) 2.0 registry were analysed after dividing into three groups (training, test, and validation). IR images of tissue sections using QCL-IR microscopes were classified by AI (convolutional neural networks [CNN]) using a two-step approach. The first CNN (modified U-Net) detected areas of cancer while the second CNN (VGG-Net) classified MSI/MSS. End-points were area under receiver operating characteristic (AUROC) and area under precision recall curve (AUPRC). RESULTS: The cancer detection in the first step was based on 629 patients (train n = 273, test n = 138, and validation n = 218). Resulting classification AUROC was 1.0 for the validation dataset. The second step classifying MSI/MSS was performed on 547 patients (train n = 331, test n = 69, and validation n = 147) reaching AUROC and AUPRC of 0.9 and 0.74, respectively, for the validation cohort. CONCLUSION: Our novel label-free digital pathology approach accurately and rapidly classifies MSI vs. MSS. The tissue sections analysed were not processed leaving the sample unmodified for subsequent analyses. Our approach demonstrates an AI-based decision support tool potentially driving improved patient stratification and precision oncology in the future.

Establishment of a human cell-based in vitro battery to assess developmental neurotoxicity hazard of chemicals.

Developmental neurotoxicity (DNT) is a major safety concern for all chemicals of the human exposome. However, DNT data from animal studies are available for only a small percentage of manufactured compounds. Test methods with a higher throughput than current regulatory guideline methods, and with improved human relevance are urgently needed. We therefore explored the feasibility of DNT hazard assessment based on new approach methods (NAMs). An in vitro battery (IVB) was assembled from ten individual NAMs that had been developed during the past years to probe effects of chemicals on various fundamental neurodevelopmental processes. All assays used human neural cells at different developmental stages. This allowed us to assess disturbances of: (i) proliferation of neural progenitor cells (NPC); (ii) migration of neural crest cells, radial glia cells, neurons and oligodendrocytes; (iii) differentiation of NPC into neurons and oligodendrocytes; and (iv) neurite outgrowth of peripheral and central neurons. In parallel, cytotoxicity measures were obtained. The feasibility of concentration-dependent screening and of a reliable biostatistical processing of the complex multi-dimensional data was explored with a set of 120 test compounds, containing subsets of pre-defined positive and negative DNT compounds. The battery provided alerts (hit or borderline) for 24 of 28 known toxicants (82% sensitivity), and for none of the 17 negative controls. Based on the results from this screen project, strategies were developed on how IVB data may be used in the context of risk assessment scenarios employing integrated approaches for testing and assessment (IATA).

A framework for falsifiable explanations of machine learning models with an application in computational pathology.

In recent years, deep learning has been the key driver of breakthrough developments in computational pathology and other image based approaches that support medical diagnosis and treatment. The underlying neural networks as inherent black boxes lack transparency and are often accompanied by approaches to explain their output. However, formally defining explainability has been a notorious unsolved riddle. Here, we introduce a hypothesis-based framework for falsifiable explanations of machine learning models. A falsifiable explanation is a hypothesis that connects an intermediate space induced by the model with the sample from which the data originate. We instantiate this framework in a computational pathology setting using hyperspectral infrared microscopy. The intermediate space is an activation map, which is trained with an inductive bias to localize tumor. An explanation is constituted by hypothesizing that activation corresponds to tumor and associated structures, which we validate by histological staining as an independent secondary experiment.

Inverse folding based pre-training for the reliable identification of intrinsic transcription terminators.

It is well-established that neural networks can predict or identify structural motifs of non-coding RNAs (ncRNAs). Yet, the neural network based identification of RNA structural motifs is limited by the availability of training data that are often insufficient for learning features of specific ncRNA families or structural motifs. Aiming to reliably identify intrinsic transcription terminators in bacteria, we introduce a novel pre-training approach that uses inverse folding to generate training data for predicting or identifying a specific family or structural motif of ncRNA. We assess the ability of neural networks to identify secondary structure by systematic in silico mutagenesis experiments. In a study to identify intrinsic transcription terminators as functionally well-understood RNA structural motifs, our inverse folding based pre-training approach significantly boosts the performance of neural network topologies, which outperform previous approaches to identify intrinsic transcription terminators. Inverse-folding based pre-training provides a simple, yet highly effective way to integrate the well-established thermodynamic energy model into deep neural networks for identifying ncRNA families or motifs. The pre-training technique is broadly applicable to a range of network topologies as well as different types of ncRNA families and motifs.

Reliable identification and quantification of neural cells in microscopic images of neurospheres.

Neurosphere cultures consisting of primary human neural stem/progenitor cells (hNPC) are used for studying the effects of substances on early neurodevelopmental processes in vitro. Differentiating hNPCs migrate and differentiate into radial glia, neurons, astrocytes, and oligodendrocytes upon plating on a suitable extracellular matrix and thus model processes of early neural development. In order to characterize alterations in hNPC development, it is thus an essential task to reliably identify the cell type of each migrated cell in the migration area of a neurosphere. To this end, we introduce and validate a deep learning approach for identifying and quantifying cell types in microscopic images of differentiated hNPC. As we demonstrate, our approach performs with high accuracy and is robust against typical potential confounders. We demonstrate that our deep learning approach reproduces the dose responses of well-established developmental neurotoxic compounds and controls, indicating its potential in medium or high throughput in vitro screening studies. Hence, our approach can be used for studying compound effects on neural differentiation processes in an automated and unbiased process.

Shape decomposition algorithms for laser capture microdissection.

BACKGROUND: In the context of biomarker discovery and molecular characterization of diseases, laser capture microdissection is a highly effective approach to extract disease-specific regions from complex, heterogeneous tissue samples. For the extraction to be successful, these regions have to satisfy certain constraints in size and shape and thus have to be decomposed into feasible fragments. RESULTS: We model this problem of constrained shape decomposition as the computation of optimal feasible decompositions of simple polygons. We use a skeleton-based approach and present an algorithmic framework that allows the implementation of various feasibility criteria as well as optimization goals. Motivated by our application, we consider different constraints and examine the resulting fragmentations. We evaluate our algorithm on lung tissue samples in comparison to a heuristic decomposition approach. Our method achieved a success rate of over 95% in the microdissection and tissue yield was increased by 10-30%. CONCLUSION: We present a novel approach for constrained shape decomposition by demonstrating its advantages for the application in the microdissection of tissue samples. In comparison to the previous decomposition approach, the proposed method considerably increases the amount of successfully dissected tissue.

A Propagated Skeleton Approach to High Throughput Screening of Neurite Outgrowth for In Vitro Parkinson's Disease Modelling.

Neuronal models of neurodegenerative diseases such as Parkinson's Disease (PD) are extensively studied in pathological and therapeutical research with neurite outgrowth being a core feature. Screening of neurite outgrowth enables characterization of various stimuli and therapeutic effects after lesion. In this study, we describe an autonomous computational assay for a high throughput skeletonization approach allowing for quantification of neurite outgrowth in large data sets from fluorescence microscopic imaging. Development and validation of the assay was conducted with differentiated SH-SY5Y cells and primary mesencephalic dopaminergic neurons (MDN) treated with the neurotoxic lesioning compound Rotenone. Results of manual annotation using NeuronJ and automated data were shown to correlate strongly (R2-value 0.9077 for SH-SY5Y cells and R2-value 0.9297 for MDN). Pooled linear regressions of results from SH-SY5Y cell image data could be integrated into an equation formula (y=0.5410·x+1792; y=0.8789·x+0.09191 for normalized results) with y depicting automated and x depicting manual data. This automated neurite length algorithm constitutes a valuable tool for modelling of neurite outgrowth that can be easily applied to evaluate therapeutic compounds with high throughput approaches.

A representation learning approach for recovering scatter-corrected spectra from Fourier-transform infrared spectra of tissue samples.

Infrared spectra obtained from cell or tissue specimen have commonly been observed to involve a significant degree of scattering effects, often Mie scattering, which probably overshadows biochemically relevant spectral information by a nonlinear, nonadditive spectral component in Fourier transform infrared (FTIR) spectroscopic measurements. Correspondingly, many successful machine learning approaches for FTIR spectra have relied on preprocessing procedures that computationally remove the scattering components from an infrared spectrum. We propose an approach to approximate this complex preprocessing function using deep neural networks. As we demonstrate, the resulting model is not just several orders of magnitudes faster, which is important for real-time clinical applications, but also generalizes strongly across different tissue types. Using Bayesian machine learning approaches, our approach unveils model uncertainty that coincides with a band shift in the amide I region that occurs when scattering is removed computationally based on an established physical model. Furthermore, our proposed method overcomes the trade-off between computation time and the corrected spectrum being biased towards an artificial reference spectrum.

Grayscale representation of infrared microscopy images by extended multiplicative signal correction for registration with histological images.

Fourier-transform infrared (FTIR) microspectroscopy is rounding the corner to become a label-free routine method for cancer diagnosis. In order to build infrared-spectral based classifiers, infrared images need to be registered with Hematoxylin and Eosin (H&E) stained histological images. While FTIR images have a deep spectral domain with thousands of channels carrying chemical and scatter information, the H&E images have only three color channels for each pixel and carry mainly morphological information. Therefore, image representations of infrared images are needed that match the morphological information in H&E images. In this paper, we propose a novel approach for representation of FTIR images based on extended multiplicative signal correction highlighting morphological features that showed to correlate well with morphological information in H&E images. Based on the obtained representations, we developed a strategy for global-to-local image registration for FTIR images and H&E stained histological images of parallel tissue sections.

Lead-seq: transcriptome-wide structure probing in vivo using lead(II) ions.

The dynamic conformation of RNA molecules within living cells is key to their function. Recent advances in probing the RNA structurome in vivo, including the use of SHAPE (Selective 2'-Hydroxyl Acylation analyzed by Primer Extension) or kethoxal reagents or DMS (dimethyl sulfate), provided unprecedented insights into the architecture of RNA molecules in the living cell. Here, we report the establishment of lead probing in a global RNA structuromics approach. In order to elucidate the transcriptome-wide RNA landscape in the enteric pathogen Yersinia pseudotuberculosis, we combined lead(II) acetate-mediated cleavage of single-stranded RNA regions with high-throughput sequencing. This new approach, termed 'Lead-seq', provides structural information independent of base identity. We show that the method recapitulates secondary structures of tRNAs, RNase P RNA, tmRNA, 16S rRNA and the rpsT 5'-untranslated region, and that it reveals global structural features of mRNAs. The application of Lead-seq to Y. pseudotuberculosis cells grown at two different temperatures unveiled the first temperature-responsive in vivo RNA structurome of a bacterial pathogen. The translation of candidate genes derived from this approach was confirmed to be temperature regulated. Overall, this study establishes Lead-seq as complementary approach to interrogate intracellular RNA structures on a global scale.

Deep representation learning for domain adaptable classification of infrared spectral imaging data.

MOTIVATION: Applying infrared microscopy in the context of tissue diagnostics heavily relies on computationally preprocessing the infrared pixel spectra that constitute an infrared microscopic image. Existing approaches involve physical models, which are non-linear in nature and lead to classifiers that do not generalize well, e.g. across different types of tissue preparation. Furthermore, existing preprocessing approaches involve iterative procedures that are computationally demanding, so that computation time required for preprocessing does not keep pace with recent progress in infrared microscopes which can capture whole-slide images within minutes. RESULTS: We investigate the application of stacked contractive autoencoders as an unsupervised approach to preprocess infrared microscopic pixel spectra, followed by supervised fine-tuning to obtain neural networks that can reliably resolve tissue structure. To validate the robustness of the resulting classifier, we demonstrate that a network trained on embedded tissue can be transferred to classify fresh frozen tissue. The features obtained from unsupervised pretraining thus generalize across the large spectral differences between embedded and fresh frozen tissue, where under previous approaches separate classifiers had to be trained from scratch. AVAILABILITY AND IMPLEMENTATION: Our implementation can be downloaded from https://github.com/arnrau/SCAE_IR_Spectral_Imaging. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Fast and Noninvasive Diagnosis of Cervical Cancer by Coherent Anti-Stokes Raman Scattering.

Cervical cancer is the fourth most common cancer in women worldwide, and early detection of its precancerous lesions can decrease mortality. Cytopathology, HPV testing, and histopathology are the most commonly used tools in clinical practice. However, these methods suffer from many limitations such as subjectivity, cost, and time. Therefore, there is an unmet clinical need to develop new noninvasive methods for the early detection of cervical cancer. Here, a novel noninvasive, fast, and label-free approach with high accuracy is presented using liquid-based cytology Pap smears. CARS and SHG/TPF imaging was performed at one wavenumber on the Pap smears from patients with specimens negative for intraepithelial lesions or malignancy (NILM), and low-grade (LSIL) and high-grade (HSIL) squamous intraepithelial lesions. The normal, LSIL, and HSIL cells were selected on the basis of the ratio of the nucleus to the cytoplasm and cell morphology. Raman spectral imaging of single cells from the same smears was also performed to provide integral biochemical information of cells. Deep convolutional neural networks (DCNNs) were trained independently with CARS, SHG/TPF, and Raman images, taking into account both morphotextural and spectral information. DCNNs based on CARS, SHG/TPF, or Raman images have discriminated between normal and cancerous Pap smears with 100% accuracy. These results demonstrate that CARS/SHG/TPF microscopy has a prospective use as a label-free imaging technique for the fast screening of a large number of cells in cytopathological samples.

Hierarchical deep convolutional neural networks combine spectral and spatial information for highly accurate Raman-microscopy-based cytopathology.

Hierarchical variants of so-called deep convolutional neural networks (DCNNs) have facilitated breakthrough results for numerous pattern recognition tasks in recent years. We assess the potential of these novel whole-image classifiers for Raman-microscopy-based cytopathology. Conceptually, DCNNs facilitate a flexible combination of spectral and spatial information for classifying cellular images as healthy or cancer-affected cells. As we demonstrate, this conceptual advantage translates into practice, where DCNNs exceed the accuracy of both conventional classifiers based on pixel spectra as well as classifiers based on morphological features extracted from Raman microscopic images. Remarkably, accuracies exceeding those of all previously proposed classifiers are obtained while using only a small fraction of the spectral information provided by the dataset. Overall, our results indicate a high potential for DCNNs in medical applications of not just Raman, but also infrared microscopy.