Deep neural network based histological scoring of lung fibrosis and inflammation in the mouse model system

Ferrocene-derived magnetic fiber-particles from diesel exhaust: enhanced pulmonary toxicity via Bach1-SAT1-polyamine depletion

BACKGROUND AND AIM

Magnetic nanoparticles are key components of air pollution. The combustion of diesel engine fuels, especially with ferrocene doping to reduce emissions, may increase exposure to these particles and related health risks. This study aimed to reveal the generation and characterization of ferrocene-derived magnetic particles (FMP) in ferrocene-doped diesel exhaust, and to investigate its toxicities and associated mechanisms in an avian model.

METHODS

FMP was observed in ferrocene-doped diesel exhaust particles, and extracted with neodymium magnets. Extracted FMP was characterized, and exposed to hatchling chickens via aerosol inhalation. Pulmonary toxicities were assessed with pathological and molecular methods. Associated mechanisms were investigated with RNA-seq, in vitro cell culture, and in vivo gene silencing.

RESULTS

FMP was characterized to be fibrous, magnetic iron-containing carbon particles. Extracted FMP could directly induce pulmonary toxicity. Mechanistic investigations revealed molecular mechanism associated with ferroptosis via Bach1, SAT1 and polyamines depletion, and further confirmed with ferroptosis inhibitor treatment, Bach1 inhibitor treatment, supplementation of polyamines or SAT1 silencing.

CONCLUSIONS

Ferrocene doping could result in formation of magnetic particles in diesel exhaust. For the first time, magnetic fiber-like particles were extracted from ferrocene-doped DE particles, which is a potential source of magnetic particles in air pollution. To better balance emission control and health effects, further investigations are necessary.

Quantitative Assessment of Pulmonary Fibrosis in a Murine Model via a Multimodal Imaging Workflow

Disease-recapitulating animal models are valuable tools in preclinical development for the study of compounds. In the case of fibrotic pulmonary diseases such as idiopathic pulmonary fibrosis (IPF), the bleomycin model of lung injury in the mouse is widely used. To evaluate bleomycin-induced changes in the lung, we employed a quantitative, multimodal approach. Using in vivo microcomputed tomography (μCT), we demonstrated radiographic changes associated with disease progression in aeration levels of the lung parenchyma. There exists an unmet need for a quantitative, high-resolution imaging probe to detect pulmonary fibrosis, particularly that can differentiate between inflammatory and fibrotic components of the disease. Matrix remodeling and overexpression of extracellular matrix (ECM) proteins such as collagen and fibronectin are hallmarks of organ fibrosis. A splice variant of fibronectin containing extra domain A (FnEDA) is of particular interest in fibrosis due to its high level of expression in diseased tissue, which is confirmed here using immunohistochemistry (IHC) in mouse and human lungs. An antibody against FnEDA was evaluated for use as an imaging tool, particularly by using in vivo single-photon emission computed tomography (SPECT) and ex vivo near-infrared (NIR) fluorescence imaging. These data were further corroborated with histological tissue staining and fibrosis quantitation based on a Modified Ashcroft (MA) score and a digital image analysis of whole slide lung tissue sections. The fusion of these different approaches represents a robust integrated workflow combining anatomical and molecular imaging technologies to enable the visualization and quantitation of disease activity and treatment response with an inhibitor of the TGFβ signaling pathway.

Optimized digital workflow for pathologist-grade evaluation in bleomycin-induced pulmonary fibrosis mouse model

Idiopathic pulmonary fibrosis (IPF) is a progressive and ultimately fatal disorder of unknown etiology, characterized by interstitial fibrosis of the lungs. Bleomycin-induced pulmonary fibrosis mouse model (BLM model) is a widely used animal model to evaluate therapeutic targets for IPF. Histopathological analysis of lung fibrosis is an important method for evaluating BLM model. However, this method requires expertise in recognizing complex visual patterns and is time-consuming, making the workflow difficult and inefficient. Therefore, we developed a new workflow for BLM model that reduces inter- and intra-observer variations and improves the evaluation process. We generated deep learning models for grading lung fibrosis that were able to achieve accuracy comparable to that of pathologists. These models incorporate complex image patterns and qualitative factors, such as collagen texture and distribution, potentially identifying drug candidates overlooked in evaluations based solely on simple area extraction. This deep learning-based fibrosis grade assessment has the potential to streamline drug development for pulmonary fibrosis by offering higher granularity and reproducibility in evaluating BLM model.

MetFinder: A Tool for Automated Quantitation of Metastatic Burden in Histological Sections From Preclinical Models

As efforts to study the mechanisms of melanoma metastasis and novel therapeutic approaches multiply, researchers need accurate, high-throughput methods to evaluate the effects on tumor burden resulting from specific interventions. We show that automated quantification of tumor content from whole slide images is a compelling solution to assess in vivo experiments. In order to increase the outflow of data collection from preclinical studies, we assembled a large dataset with annotations and trained a deep neural network for the quantitative analysis of melanoma tumor content on histopathological sections of murine models. After assessing its performance in segmenting these images, the tool obtained consistent results with an orthogonal method (bioluminescence) of measuring metastasis in an experimental setting. This AI-based algorithm, made freely available to academic laboratories through a web-interface called MetFinder, promises to become an asset for melanoma researchers and pathologists interested in accurate, quantitative assessment of metastasis burden.

PDE4B inhibition by nerandomilast: Effects on lung fibrosis and transcriptome in fibrotic rats and on biomarkers in human lung epithelial cells

BACKGROUND AND PURPOSE

The PDE4 family is considered a prime target for therapeutic intervention in several fibro-inflammatory diseases. We have investigated the molecular mechanisms of nerandomilast (BI 1015550), a preferential PDE4B inhibitor.

EXPERIMENTAL APPROACH

In addition to clinically relevant parameters of idiopathic pulmonary fibrosis (IPF; lung function measurement/high-resolution computed tomography scan/AI-Ashcroft score), whole-lung homogenates from a therapeutic male Wistar rat model of pulmonary fibrosis were analysed by next-generation sequencing (NGS). Data were matched with public domain data derived from human IPF samples to investigate how well the rat model reflected human IPF. We scored the top counter-regulated genes following treatment with nerandomilast in human single cells and validated disease markers discovered in the rat model using a human disease-relevant in vitro assay of IPF.

KEY RESULTS

Nerandomilast improved the decline of lung function parameters in bleomycin-treated animals. In the NGS study, most transcripts deregulated by bleomycin treatment were normalised by nerandomilast treatment. Most notably, a significant number of deregulated transcripts that were identified in human IPF disease were also found in the animal model and reversed by nerandomilast. Mapping to single-cell data revealed the strongest effects on mesenchymal, epithelial and endothelial cell populations. In a primary human epithelial cell culture system, several disease-related (bio)markers were inhibited by nerandomilast in a concentration-dependent manner.

CONCLUSIONS AND IMPLICATIONS

This study further supports the available knowledge about the anti-inflammatory/antifibrotic mechanisms of nerandomilast and provides novel insights into the mode of action and signalling pathways influenced by nerandomilast treatment of lung fibrosis.

Quantification of fibrosis extend and airspace availability in lung: A semi-automatic ImageJ/Fiji toolbox

The evaluation of the structural integrity of mechanically dynamic organs such as lungs is critical for the diagnosis of numerous pathologies and the development of therapies. This task is classically performed by histology experts in a qualitative or semi-quantitative manner. Automatic digital image processing methods appeared in the last decades, and although immensely powerful, tools are highly specialized and lack the versatility required in various experimental designs. Here, a set of scripts for the image processing software ImageJ/Fiji to easily quantify fibrosis extend and alveolar airspace availability in Sirius Red or Masson's trichrome stained samples is presented. The toolbox consists in thirteen modules: sample detection, particles filtration (automatic and manual), border definition, air ducts identification, air ducts walls definition, parenchyma extraction, MT-staining specific pre-processing, fibrosis detection, fibrosis particles filtration, airspace detection, and visualizations (tissue only or tissue and airspace). While the process is largely automated, critical parameters are accessible to the user for increased adaptability. The modularity of the protocol allows for its adjustment to alternative experimental settings. Fibrosis and airspace can be combined as an evaluation of the structural integrity of the organ. All settings and intermediate states are saved to ensure reproducibility. These new analysis scripts allow for a rapid quantification of fibrosis and airspace in a large variety of experimental settings.

Monocyte-derived alveolar macrophages are key drivers of smoke-induced lung inflammation and tissue remodeling

Smoking is a leading risk factor of chronic obstructive pulmonary disease (COPD), that is characterized by chronic lung inflammation, tissue remodeling and emphysema. Although inflammation is critical to COPD pathogenesis, the cellular and molecular basis underlying smoking-induced lung inflammation and pathology remains unclear. Using murine smoke models and single-cell RNA-sequencing, we show that smoking establishes a self-amplifying inflammatory loop characterized by an influx of molecularly heterogeneous neutrophil subsets and excessive recruitment of monocyte-derived alveolar macrophages (MoAM). In contrast to tissue-resident AM, MoAM are absent in homeostasis and characterized by a pro-inflammatory gene signature. Moreover, MoAM represent 46% of AM in emphysematous mice and express markers causally linked to emphysema. We also demonstrate the presence of pro-inflammatory and tissue remodeling associated MoAM orthologs in humans that are significantly increased in emphysematous COPD patients. Inhibition of the IRAK4 kinase depletes a rare inflammatory neutrophil subset, diminishes MoAM recruitment, and alleviates inflammation in the lung of cigarette smoke-exposed mice. This study extends our understanding of the molecular signaling circuits and cellular dynamics in smoking-induced lung inflammation and pathology, highlights the functional consequence of monocyte and neutrophil recruitment, identifies MoAM as key drivers of the inflammatory process, and supports their contribution to pathological tissue remodeling.

The potential role of artificial intelligence in the clinical practice of interstitial lung disease

Artificial intelligence (AI) is being widely applied in the field of medicine, in areas such as drug discovery, diagnostic support, and assistance with medical practice. Among these, medical imaging is an area where AI is expected to make a significant contribution. In Japan, as of November 2022, 23 AI medical devices have received regulatory approval; all these devices are related to image analysis. In interstitial lung diseases, technologies have been developed that use AI to analyze high-resolution computed tomography and pathological images, and gene expression patterns in tissue taken from transbronchial lung biopsies to assist in the diagnosis of idiopathic pulmonary fibrosis. Some of these technologies are already being used in clinical practice in the United States. AI is expected to reduce the burden on physicians, improve reproducibility, and advance personalized medicine. Obtaining sufficient data for diseases with a small number of patients is difficult. Additionally, certain issues must be addressed in order for AI to be applied in healthcare. These issues include taking responsibility for the AI results output, updating software after the launch of technology, and adapting to new imaging technologies. Establishing research infrastructures such as large-scale databases and common platforms is important for the development of AI technology: their use requires an understanding of the characteristics and limitations of the systems. CLINICAL TRIAL REGISTRATION: Not applicable.

Quantifying acute kidney injury in an Ischaemia-Reperfusion Injury mouse model using deep-learning-based semantic segmentation in histology

This study focuses on ischaemia-reperfusion injury (IRI) in kidneys, a cause of acute kidney injury (AKI) and end-stage kidney disease (ESKD). Traditional kidney damage assessment methods are semi-quantitative and subjective. This study aims to use a convolutional neural network (CNN) to segment murine kidney structures after IRI, quantify damage via CNN-generated pathological measurements, and compare this to conventional scoring. The CNN was able to accurately segment the different pathological classes, such as Intratubular casts and Tubular necrosis, with an F1 score of over 0.75. Some classes, such as Glomeruli and Proximal tubules, had even higher statistical values with F1 scores over 0.90. The scoring generated based on the segmentation approach statistically correlated with the semiquantitative assessment (Spearman's rank correlation coefficient=0.94). The heatmap approach localised the intratubular necrosis mainly in the outer stripe of the outer medulla, while the tubular casts were also present in more superficial or deeper portions of the cortex and medullary areas. This study presents a CNN model capable of segmenting multiple classes of interest, including acute IRI-specific pathological changes, in a whole mouse kidney section and can provide insights into the distribution of pathological classes within the whole mouse kidney section.

The state of the art for artificial intelligence in lung digital pathology

Lung diseases carry a significant burden of morbidity and mortality worldwide. The advent of digital pathology (DP) and an increase in computational power have led to the development of artificial intelligence (AI)-based tools that can assist pathologists and pulmonologists in improving clinical workflow and patient management. While previous works have explored the advances in computational approaches for breast, prostate, and head and neck cancers, there has been a growing interest in applying these technologies to lung diseases as well. The application of AI tools on radiology images for better characterization of indeterminate lung nodules, fibrotic lung disease, and lung cancer risk stratification has been well documented. In this article, we discuss methodologies used to build AI tools in lung DP, describing the various hand-crafted and deep learning-based unsupervised feature approaches. Next, we review AI tools across a wide spectrum of lung diseases including cancer, tuberculosis, idiopathic pulmonary fibrosis, and COVID-19. We discuss the utility of novel imaging biomarkers for different types of clinical problems including quantification of biomarkers like PD-L1, lung disease diagnosis, risk stratification, and prediction of response to treatments such as immune checkpoint inhibitors. We also look briefly at some emerging applications of AI tools in lung DP such as multimodal data analysis, 3D pathology, and transplant rejection. Lastly, we discuss the future of DP-based AI tools, describing the challenges with regulatory approval, developing reimbursement models, planning clinical deployment, and addressing AI biases. © 2022 The Authors. The Journal of Pathology published by John Wiley & Sons Ltd on behalf of The Pathological Society of Great Britain and Ireland.

Artificial Intelligence-based Tumor Segmentation in Mouse Models of Lung Adenocarcinoma

BACKGROUND

Mouse models are highly effective for studying the pathophysiology of lung adenocarcinoma and evaluating new treatment strategies. Treatment efficacy is primarily determined by the total tumor burden measured on excised tumor specimens. The measurement process is time-consuming and prone to human errors. To address this issue, we developed a novel deep learning model to segment lung tumor foci on digitally scanned hematoxylin and eosin (H&E) histology slides.

METHODS

Digital slides of 239 mice from 9 experimental cohorts were split into training (n=137), validation (n=37), and testing cohorts (n=65). Image patches of 500×500 pixels were extracted from 5× and 10× magnifications, along with binary masks of expert annotations representing ground-truth tumor regions. Deep learning models utilizing DeepLabV3+ and UNet architectures were trained for binary segmentation of tumor foci under varying stain normalization conditions. The performance of algorithm segmentation was assessed by Dice Coefficient, and detection was evaluated by sensitivity and positive-predictive value (PPV).

RESULTS

The best model on patch-based validation was DeepLabV3+ using a Resnet-50 backbone, which achieved Dice 0.890 and 0.873 on validation and testing cohort, respectively. This result corresponded to 91.3 Sensitivity and 51.0 PPV in the validation cohort and 93.7 Sensitivity and 51.4 PPV in the testing cohort. False positives could be reduced 10-fold with thresholding artificial intelligence (AI) predicted output by area, without negative impact on Dice Coefficient. Evaluation at various stain normalization strategies did not demonstrate improvement from the baseline model.

CONCLUSIONS

A robust AI-based algorithm for detecting and segmenting lung tumor foci in the pre-clinical mouse models was developed. The output of this algorithm is compatible with open-source software that researchers commonly use.

An EZH2-dependent transcriptional complex promotes aberrant epithelial remodelling after injury

Unveiling the molecular mechanisms of tissue remodelling following injury is imperative to elucidate its regenerative capacity and aberrant repair in disease. Using different omics approaches, we identified enhancer of zester homolog 2 (EZH2) as a key regulator of fibrosis in injured lung epithelium. Epithelial injury drives an enrichment of nuclear transforming growth factor-β-activated kinase 1 (TAK1) that mediates EZH2 phosphorylation to facilitate its liberation from polycomb repressive complex 2 (PRC2). This process results in the establishment of a transcriptional complex of EZH2, RNA-polymerase II (POL2) and nuclear actin, which orchestrates aberrant epithelial repair programmes. The liberation of EZH2 from PRC2 is accompanied by an EZH2-EZH1 switch to preserve H3K27me3 deposition at non-target genes. Loss of epithelial TAK1, EZH2 or blocking nuclear actin influx attenuates the fibrotic cascade and restores respiratory homeostasis. Accordingly, EZH2 inhibition significantly improves outcomes in a pulmonary fibrosis mouse model. Our results reveal an important non-canonical function of EZH2, paving the way for new therapeutic interventions in fibrotic lung diseases.

Automated Digital Quantification of Pulmonary Fibrosis in Human Histopathology Specimens

Pulmonary fibrosis is characterized by abnormal interstitial extracellular matrix and cellular accumulations. Methods quantifying fibrosis severity in lung histopathology samples are semi-quantitative, subjective, and analyze only portions of sections. We sought to determine whether automated computerized imaging analysis shown to continuously measure fibrosis in mice could also be applied in human samples. A pilot study was conducted to analyze a small number of specimens from patients with Hermansky-Pudlak syndrome pulmonary fibrosis (HPSPF) or idiopathic pulmonary fibrosis (IPF). Digital images of entire lung histological serial sections stained with picrosirius red and alcian blue or anti-CD68 antibody were analyzed using dedicated software to automatically quantify fibrosis, collagen, and macrophage content. Automated fibrosis quantification based on parenchymal tissue density and fibrosis score measurements was compared to pulmonary function values or Ashcroft score. Automated fibrosis quantification of HPSPF lung explants was significantly higher than that of IPF lung explants or biopsies and was also significantly higher in IPF lung explants than in IPF biopsies. A high correlation coefficient was found between some automated quantification measurements and lung function values for the three sample groups. Automated quantification of collagen content in lung sections used for digital image analyses was similar in the three groups. CD68 immunolabeled cell measurements were significantly higher in HPSPF explants than in IPF biopsies. In conclusion, computerized image analysis provides access to accurate, reader-independent pulmonary fibrosis quantification in human histopathology samples. Fibrosis, collagen content, and immunostained cells can be automatically and individually quantified from serial sections. Robust automated digital image analysis of human lung samples enhances the available tools to quantify and study fibrotic lung disease.

Molecular Imaging of Inflammatory Disease

Inflammatory diseases include a wide variety of highly prevalent conditions with high mortality rates in severe cases ranging from cardiovascular disease, to rheumatoid arthritis, to chronic obstructive pulmonary disease, to graft vs. host disease, to a number of gastrointestinal disorders. Many diseases that are not considered inflammatory per se are associated with varying levels of inflammation. Imaging of the immune system and inflammatory response is of interest as it can give insight into disease progression and severity. Clinical imaging technologies such as computed tomography (CT) and magnetic resonance imaging (MRI) are traditionally limited to the visualization of anatomical information; then, the presence or absence of an inflammatory state must be inferred from the structural abnormalities. Improvement in available contrast agents has made it possible to obtain functional information as well as anatomical. In vivo imaging of inflammation ultimately facilitates an improved accuracy of diagnostics and monitoring of patients to allow for better patient care. Highly specific molecular imaging of inflammatory biomarkers allows for earlier diagnosis to prevent irreversible damage. Advancements in imaging instruments, targeted tracers, and contrast agents represent a rapidly growing area of preclinical research with the hopes of quick translation to the clinic.

Artificial Intelligence in Toxicologic Pathology: Quantitative Evaluation of Compound-Induced Hepatocellular Hypertrophy in Rats

Digital pathology evolved rapidly, enabling more systematic usage of image analysis and development of artificial intelligence (AI) applications. Here, combined AI models were developed to evaluate hepatocellular hypertrophy in rat liver, using commercial AI-based software on hematoxylin and eosin-stained whole slide images. In a first approach, deep learning-based identification of critical tissue zones (centrilobular, midzonal, and periportal) enabled evaluation of region-specific cell size. Mean cytoplasmic area of hepatocytes was calculated via several sequential algorithms including segmentation in microanatomical structures (separation of sinusoids and vessels from hepatocytes), nuclear detection, and area measurements. An increase in mean cytoplasmic area could be shown in groups given phenobarbital, known to induce hepatocellular hypertrophy when compared to control groups, in multiple studies. Quantitative results correlated with the gold standard: observation and grading performed by board-certified veterinary pathologists, liver weights, and gene expression. Furthermore, as a second approach, we introduce for the first time deep learning-based direct detection of hepatocellular hypertrophy with similar results. Cell hypertrophy is challenging to pick up, particularly in milder cases. Additional evaluation of mean cytoplasmic area or direct detection of hypertrophy, combined with histopathological observations and liver weights, is expected to increase accuracy and repeatability of diagnoses and grading by pathologists.

Deep Learning-Based Detection of Endothelial Tip Cells in the Oxygen-Induced Retinopathy Model

Proliferative retinopathies, such as diabetic retinopathy and retinopathy of prematurity, are leading causes of vision impairment. A common feature is a loss of retinal capillary vessels resulting in hypoxia and neuronal damage. The oxygen-induced retinopathy model is widely used to study revascularization of an ischemic area in the mouse retina. The presence of endothelial tip cells indicates vascular recovery; however, their quantification relies on manual counting in microscopy images of retinal flat mount preparations. Recent advances in deep neural networks (DNNs) allow the automation of such tasks. We demonstrate a workflow for detection of tip cells in retinal images using the DNN-based Single Shot Detector (SSD). The SSD was designed for detection of objects in natural images. We adapt the SSD architecture and training procedure to the tip cell detection task and retrain the DNN using labeled tip cells in images of fluorescently stained retina flat mounts. Transferring knowledge from the pretrained DNN and extensive data augmentation reduced the amount of required labeled data. Our system shows a performance comparable to the human level, while providing highly consistent results. Therefore, such a system can automate counting of tip cells, a readout frequently used in retinopathy research, thereby reducing routine work for biomedical experts.

Digital Image Analysis of Picrosirius Red Staining: A Robust Method for Multi-Organ Fibrosis Quantification and Characterization

Current understanding of fibrosis remains incomplete despite the increasing burden of related diseases. Preclinical models are used to dissect the pathogenesis and dynamics of fibrosis, and to evaluate anti-fibrotic therapies. These studies require objective and accurate measurements of fibrosis. Existing histological quantification methods are operator-dependent, organ-specific, and/or need advanced equipment. Therefore, we developed a robust, minimally operator-dependent, and tissue-transposable digital method for fibrosis quantification. The proposed method involves a novel algorithm for more specific and more sensitive detection of collagen fibers stained by picrosirius red (PSR), a computer-assisted segmentation of histological structures, and a new automated morphological classification of fibers according to their compactness. The new algorithm proved more accurate than classical filtering using principal color component (red-green-blue; RGB) for PSR detection. We applied this new method on established mouse models of liver, lung, and kidney fibrosis and demonstrated its validity by evidencing topological collagen accumulation in relevant histological compartments. Our data also showed an overall accumulation of compact fibers concomitant with worsening fibrosis and evidenced topological changes in fiber compactness proper to each model. In conclusion, we describe here a robust digital method for fibrosis analysis allowing accurate quantification, pattern recognition, and multi-organ comparisons useful to understand fibrosis dynamics.

Artificial intelligence identifies inflammation and confirms fibroblast foci as prognostic tissue biomarkers in idiopathic pulmonary fibrosis

A large number of fibroblast foci (FF) predict mortality in idiopathic pulmonary fibrosis (IPF). Other prognostic histological markers have not been identified. Artificial intelligence (AI) offers a possibility to quantitate possible prognostic histological features in IPF. We aimed to test the use of AI in IPF lung tissue samples by quantitating FF, interstitial mononuclear inflammation, and intra-alveolar macrophages with a deep convolutional neural network (CNN). Lung tissue samples of 71 patients with IPF from the FinnishIPF registry were analyzed by an AI model developed in the Aiforia® platform. The model was trained to detect tissue, air spaces, FF, interstitial mononuclear inflammation, and intra-alveolar macrophages with 20 samples. For survival analysis, cut-point values for high and low values of histological parameters were determined with maximally selected rank statistics. Survival was analyzed using the Kaplan-Meier method. A large area of FF predicted poor prognosis in IPF (p = 0.01). High numbers of interstitial mononuclear inflammatory cells and intra-alveolar macrophages were associated with prolonged survival (p = 0.01 and p = 0.01, respectively). Of lung function values, low diffusing capacity for carbon monoxide was connected to a high density of FF (p = 0.03) and a high forced vital capacity of predicted was associated with a high intra-alveolar macrophage density (p = 0.03). The deep CNN detected histological features that are difficult to quantitate manually. Interstitial mononuclear inflammation and intra-alveolar macrophages were novel prognostic histological biomarkers in IPF. Evaluating histological features with AI provides novel information on the prognostic estimation of IPF.

Deep learning enables pathologist-like scoring of NASH models

Non-alcoholic fatty liver disease (NAFLD) and the progressive form of non-alcoholic steatohepatitis (NASH) are diseases of major importance with a high unmet medical need. Efficacy studies on novel compounds to treat NAFLD/NASH using disease models are frequently evaluated using established histological feature scores on ballooning, inflammation, steatosis and fibrosis. These features are assessed by a trained pathologist using microscopy and assigned discrete scores. We demonstrate how to automate these scores with convolutional neural networks (CNNs). Whole slide images of stained liver sections are analyzed using two different scales with four CNNs, each specialized for one of four histopathological features. A continuous value is obtained to quantify the extent of each feature, which can be used directly to provide a high resolution readout. In addition, the continuous values can be mapped to obtain the established discrete pathologist-like scores. The automated deep learning-based scores show good agreement with the trainer - a human pathologist.

Society of Toxicologic Pathology Digital Pathology and Image Analysis Special Interest Group Article*: Opinion on the Application of Artificial Intelligence and Machine Learning to Digital Toxicologic Pathology

Toxicologic pathology is transitioning from analog to digital methods. This transition seems inevitable due to a host of ongoing social and medical technological forces. Of these, artificial intelligence (AI) and in particular machine learning (ML) are globally disruptive, rapidly growing sectors of technology whose impact on the long-established field of histopathology is quickly being realized. The development of increasing numbers of algorithms, peering ever deeper into the histopathological space, has demonstrated to the scientific community that AI pathology platforms are now poised to truly impact the future of precision and personalized medicine. However, as with all great technological advances, there are implementation and adoption challenges. This review aims to define common and relevant AI and ML terminology, describe data generation and interpretation, outline current and potential future business cases, discuss validation and regulatory hurdles, and most importantly, propose how overcoming the challenges of this burgeoning technology may shape toxicologic pathology for years to come, enabling pathologists to contribute even more effectively to answering scientific questions and solving global health issues. [Box: see text]