Modern Imaging Technologies in Toxicologic Pathology: An Overview

Convolutional Neural Networks for Classifying Cervical Cancer Types Using Histological Images

Cervical cancer is the most common cancer among women worldwide. The diagnosis and classification of cancer are extremely important, as it influences the optimal treatment and length of survival. The objective was to develop and validate a diagnosis system based on convolutional neural networks (CNN) that identifies cervical malignancies and provides diagnostic interpretability. A total of 8496 labeled histology images were extracted from 229 cervical specimens (cervical squamous cell carcinoma, SCC, n = 37; cervical adenocarcinoma, AC, n = 8; nonmalignant cervical tissues, n = 184). AlexNet, VGG-19, Xception, and ResNet-50 with five-fold cross-validation were constructed to distinguish cervical cancer images from nonmalignant images. The performance of CNNs was quantified in terms of accuracy, precision, recall, and the area under the receiver operating curve (AUC). Six pathologists were recruited to make a comparison with the performance of CNNs. Guided Backpropagation and Gradient-weighted Class Activation Mapping (Grad-CAM) were deployed to highlight the area of high malignant probability. The Xception model had excellent performance in identifying cervical SCC and AC in test sets. For cervical SCC, AUC was 0.98 (internal validation) and 0.974 (external validation). For cervical AC, AUC was 0.966 (internal validation) and 0.958 (external validation). The performance of CNNs falls between experienced and inexperienced pathologists. Grad-CAM and Guided Gard-CAM ensured diagnoses interpretability by highlighting morphological features of malignant changes. CNN is efficient for histological image classification tasks of distinguishing cervical malignancies from benign tissues and could highlight the specific areas of concern. All these findings suggest that CNNs could serve as a diagnostic tool to aid pathologic diagnosis.

An improved residual U-Net with morphological-based loss function for automatic liver segmentation in computed tomography

This paper proposes an improved ResU-Net framework for automatic liver CT segmentation. By employing a new loss function and data augmentation strategy, the accuracy of liver segmentation is improved, and the performance is verified on two public datasets LiTS17 and SLiver07. Firstly, to speed up the convergence of the model, the residual module is used to replace the original convolution module of U-Net. Secondly, to suppress the problem of pixel imbalance, the opposite number of Dice is proposed to replace the cross-entropy loss function, and the morphological method is introduced to weigh the pixels. Finally, to improve the generalization ability of the model, random affine transformation and random elastic deformation are employed for data augmentation. From 20 training datasets of Sliver07, 16 sets were selected as the training set, two sets were used for verification, and two sets were used for the test; meanwhile, from 131 training datasets of LiTS2017, eight sets were selected as the test set. In the experiment, four evaluation metrics, including DICE global, DICE per case, VOE, and RVD, were calculated, with the accuracies of 94.28, 94.24 ± 2.07, 10.83 ± 3.70, and -0.25 ± 2.74, respectively. Compared with U-Net and ResU-Net, the performance of the proposed method is significantly improved. The experimental results show that, although the method's complexity is high, it has a faster convergence speed and stronger generalization ability. The segmentation effect on the 2D image is significantly improved, and the scalability on 3D data is also robust. In addition, the proposed method performs well in the case of low-contrast neighboring organs, which proves the robustness of the proposed method.

[18F]F-ET-OTSSP167 Targets Maternal Embryo Leucine Zipper Kinase for PET Imaging of Triple-Negative Breast Cancer

Maternal embryo leucine zipper kinase (MELK) is a serine/threonine kinase and is highly expressed in triple-negative breast cancer (TNBC). This study aimed to develop a ¹⁸F-radiolabeled tracer based on the structure of a small-molecule MELK inhibitor OTSSP167 and evaluate its application for PET imaging of MELK expression in TNBC. OTSSP167 was modified with ethylene glycol to adjust its pharmacokinetics and was then radiolabeled with ¹⁸F to obtain [¹⁸F]F-ET-OTSSP167 at a labeling yield of 7.14 ± 2.19% and a molar activity of 16.23 ± 1.13 MBq/nmol. In vitro binding assays showed differentiated binding affinities of [¹⁸F]F-ET-OTSSP167 in different breast cancer cell lines, with high uptake in MDA-MB-231 (mild MELK expression) and low uptake in MCF-7 (negative MELK expression). PET imaging revealed that MDA-MB-231 tumors could be clearly delineated in vivo, while low tracer uptake was observed in MCF-7 tumors. These findings were confirmed by ex vivo biodistribution studies and were consistent with the immunohistochemistry and tissue staining results. Tracer accumulation in MDA-MB-231 tumors was significantly inhibited by excess amounts of OTSSP167, indicating high specificity of the tracer. In summary, [¹⁸F]F-ET-OTSSP167, an easily-prepared probe, can be used to visualize MELK positive tumors, demonstrating its promising clinical potential in selecting patients for MELK inhibitor therapy.

Quantitative neurotoxicology: Potential role of artificial intelligence/deep learning approach

Neurotoxicity studies are important in the preclinical stages of drug development process, because exposure to certain compounds that may enter the brain across a permeable blood brain barrier damages neurons and other supporting cells such as astrocytes. This could, in turn, lead to various neurological disorders such as Parkinson's or Huntington's disease as well as various dementias. Toxicity assessment is often done by pathologists after these exposures by qualitatively or semiquantitatively grading the severity of neurotoxicity in histopathology slides. Quantification of the extent of neurotoxicity supports qualitative histopathological analysis and provides a better understanding of the global extent of brain damage. Stereological techniques such as the utilization of an optical fractionator provide an unbiased quantification of the neuronal damage; however, the process is time-consuming. Advent of whole slide imaging (WSI) introduced digital image analysis which made quantification of neurotoxicity automated, faster and with reduced bias, making statistical comparisons possible. Although automated to a certain level, simple digital image analysis requires manual efforts of experts which is time-consuming and limits analysis of large datasets. Digital image analysis coupled with a deep learning artificial intelligence model provides a good alternative solution to time-consuming stereological and simple digital analysis. Deep learning models could be trained to identify damaged or dead neurons in an automated fashion. This review has focused on and discusses studies demonstrating the role of deep learning in segmentation of brain regions, toxicity detection and quantification of degenerated neurons as well as the estimation of area/volume of degeneration.

Deep learning in digital pathology image analysis: a survey

Deep learning (DL) has achieved state-of-the-art performance in many digital pathology analysis tasks. Traditional methods usually require hand-crafted domain-specific features, and DL methods can learn representations without manually designed features. In terms of feature extraction, DL approaches are less labor intensive compared with conventional machine learning methods. In this paper, we comprehensively summarize recent DL-based image analysis studies in histopathology, including different tasks (e.g., classification, semantic segmentation, detection, and instance segmentation) and various applications (e.g., stain normalization, cell/gland/region structure analysis). DL methods can provide consistent and accurate outcomes. DL is a promising tool to assist pathologists in clinical diagnosis.

Digital Image Analyses on Whole-Lung Slides in Mouse Models of Acute Pneumonia

Descriptive histopathology of mouse models of pneumonia is essential in assessing the outcome of infections, molecular manipulations, or therapies in the context of whole lungs. Quantitative comparisons between experimental groups, however, have been limited to laborious stereology or ill-defined scoring systems that depend on the subjectivity of a more or less experienced observer. Here, we introduce self-learning digital image analyses that allow us to transform optical information from whole mouse lung sections into statistically testable data. A pattern-recognition-based software and a nuclear count algorithm were adopted to quantify user-defined pathologies from whole slide scans of lungs infected with Streptococcus pneumoniae or influenza A virus compared with PBS-challenged lungs. The readout parameters "relative area affected" and "nuclear counts per area" are proposed as relevant criteria for the quantification of lesions from hematoxylin and eosin-stained sections, also allowing for the generation of a heat map of, for example, immune cell infiltrates with anatomical assignments across entire lung sections. Moreover, when combined with immunohistochemical labeling of marker proteins, both approaches are useful for the identification and counting of, for example, immune cell populations, as validated here by direct comparisons with flow cytometry data. The solutions can easily and flexibly be adjusted to specificities of different models or pathogens. Automated digital analyses of whole mouse lung sections may set a new standard for the user-defined, high-throughput comparative quantification of histological and immunohistochemical images. Still, our algorithms established here are only a start, and need to be tested in additional studies and other applications in the future.

Meeting Report: Tissue-based Image Analysis

Quantitative image analysis (IA) is a rapidly evolving area of digital pathology. Although not a new concept, the quantification of histological features on photomicrographs used to be cumbersome, resource-intensive, and limited to specialists and specialized laboratories. Recent technological advances like highly efficient automated whole slide digitizer (scanner) systems, innovative IA platforms, and the emergence of pathologist-friendly image annotation and analysis systems mean that quantification of features on histological digital images will become increasingly prominent in pathologists' daily professional lives. The added value of quantitative IA in pathology includes confirmation of equivocal findings noted by a pathologist, increasing the sensitivity of feature detection, quantification of signal intensity, and improving efficiency. There is no denying that quantitative IA is part of the future of pathology; however, there are also several potential pitfalls when trying to estimate volumetric features from limited 2-dimensional sections. This continuing education session on quantitative IA offered a broad overview of the field; a hands-on toxicologic pathologist experience with IA principles, tools, and workflows; a discussion on how to apply basic stereology principles in order to minimize bias in IA; and finally, a reflection on the future of IA in the toxicologic pathology field.

Evaluation of 99mTc-HYNIC-VCAM-1scFv as a Potential Qualitative and Semiquantitative Probe Targeting Various Tumors

Vascular cell adhesion molecule 1 (VCAM-1) is overexpressed in varieties of cancers. This study aimed to evaluate the application of a single chain variable fragment (scFv) of anti-VCAM-1 antibody labeled with ^99mTc as a possible imaging agent in several tumors. VCAM-1 scFv was labeled with ^99mTc using succinimidyl 6-hydrazinium nicotinate hydrochloride, and ^99mTc-HYNIC-VCAM-1_scFv was successfully synthesized with a high radiolabeling yield. VCAM-1 expression was evaluated in six cell lines by immunofluorescence staining. In vitro binding assays showed different binding affinities of ^99mTc-HYNIC-VCAM-1_scFv in different tumor cell lines, with high uptake in B16F10 melanoma and HT1080 fibrosarcoma cells, which was consistent with immunofluorescence staining results. In vivo SPECT planar imaging demonstrated that B16F10 and HT1080 tumors could be clearly visualized. Less intense uptake was observed in human SKOV3.ip ovarian tumor, and weak uptake was observed in human A375m melanoma, MDA-MB-231 breast cancer, and 786-O renal tumors. These findings were confirmed by biodistribution and immunofluorescence studies. High uptake by B16F10 tumors was inhibited by excess unlabeled VCAM-1_scFv. ^99mTc-HYNIC-VCAM-1_scFv, which selectively binds to VCAM-1, can provide a qualitative and semiquantitative method for noninvasive evaluation of VCAM-1 expression by tumors.

Compact Magnetic Resonance Imaging Systems-Novel Cost-Effective Tools for Preclinical Drug Safety and Efficacy Evaluation

Practical magnetic resonance imaging for use in investigative and preclinical toxicology studies is now feasible. Newly developed, self-containing imaging systems provide an efficient and cost-effective means to rapidly obtain in vivo and ex vivo magnetic resonance imaging images to improve how we perform toxicology and toxicologic pathology.

Fluorescence Detection of Cell Death in Liver of Mice Treated with Thioacetamide

The purpose of this study was to detect cell death in the liver of mice treated with thioacetamide (TAA) using fluorescence bioimaging and compare this outcome with that using conventional histopathological examination. At 6 weeks of age, 24 mice were randomly divided into three groups: group 1 (G1), control group; group 2 (G2), fluorescence probe control group; group 3 (G3), TAA-treated group. G3 mice were treated with TAA. Twenty-two hours after TAA treatment, G2 and G3 mice were treated with Annexin-Vivo 750. Fluorescence in vivo bioimaging was performed by fluorescence molecular tomography at two hours after Annexin-Vivo 750 treatment, and fluorescence ex vivo bioimaging of the liver was performed. Liver damage was validated by histopathological examination. In vivo bioimaging showed that the fluorescence intensity was increased in the right upper part of G3 mice compared with that in G2 mice, whereas G1 mice showed no signal. Additionally ex vivo bioimaging showed that the fluorescence intensity was significantly increased in the livers of G3 mice compared with those in G1 or G2 mice (p < 0.05). Histopathological examination of the liver showed no cell death in G1 and G2 mice. However, in G3 mice, there was destruction of hepatocytes and increased cell death. Terminal deoxynucleotidyl transferase dUTP nick end labeling staining confirmed many cell death features in the liver of G3 mice, whereas no pathological findings were observed in the liver of G1 and G2 mice. Taken together, fluorescence bioimaging in this study showed the detection of cell death and made it possible to quantify the level of cell death in male mice. The outcome was correlated with conventional biomedical examination. As it was difficult to differentiate histological location by fluorescent bioimaging, it is necessary to develop specific fluorescent dyes for monitoring hepatic disease progression and to exploit new bioimaging techniques without dye-labeling.

Advances in the design of cell-permeable fluorescent probes for applications in live cell imaging

Advances in the design strategy of cell-permeable small fluorescent probes are discussed. Their applications in imaging specific cell types and intracellular bioanalytes, as well as the cellular environment in live conditions, are presented.

Magnetic Resonance Imaging as a Noninvasive Method for Longitudinal Monitoring of Infusion Site Reactions Following Administration of a Novel Apomorphine Formulation

Infusion site reactions are common following subcutaneous infusion of drugs. Such reactions can lead to discontinuation of the treatment. Therefore, assessment of such reactions is essential during preclinical safety studies, and magnetic resonance imaging (MRI) can assist in evaluation. Here, in vivo and ex vivo MRI evaluations were used in addition to classical histopathology to assess the infusion site reaction to ND0701, a novel formulation of apomorphine base developed for the treatment of Parkinson’s disease, in comparison to the commercial apomorphine hydrochloride (HCl) formulation. Both formulations, each at two concentrations, were continuously administered subcutaneously for 20 hr to each of 3 male and 3 female domestic pigs. Based on MRI evaluations, there was a gradual decrease in the volume of the subcutaneous lesions over 4 weeks, with smaller lesions and quicker resolution with ND0701 at concentrations 2.5- to 5-fold higher when compared to the commercial apomorphine HCl formulation. Histopathological evaluation of ND0701 revealed only minimal inflammation at the sites of infusion, whereas the commercial apomorphine HCl caused persistent inflammatory reactions and necrosis. This study provides support to the use of MRI in preclinical testing of subcutaneous drugs when evaluating local site reactions.

Assessment of collagen antibody-induced arthritis in BALB/c mice using bioimaging analysis and histopathological examination

The aim of this study was to examine the therapeutic potential of sulfasalazine and prednisolone in a mouse collagen antibody-induced arthritis (CAIA) model. Twenty-five male BALB/c mice were randomly divided into five groups: group 1 (G1): control, group 2 (G2): probe control, group 3 (G3): CAIA, group 4 (G4): CAIA+sulfasalazine (10 mg/kg, oral), and group 5 (G5): CAIA+prednisolone (100 mg/kg, oral). Fluorescence bioimaging was performed in vivo 24 and 48 h after treatment with a fluorescence probe (OsteoSense® 680 EX), and all mice were sacrificed. The hind knee joints were fixed in 10% neutral phosphate-buffered formalin, and micro-computed tomography (micro-CT) and histopathological analyses were performed. The paw thickness increased in a time-dependent manner in G3 mice, but trended toward a decrease in both G4 and G5 mice. Fluorescence intensity increased in G3 mice at 24 and 48 h after fluorescence probe treatment, but the fluorescence intensity in G4 and G5 mice was lower than that in G3. Micro-CT analyses showed that the joint surfaces of G3 mice had a rough and irregular articular appearance, but the occurrence of these irregularities was lower in G4 and G5. Hematoxylin and eosin and Safranin O-fast green staining confirmed that destruction of the cartilage and bony structures, synovial hyperplasia, and inflammatory cell infiltration all occurred in G3, and that the occurrence of these phenomena was lower in G4 and G5 than in G3. Taken together, these results suggest that sulfasalazine and prednisolone can reduce acute rheumatoid arthritis in mice.

Magnetic resonance imaging assessment of the ventricular system in the brains of adult and juvenile beagle dogs treated with posaconazole IV Solution

INTRODUCTION

Noxafil® (posaconazole; POS) is a potent, selective triazole antifungal approved for use in adults as an oral suspension, oral tablet and intravenous (IV) Solution. In support of pediatric administration of POS IV Solution to children<two years of age, a nonclinical study in juvenile pre-weaning Beagle dogs was conducted, which showed enlarged lateral ventricles in the brain at the conclusion of a 6 week dosing period.

METHODS

To evaluate the impact of this finding on older age dogs, which would support administration to children>two years of age, two studies were undertaken using magnetic resonance imaging (MRI) to monitor brain ventricle size longitudinally during three months administration of POS IV in adult and juvenile dogs. Necropsy was performed on all animals at the end of the studies. From the baseline MRI images, great variability in ventricle size was noted in both the adult and juvenile dogs; these images were used to distribute differently sized ventricles between treatment and vehicle groups as to not skew group means during the course of the study.

RESULTS

POS IV Solution had no effect on ventricle volume at any timepoint during dosing in either the adult or the juvenile dogs. Further, no gross or histomorphologic differences between groups were observed in either study. Compared to juvenile dogs, MRI analysis showed that adult dogs had larger ventricles, lower variability in all ventricle volumes, and a greater rate of increase in total ventricle volume.

DISCUSSION

Information on growth and development of brains is one of the few areas in which more detailed information is available about humans than about the standard laboratory animals used to model disease and predict toxicities. The use of MRI helped elucidate large natural variabilities in the dog brain, which could have altered the interpretation of this de-risking study, and provided a valuable noninvasive means to monitor the brain ventricles longitudinally.

Development of a computational high-throughput tool for the quantitative examination of dose-dependent histological features

High-resolution digitalizing of histology slides facilitates the development of computational alternatives to manual quantitation of features of interest. We developed a MATLAB-based quantitative histological analysis tool (QuHAnT) for the high-throughput assessment of distinguishable histological features. QuHAnT validation was demonstrated by comparison with manual quantitation using liver sections from mice orally gavaged with sesame oil vehicle or 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD; 0.001-30 μg/kg) every 4 days for 28 days, which elicits hepatic steatosis with mild fibrosis. A quality control module of QuHAnT reduced the number of quantifiable Oil Red O (ORO)-stained images from 3,123 to 2,756. Increased ORO staining was measured at 10 and 30 μg/kg TCDD with a high correlation between manual and computational volume densities (Vv ), although the dynamic range of QuHAnT was 10-fold greater. Additionally, QuHAnT determined the size of each ORO vacuole, which could not be accurately quantitated by visual examination or manual point counting. PicroSirius Red quantitation demonstrated superior collagen deposition detection due to the ability to consider all images within each section. QuHAnT dramatically reduced analysis time and facilitated the comprehensive assessment of features improving accuracy and sensitivity and represents a complementary tool for tissue/cellular features that are difficult and tedious to assess via subjective or semiquantitative methods.

Four-dimensional MRI of renal function in the developing mouse

The major roles of filtration, metabolism and high blood flow make the kidney highly vulnerable to drug-induced toxicity and other renal injuries. A method to follow kidney function is essential for the early screening of toxicity and malformations. In this study, we acquired high spatiotemporal resolution (four dimensional) datasets of normal mice to follow changes in kidney structure and function during development. The data were acquired with dynamic contrast-enhanced MRI (via keyhole imaging) and a cryogenic surface coil, allowing us to obtain a full three-dimensional image (isotropic resolution, 125 microns) every 7.7 s over a 50-min scan. This time course permitted the demonstration of both contrast enhancement and clearance. Functional changes were measured over a 17-week course (at 3, 5, 7, 9, 13 and 17 weeks). The time dimension of the MRI dataset was processed to produce unique image contrasts to segment the four regions of the kidney: cortex (CO), outer stripe (OS) of the outer medulla (OM), inner stripe (IS) of the OM and inner medulla (IM). Local volumes, time-to-peak (TTP) values and decay constants (DC) were measured in each renal region. These metrics increased significantly with age, with the exception of DC values in the IS and OS. These data will serve as a foundation for studies of normal renal physiology and future studies of renal diseases that require early detection and intervention.

High-throughput, automated quantification of white matter neurons in mild malformation of cortical development in epilepsy

INTRODUCTION

In epilepsy, the diagnosis of mild Malformation of Cortical Development type II (mMCD II) predominantly relies on the histopathological assessment of heterotopic neurons in the white matter. The exact diagnostic criteria for mMCD II are still ill-defined, mainly because findings from previous studies were contradictory due to small sample size, and the use of different stains and quantitative systems. Advance in technology leading to the development of whole slide imaging with high-throughput, automated quantitative analysis (WSA) may overcome these differences, and may provide objective, rapid, and reliable quantitation of white matter neurons in epilepsy. This study quantified the density of NeuN immunopositive neurons in the white matter of up to 142 epilepsy and control cases using WSA. Quantitative data from WSA was compared to two other systems, semi-automated quantitation, and the widely accepted method of stereology, to assess the reliability and quality of results from WSA.

RESULTS

All quantitative systems showed a higher density of white matter neurons in epilepsy cases compared to controls (P = 0.002). We found that, in particular, WSA with user-defined region of interest (manual) was superior in terms of larger sampled size, ease of use, time consumption, and accuracy in region selection and cell recognition compared to other methods. Using results from WSA manual, we proposed a threshold value for the classification of mMCD II, where 78% of patients now classified with mMCD II were seizure-free at the second post-operatively follow up.

CONCLUSION

This study confirms the potential role of WSA in future quantitative diagnostic histology, especially for the histopathological diagnosis of mMCD.

Chickens from lines artificially selected for juvenile low and high body weight differ in glucose homeostasis and pancreas physiology

Artificial selection of White Plymouth Rock chickens for juvenile (day 56) body weight resulted in two divergent genetic lines: hypophagic low weight (LWS) chickens and hyperphagic obese high weight (HWS) chickens, with the latter more than 10-fold heavier than the former at selection age. A study was designed to investigate glucose regulation and pancreas physiology at selection age in LWS chickens and HWS chickens. Oral glucose tolerance and insulin sensitivity tests revealed differences in threshold sensitivity to insulin and glucose clearance rate between the lines. Results from real-time PCR showed greater pancreatic mRNA expression of four glucose regulatory genes (preproinsulin, PPI; preproglucagon, PPG; glucose transporter 2, GLUT2; and pancreatic duodenal homeobox 1, Pdx1) in LWS chickens, than HWS chickens. Histological analysis of the pancreas revealed that HWS chickens have larger pancreatic islets, less pancreatic islet mass, and more pancreatic inflammation than LWS chickens, all of which presumably contribute to impaired glucose metabolism.

Whole-slide imaging and automated image analysis: considerations and opportunities in the practice of pathology

Digital pathology, the practice of pathology using digitized images of pathologic specimens, has been transformed in recent years by the development of whole-slide imaging systems, which allow for the evaluation and interpretation of digital images of entire histologic sections. Applications of whole-slide imaging include rapid transmission of pathologic data for consultations and collaborations, standardization and distribution of pathologic materials for education, tissue specimen archiving, and image analysis of histologic specimens. Histologic image analysis allows for the acquisition of objective measurements of histomorphologic, histochemical, and immunohistochemical properties of tissue sections, increasing both the quantity and quality of data obtained from histologic assessments. Currently, numerous histologic image analysis software solutions are commercially available. Choosing the appropriate solution is dependent on considerations of the investigative question, computer programming and image analysis expertise, and cost. However, all studies using histologic image analysis require careful consideration of preanalytical variables, such as tissue collection, fixation, and processing, and experimental design, including sample selection, controls, reference standards, and the variables being measured. The fields of digital pathology and histologic image analysis are continuing to evolve, and their potential impact on pathology is still growing. These methodologies will increasingly transform the practice of pathology, allowing it to mature toward a quantitative science. However, this maturation requires pathologists to be at the forefront of the process, ensuring their appropriate application and the validity of their results. Therefore, histologic image analysis and the field of pathology should co-evolve, creating a symbiotic relationship that results in high-quality reproducible, objective data.

Toxicologic Pathology in the 21st Century

Toxicology is and will be heavily influenced by advances in many scientific disciplines. For toxicologic pathology, particularly relevant are the increasing array of molecular methods providing deeper insights into toxicity pathways, in vivo imaging techniques visualizing toxicodynamics and more powerful computers anticipated to allow (partly) automated morphological diagnoses. It appears unlikely that, in a foreseeable future, animal studies can be replaced by in silico and in vitro studies or longer term in vivo studies by investigations of biomarkers including toxicogenomics of shorter term studies, though the importance of such approaches will continue to increase. In addition to changes based on scientific progress, the work of toxicopathologists is and will be affected by social and financial factors, among them stagnating budgets, globalization, and outsourcing. The number of toxicopathologists in North America, Europe, and the Far East is not expected to grow. Many toxicopathologists will likely spend less time at the microscope but will be more heavily involved in early research activities, imaging, and as generalists with a broad biological understanding in evaluation and management of toxicity. Toxicologic pathology will remain important and is indispensable for validation of new methods, quality assurance of established methods, and for areas without good alternative methods.

Investigation into diagnostic agreement using automated computer-assisted histopathology pattern recognition image analysis

The extent to which histopathology pattern recognition image analysis (PRIA) agrees with microscopic assessment has not been established. Thus, a commercial PRIA platform was evaluated in two applications using whole-slide images. Substantial agreement, lacking significant constant or proportional errors, between PRIA and manual morphometric image segmentation was obtained for pulmonary metastatic cancer areas (Passing/Bablok regression). Bland-Altman analysis indicated heteroscedastic measurements and tendency toward increasing variance with increasing tumor burden, but no significant trend in mean bias. The average between-methods percent tumor content difference was -0.64. Analysis of between-methods measurement differences relative to the percent tumor magnitude revealed that method disagreement had an impact primarily in the smallest measurements (tumor burden <3%). Regression-based 95% limits of agreement indicated substantial agreement for method interchangeability. Repeated measures revealed concordance correlation of >0.988, indicating high reproducibility for both methods, yet PRIA reproducibility was superior (C.V.: PRIA = 7.4, manual = 17.1). Evaluation of PRIA on morphologically complex teratomas led to diagnostic agreement with pathologist assessments of pluripotency on subsets of teratomas. Accommodation of the diversity of teratoma histologic features frequently resulted in detrimental trade-offs, increasing PRIA error elsewhere in images. PRIA error was nonrandom and influenced by variations in histomorphology. File-size limitations encountered while training algorithms and consequences of spectral image processing dominance contributed to diagnostic inaccuracies experienced for some teratomas. PRIA appeared better suited for tissues with limited phenotypic diversity. Technical improvements may enhance diagnostic agreement, and consistent pathologist input will benefit further development and application of PRIA.

High throughput object-based image analysis of β-amyloid plaques in human and transgenic mouse brain

Advances in imaging technology have enabled automated approaches for quantitative image analysis. In this study, a high content object based image analysis method was developed for quantification of β-amyloid (Aβ) plaques in postmortem brains of Alzheimer's disease (AD) subjects and in transgenic mice over overexpressing Aβ. Digital images acquired from immunohistochemically stained sections of the superior frontal gyrus were analyzed for Aβ plaque burden using a Definiens object-based segmentation approach. Blinded evaluation of Aβ stained sections from AD and aged matched human subjects accurately identified AD cases with one exception. Brains from transgenic mice overexpressing Aβ (PS1APP mice) were also evaluated by our Definiens object based image analysis approach. We observed an age-dependent increase in the amount of Aβ plaque load that we quantified in both the hippocampus and cortex. From the contralateral hemisphere, we measured the amount of Aβ in brain homogenates biochemically and observed a significant correlation between our biochemical measurements and those that we measured by our object based Definiens system in the hippocampus. Assessment of Aβ plaque load in PS1APP mice using a manual segmentation technique (Image-Pro Plus) confirmed the results of our object-based image analysis approach. Image acquisition and analysis of 32 stained human slides and 100 mouse slides were executed in 8 h and 22 h, respectively supporting the relatively high throughput features of the Definiens platform. The data show that digital imaging combined with object based image analysis is a reliable and efficient approach to quantifying Aβ plaques in human and mouse brain.

The Use of Immunohistochemistry for Biomarker Assessment—Can It Compete with Other Technologies?

A morphology-based assay such as immunohistochemistry (IHC) should be a highly effective means to define the expression of a target molecule of interest, especially if the target is a protein. However, over the past decade, IHC as a platform for biomarkers has been challenged by more quantitative molecular assays with reference standards but that lack morphologic context. For IHC to be considered a “top-tier” biomarker assay, it must provide truly quantitative data on par with non-morphologic assays, which means it needs to be run with reference standards. However, creating such standards for IHC will require optimizing all aspects of tissue collection, fixation, section thickness, morphologic criteria for assessment, staining processes, digitization of images, and image analysis. This will also require anatomic pathology to evolve from a discipline that is descriptive to one that is quantitative. A major step in this transformation will be replacing traditional ocular microscopes with computer monitors and whole slide images, for without digitization, there can be no accurate quantitation; without quantitation, there can be no standardization; and without standardization, the value of morphology-based IHC assays will not be realized.

Extended Field Laser Confocal Microscopy (EFLCM): combining automated Gigapixel image capture with in silico virtual microscopy

BACKGROUND

Confocal laser scanning microscopy has revolutionized cell biology. However, the technique has major limitations in speed and sensitivity due to the fact that a single laser beam scans the sample, allowing only a few microseconds signal collection for each pixel. This limitation has been overcome by the introduction of parallel beam illumination techniques in combination with cold CCD camera based image capture.

METHODS

Using the combination of microlens enhanced Nipkow spinning disc confocal illumination together with fully automated image capture and large scale in silico image processing we have developed a system allowing the acquisition, presentation and analysis of maximum resolution confocal panorama images of several Gigapixel size. We call the method Extended Field Laser Confocal Microscopy (EFLCM).

RESULTS

We show using the EFLCM technique that it is possible to create a continuous confocal multi-colour mosaic from thousands of individually captured images. EFLCM can digitize and analyze histological slides, sections of entire rodent organ and full size embryos. It can also record hundreds of thousands cultured cells at multiple wavelength in single event or time-lapse fashion on fixed slides, in live cell imaging chambers or microtiter plates.

CONCLUSION

The observer independent image capture of EFLCM allows quantitative measurements of fluorescence intensities and morphological parameters on a large number of cells. EFLCM therefore bridges the gap between the mainly illustrative fluorescence microscopy and purely quantitative flow cytometry. EFLCM can also be used as high content analysis (HCA) instrument for automated screening processes.