On the concept of objectivity in digital image analysis in pathology

Reliability and Variability of Ki-67 Digital Image Analysis Methods for Clinical Diagnostics in Breast Cancer

Ki-67 is a nuclear protein associated with proliferation, and a strong potential biomarker in breast cancer, but is not routinely measured in current clinical management owing to a lack of standardization. Digital image analysis (DIA) is a promising technology that could allow high-throughput analysis and standardization. There is a dearth of data on the clinical reliability as well as intra- and interalgorithmic variability of different DIA methods. In this study, we scored and compared a set of breast cancer cases in which manually counted Ki-67 has already been demonstrated to have prognostic value (n = 278) to 5 DIA methods, namely Aperio ePathology (Lieca Biosystems), Definiens Tissue Studio (Definiens AG), Qupath, an unsupervised immunohistochemical color histogram algorithm, and a deep-learning pipeline piNET. The piNET system achieved high agreement (interclass correlation coefficient: 0.850) and correlation (R = 0.85) with the reference score. The Qupath algorithm exhibited a high degree of reproducibility among all rater instances (interclass correlation coefficient: 0.889). Although piNET performed well against absolute manual counts, none of the tested DIA methods classified common Ki-67 cutoffs with high agreement or reached the clinically relevant Cohen's κ of at least 0.8. The highest agreement achieved was a Cohen's κ statistic of 0.73 for cutoffs 20% and 25% by the piNET system. The main contributors to interalgorithmic variation and poor cutoff characterization included heterogeneous tumor biology, varying algorithm implementation, and setting assignments. It appears that image segmentation is the primary explanation for semiautomated intra-algorithmic variation, which involves significant manual intervention to correct. Automated pipelines, such as piNET, may be crucial in developing robust and reproducible unbiased DIA approaches to accurately quantify Ki-67 for clinical diagnosis in the future.

β-catenin, Pax2, and Pten Panel Identifies Precancers Among Histologically Subdiagnostic Endometrial Lesions

Despite refinements in histologic criteria for the diagnosis of endometrioid precancers, many challenging cases are encountered in daily practice, creating diagnostic uncertainty and suboptimal patient management. Recently, an immunohistochemical 3-marker panel consisting of β-catenin, Pax2, and Pten was identified as a useful diagnostic adjunct. However, previous studies focused either on cancers or diagnostically unambiguous precancers, leaving questions about the applicability and utility of the panel in endometria with architectural features near or below the threshold of accepted histologic criteria for endometrioid precancers. Here, in a retrospective study of 90 patients, we evaluated the performance of the 3-marker panel. Notably, the panel detected a subset of disordered proliferative endometria (8/44, 18%), nonatypical hyperplasias (19/40, 48%), and cases with ambiguous features (3/6, 50%) with aberrancy for ≥1 markers. Marker-aberrant cases were more likely to progress to endometrioid precancer or cancer ( P =0.0002). Patterns of marker aberrancy in the index and progressor cases from individual patients provided evidence for origin in a common precursor, and next-generation sequencing of the progressor cases rationalized marker aberrancy for β-catenin and Pten. The results unequivocally demonstrate that some lesions that do not approach current histologic thresholds are bona fide neoplastic precursors with clinically-relevant driver events that can be detected by the 3-marker panel. The findings provide further validation for the diagnostic utility of the panel in clinical practice and its application in difficult or ambiguous cases.

Local Antibiotic Delivery Ceramic Bone Substitutes for the Treatment of Infected Bone Cavities and Bone Regeneration: A Systematic Review on What We Have Learned from Animal Models

AIMS

the focus of this study is to evaluate if the combination of an antibiotic with a ceramic biomaterial is effective in treating osteomyelitis in an infected animal model and to define which model and protocol are best suited for in vivo experiments of local bone infection treatment.

METHODS

a systematic review was carried out based on PRISMA statement guidelines. A PubMed search was conducted to find original papers on animal models of bone infections using local antibiotic delivery systems with the characteristics of bone substitutes. Articles without a control group, differing from the experimental group only by the addition of antibiotics to the bone substitute, were excluded.

RESULTS

a total of 1185 records were retrieved, and after a three-step selection, 34 papers were included. Six manuscripts studied the effect of antibiotic-loaded biomaterials on bone infection prevention. Five articles studied infection in the presence of foreign bodies. In all but one, the combination of an antibiotic with bioceramic bone substitutes tended to prevent or cure bone infection while promoting biomaterial osteointegration.

CONCLUSIONS

this systematic review shows that the combination of antibiotics with bioceramic bone substitutes may be appropriate to treat bone infection when applied locally. The variability of the animal models, time to develop an infection, antibiotic used, way of carrying and releasing antibiotics, type of ceramic material, and endpoints limits the conclusions on the ideal therapy, enhancing the need for consistent models and guidelines to develop an adequate combination of material and antimicrobial agent leading to an effective human application.

Techniques for digital histological morphometry of the pineal gland

INTRODUCTION

The pineal gland is a small photo-neuro-endocrine organ. This study used human post-mortem pineal glands to microscopically assess immunohistochemical marker intensity and percentage of positivity using known and novel digital techniques.

MATERIALS AND METHODS

An experimental non-inferiority study has been performed on 72 pineal glands harvested from post-mortem examinations. The glands have been stained with glial fibrillary acidic protein (GFAP), synaptophysin (SYN), neuron-specific enolase (NSE), and neurofilament (NF). Slides were digitally scanned. Morphometric data were obtained using optical analysis, CaseViewer, ImageJ, and MorphoRGB RESULTS: Strong and statistically significant correlations were found and plotted using Bland-Altman diagrams between the two image analysis software in the case of mean percentage and intensity of GFAP, NSE, NF, and SYN.

DISCUSSIONS

Software such as SlideViewer and ImageJ, with our novel software MorphoRGB were used to perform histological morphometry of the pineal gland. Digital morphometry of a small organ such as the pineal gland is easy to do by using whole slide imaging (WSI) and digital image analysis software, with potential use in clinical settings. MorphoRGB provides slightly more accurate data than ImageJ and is more user-friendly regarding measurements of parenchyma percentage stained by immunohistochemistry. The results show that MorphoRGB is not inferior in functionality.

CONCLUSIONS

The described morphometric techniques have potential value in current practice, experimental small animal models and human pineal glands, or other small endocrine organs that can be fully included in a whole slide image. The software we used has applications in quantifying immunohistochemical stains.

Comparison of Semi-Quantitative Scoring and Artificial Intelligence Aided Digital Image Analysis of Chromogenic Immunohistochemistry

Semi-quantitative scoring is a method that is widely used to estimate the quantity of proteins on chromogen-labelled immunohistochemical (IHC) tissue sections. However, it suffers from several disadvantages, including its lack of objectivity and the fact that it is a time-consuming process. Our aim was to test a recently established artificial intelligence (AI)-aided digital image analysis platform, Pathronus, and to compare it to conventional scoring by five observers on chromogenic IHC-stained slides belonging to three experimental groups. Because Pathronus operates on grayscale 0-255 values, we transformed the data to a seven-point scale for use by pathologists and scientists. The accuracy of these methods was evaluated by comparing statistical significance among groups with quantitative fluorescent IHC reference data on subsequent tissue sections. The pairwise inter-rater reliability of the scoring and converted Pathronus data varied from poor to moderate with Cohen’s kappa, and overall agreement was poor within every experimental group using Fleiss’ kappa. Only the original and converted that were obtained from Pathronus original were able to reproduce the statistical significance among the groups that were determined by the reference method. In this study, we present an AI-aided software that can identify cells of interest, differentiate among organelles, protein specific chromogenic labelling, and nuclear counterstaining after an initial training period, providing a feasible and more accurate alternative to semi-quantitative scoring.

Ki-67 as a Prognostic Biomarker in Invasive Breast Cancer

Simple Summary

In breast cancer development, the expression of Ki-67 is strongly associated with cancer proliferation and is a known indicator of prognosis and outcome. Ki-67 expression levels are also useful to inform treatment decision making in some cases. As a result, routine measurement of Ki-67 is now widely performed during pathological tumour evaluation. However, the Ki-67 appraisal is not without its limitations and shortcomings—the aim of this study was to provide an overview of Ki-67 use in the clinical setting, the current challenges associated with its measurement, and the novel strategies that will hopefully enhance Ki-67 proliferation indices for prospective breast cancer patients.

Abstract

The advent of molecular medicine has transformed breast cancer management. Breast cancer is now recognised as a heterogenous disease with varied morphology, molecular features, tumour behaviour, and response to therapeutic strategies. These parameters are underpinned by a combination of genomic and immunohistochemical tumour factors, with estrogen receptor (ER) status, progesterone receptor (PgR) status, human epidermal growth factor receptor-2 (HER2) status, Ki-67 proliferation indices, and multigene panels all playing a contributive role in the substratification, prognostication and personalization of treatment modalities for each case. The expression of Ki-67 is strongly linked to tumour cell proliferation and growth and is routinely evaluated as a proliferation marker. This review will discuss the clinical utility, current pitfalls, and promising strategies to augment Ki-67 proliferation indices in future breast oncology.

What makes AI 'intelligent' and 'caring'? Exploring affect and relationality across three sites of intelligence and care

This paper scrutinises how AI and robotic technologies are transforming the relationships between people and machines in new affective, embodied and relational ways. Through investigating what it means to exist as human 'in relation' to AI across health and care contexts, we aim to make three main contributions. (1) We start by highlighting the complexities of philosophical issues surrounding the concepts of "artificial intelligence" and "ethical machines." (2) We outline some potential challenges and opportunities that the creation of such technologies may bring in the health and care settings. We focus on AI applications that interface with health and care via examples where AI is explicitly designed as an 'augmenting' technology that can overcome human bodily and cognitive as well as socio-economic constraints. We focus on three dimensions of 'intelligence' - physical, interpretive, and emotional - using the examples of robotic surgery, digital pathology, and robot caregivers, respectively. Through investigating these areas, we interrogate the social context and implications of human-technology interaction in the interrelational sphere of care practice. (3) We argue, in conclusion, that there is a need for an interdisciplinary mode of theorising 'intelligence' as relational and affective in ways that can accommodate the fragmentation of both conceptual and material boundaries between human and AI, and human and machine. Our aim in investigating these sociological, philosophical and ethical questions is primarily to explore the relationship between affect, relationality and 'intelligence,' the intersection and integration of 'human' and 'artificial' intelligence, through an examination of how AI is used across different dimensions of intelligence. This allows us to scrutinise how 'intelligence' is ultimately conveyed, understood and (technologically or algorithmically) configured in practice through emerging relationships that go beyond the conceptual divisions between humans and machines, and humans vis-à-vis artificial intelligence-based technologies.

Digital Image Analysis of Ki-67 Stained Tissue Microarrays and Recurrence in Tamoxifen-Treated Breast Cancer Patients

Purpose

The proliferation marker Ki-67 has been used as a prognostic marker to separate low- and high-risk breast cancer subtypes and guide treatment decisions for adjuvant chemotherapy. The association of Ki-67 with response to tamoxifen therapy is unclear. High-throughput automated scoring of Ki-67 might enable standardization of quantification and definition of clinical cut-off values. We hypothesized that digital image analysis (DIA) of Ki-67 can be used to evaluate proliferation in breast cancer tumors, and that Ki-67 may be associated with tamoxifen resistance in early-stage breast cancer.

Patients and Methods

Here, we apply DIA technology from Visiopharm using a custom designed algorithm for quantifying the expression of Ki-67, in a case–control study nested in the Danish Breast Cancer Group clinical database, consisting of stages I, II, or III breast cancer patients of 35–69 years of age, diagnosed during 1985–2001, in the Jutland peninsula, Denmark. We assessed DIA-Ki-67 score on tissue microarrays (TMAs) from breast cancer patients in a case–control study including 541 ER-positive and 300 ER-negative recurrent cases and their non-recurrent controls, matched on ER-status, cancer stage, menopausal status, year of diagnosis, and county of residence. We used logistic regression to estimate odds ratios and associated 95% confidence intervals to determine the association of Ki-67 expression with recurrence risk, adjusting for matching factors, chemotherapy, type of surgery, receipt of radiation therapy, age category, and comorbidity.

Results

Ki-67 was not associated with increased risk of recurrence in tamoxifen-treated patients (ORadj =0.72, 95% CI 0.54, 0.96) or ER-negative patients (ORadj =0.85, 95% CI 0.54, 1.34).

Conclusion

Our findings suggest that Ki-67 digital image analysis in TMAs is not associated with increased risk of recurrence among tamoxifen-treated ER-positive breast cancer or ER-negative breast cancer patients. Overall, our findings do not support an increased risk of recurrence associated with Ki-67 expression.

Current status of pathological image analysis technology in pharmaceutical companies: a questionnaire survey of the Japan Pharmaceutical Manufacturers Association

The Japan Pharmaceutical Manufacturers Association (JPMA) has instituted a task force (TF) for the "development of image analysis technology for histopathological changes" as part of the collaboration for realizing cutting-edge drug development since 2016. In recent years, there has been progress in the digital pathology technology; however, few applications in nonclinical drug development studies have been observed. Therefore, TF performed a questionnaire survey to investigate the current status, needs, possibility, and development of image analysis. The subjects were 35 member companies of the JPMA. The questionnaire was set to assess the efficacy and/or safety of researchers engaged in pathological evaluations for each company. The questions focused on the experiences, implementation, and issues regarding histopathological examinations; the need for image analysis software; and future views. Valid responses were obtained from 26 companies. Most companies assumed that the beneficial aspect of image analysis is to gain objectivity and persuasiveness; however, challenges in the analysis conditions with regard to accuracy and without subjectivity persist. Additionally, there seems to be a need for image analysis software with advanced digital pathology technology, with most companies believing that, in the future, pathological evaluations will be partly performed by computers. In conclusion, in this questionnaire survey, TF extracted the current status of image analysis in nonclinical studies performed by pharmaceutical companies and collected opinions on future prospects regarding the development of image analysis software with advanced digital pathology technology.

Deep Learning-Based Histopathologic Assessment of Kidney Tissue

BACKGROUND

The development of deep neural networks is facilitating more advanced digital analysis of histopathologic images. We trained a convolutional neural network for multiclass segmentation of digitized kidney tissue sections stained with periodic acid-Schiff (PAS).

METHODS

We trained the network using multiclass annotations from 40 whole-slide images of stained kidney transplant biopsies and applied it to four independent data sets. We assessed multiclass segmentation performance by calculating Dice coefficients for ten tissue classes on ten transplant biopsies from the Radboud University Medical Center in Nijmegen, The Netherlands, and on ten transplant biopsies from an external center for validation. We also fully segmented 15 nephrectomy samples and calculated the network's glomerular detection rates and compared network-based measures with visually scored histologic components (Banff classification) in 82 kidney transplant biopsies.

RESULTS

The weighted mean Dice coefficients of all classes were 0.80 and 0.84 in ten kidney transplant biopsies from the Radboud center and the external center, respectively. The best segmented class was "glomeruli" in both data sets (Dice coefficients, 0.95 and 0.94, respectively), followed by "tubuli combined" and "interstitium." The network detected 92.7% of all glomeruli in nephrectomy samples, with 10.4% false positives. In whole transplant biopsies, the mean intraclass correlation coefficient for glomerular counting performed by pathologists versus the network was 0.94. We found significant correlations between visually scored histologic components and network-based measures.

CONCLUSIONS

This study presents the first convolutional neural network for multiclass segmentation of PAS-stained nephrectomy samples and transplant biopsies. Our network may have utility for quantitative studies involving kidney histopathology across centers and provide opportunities for deep learning applications in routine diagnostics.

Assessing activated sludge morphology and oxygen transfer performance using image analysis

The morphology of the microbial communities can have dramatic impacts on not only the treatment performance, but also the energy use performance of an activated sludge process. In this research, we developed and calibrated an image analysis technique to determine key morphological parameters such as the floc diameter and the specific filament length (SFL) and discovered that the SFL has significant impacts on sludge floc size, the specific extracellular polymeric substances production, the settleability, mixed liquor viscosity, and oxygen transfer efficiency. When the SFL increased from 2.5 × 10⁹ μm g^-1 to 6.0 × 10¹⁰ μm g^-1, the apparent viscosity normalized by the mixed liquor suspended solids concentration increased by 67%, and the oxygen transfer efficiency decreased by 29%. A long solids retention time (SRT) of 40 day reduced SFL, improved sludge settling performance, and improved oxygen transfer efficiency as compared to shorter SRTs of 10 and 20 day. The findings underscore the need to assess microbial morphology when quantifying the treatment performance and energy performance of activated sludge processes.

Introduction to Digital Image Analysis in Whole-slide Imaging: A White Paper from the Digital Pathology Association

The advent of whole-slide imaging in digital pathology has brought about the advancement of computer-aided examination of tissue via digital image analysis. Digitized slides can now be easily annotated and analyzed via a variety of algorithms. This study reviews the fundamentals of tissue image analysis and aims to provide pathologists with basic information regarding the features, applications, and general workflow of these new tools. The review gives an overview of the basic categories of software solutions available, potential analysis strategies, technical considerations, and general algorithm readouts. Advantages and limitations of tissue image analysis are discussed, and emerging concepts, such as artificial intelligence and machine learning, are introduced. Finally, examples of how digital image analysis tools are currently being used in diagnostic laboratories, translational research, and drug development are discussed.

Interlaboratory variability of Ki67 staining in breast cancer

BACKGROUND

Postanalytic issues of Ki67 assessment in breast cancers like counting method standardisation and interrater bias have been subject of various studies, but little is known about analytic variability of Ki67 staining between pathology labs. Our aim was to study interlaboratory variability of Ki67 staining in breast cancer using tissue microarrays (TMAs) and central assessment to minimise preanalytic and postanalytic influences.

METHODS

Thirty European pathology labs stained serial slides of a TMA set of breast cancer tissues with Ki67 according to their routine in-house protocol. The Ki67-labelling index (Ki67-LI) of 70 matched samples was centrally assessed by one observer who counted all cancer cells per sample. We then tested for differences between the labs in Ki67-LI medians by analysing variance on ranks and in proportions of tumours classified as luminal A after dichotomising oestrogen receptor-positive cancers into cancers showing low (<14%, luminal A) and high (≥14%, luminal B HER2 negative) Ki67-LI using Cochran's Q.

RESULTS

Substantial differences between the 30 labs were indicated for median Ki67-LI (0.65%-33.0%, p < 0.0001) and proportion of cancers classified as luminal A (17%-57%, p < 0.0001). The differences remained significant when labs using the same antibody (MIB-1, SP6, or 30-9) were analysed separately or labs without prior participation in external quality assurance programs were excluded (p < 0.0001, respectively).

CONCLUSION

Substantial variability in Ki67 staining of breast cancer tissue was found between 30 routine pathology labs. Clinical use of the Ki67-LI for therapeutic decisions should be considered only fully aware of lab-specific reference values.

In vivo and clinical application of strontium-enriched biomaterials for bone regeneration: A systematic review

OBJECTIVES

This systematic review aimed to assess the in vivo and clinical effect of strontium (Sr)-enriched biomaterials in bone formation and/or remodelling.

METHODS

A systematic search was performed in Pubmed, followed by a two-step selection process. We included in vivo original studies on Sr-containing biomaterials used for bone support or regeneration, comparing at least two groups that only differ in Sr addition in the experimental group.

RESULTS

A total of 572 references were retrieved and 27 were included. Animal models were used in 26 articles, and one article described a human study. Osteoporotic models were included in 11 papers. All articles showed similar or increased effect of Sr in bone formation and/or regeneration, in both healthy and osteoporotic models. No study found a decreased effect. Adverse effects were assessed in 17 articles, 13 on local and four on systemic adverse effects. From these, only one reported a systemic impact from Sr addition. Data on gene and/or protein expression were available from seven studies.

CONCLUSIONS

This review showed the safety and effectiveness of Sr-enriched biomaterials for stimulating bone formation and remodelling in animal models. The effect seems to increase over time and is impacted by the concentration used. However, included studies present a wide range of study methods. Future work should focus on consistent models and guidelines when developing a future clinical application of this element.Cite this article: N. Neves, D. Linhares, G. Costa, C. C. Ribeiro, M. A. Barbosa. In vivo and clinical application of strontium-enriched biomaterials for bone regeneration: A systematic review. Bone Joint Res 2017;6:366-375. DOI: 10.1302/2046-3758.66.BJR-2016-0311.R1.

Ki67/SATB1 ratio is an independent prognostic factor of overall survival in patients with early hormone receptor-positive invasive ductal breast carcinoma

Biological diversity of breast cancer presents challenges for personalized therapy and necessitates multiparametric approaches to understand and manage the disease. Multiple protein biomarkers tested by immunohistochemistry (IHC), followed by digital image analysis and multivariate statistics of the data, have been shown to be effective in exploring latent profiles of tumor tissue immunophenotype. In this study, based on tissue microarrays of 107 patients with hormone receptor (HR) positive invasive ductal breast carcinoma, we investigated the prognostic value of the integrated immunophenotype to predict overall survival (OS) of the patients. A set of 10 IHC markers (ER, PR, HER2, Ki67, AR, BCL2, HIF-1α, SATB1, p53, and p16) was used. The main factor of the variance was characterized by opposite loadings of ER/PR/AR/BCL2 and Ki67/HIF-1α; it was associated with histological grade but did not predict OS. The second factor was driven by SATB1 expression along with moderate positive HIF-1α and weak negative Ki67 loadings. Importantly, this factor did not correlate with any clinicopathologic parameters, but was an independent predictor of better OS. Ki67 and SATB1 did not reach statistical significance as single predictors; however, high Ki67/SATB1 ratio was an independent predictor of worse OS. In addition, our data indicate potential double prognostic meaning of HIF-1α expression in breast cancer and necessitate focused studies, taking into account the immunophenotype interactions and tissue heterogeneity aspects.

Commentary

Historically, pathologists perform manual evaluation of H&E- or immunohistochemically-stained slides, which can be subjective, inconsistent, and, at best, semiquantitative. As the complexity of staining and demand for increased precision of manual evaluation increase, the pathologist’s assessment will include automated analyses (i.e., “digital pathology”) to increase the accuracy, efficiency, and speed of diagnosis and hypothesis testing and as an important biomedical research and diagnostic tool. This commentary introduces the many roles for pathologists in designing and conducting high-throughput digital image analysis. Pathology review is central to the entire course of a digital pathology study, including experimental design, sample quality verification, specimen annotation, analytical algorithm development, and report preparation. The pathologist performs these roles by reviewing work undertaken by technicians and scientists with training and expertise in image analysis instruments and software. These roles require regular, face-to-face interactions between team members and the lead pathologist. Traditional pathology training is suitable preparation for entry-level participation on image analysis teams. The future of pathology is very exciting, with the expanding utilization of digital image analysis set to expand pathology roles in research and drug development with increasing and new career opportunities for pathologists.

Comprehensive Immunohistochemistry: Digital, Analytical and Integrated

Immunohistochemistry (IHC) is widely used in contemporary pathology as a diagnostic and, increasingly, as a prognostic and predictive tool. The main value of the method today comes from a sensitive and specific detection of a protein of interest in the context of tissue architecture and cell populations. One of the major limitations of conventional IHC is related to the fact that the results are usually obtained by visual qualitative or semiquantitative evaluation. While this is sufficient for diagnostic purposes, measurement of prognostic and predictive biomarkers requires better accuracy and reproducibility. Also, objective evaluation of the spatial heterogeneity of biomarker expression as well as the development of combined/integrated biomarkers are in great demand. On the other end of the scale, the rapid development of tissue proteomics accounting for 2D spatial aspects has led to a disruptive concept of next-generation IHC, promising high multiplexing and broad dynamic range quantitative/spatial data on tissue protein expression. This 'evolutionary gap' between conventional and next-generation IHC can be filled by comprehensive IHC based on digital technologies (empowered by quantification and spatial and multiparametric analytics) and integrated into the pathology workflow and information systems. In this paper, we share our perspectives on a comprehensive IHC road map as a multistep development process.

A Heuristic Framework for Image Filtering and Segmentation: Application to Blood Vessel Immunohistochemistry

The blood vessel density in a cancerous tissue sample may represent increased levels of tumor growth. However, identifying blood vessels in the histological (tissue) image is difficult and time-consuming and depends heavily on the observer's experience. To overcome this drawback, computer-aided image analysis frameworks have been investigated in order to boost object identification in histological images. We present a novel algorithm to automatically abstract the salient regions in blood vessel images. Experimental results show that the proposed framework is capable of deriving vessel boundaries that are comparable to those demarcated manually, even for vessel regions with weak contrast between the object boundaries and background clutter.

A methodology for comprehensive breast cancer Ki67 labeling index with intra-tumor heterogeneity appraisal based on hexagonal tiling of digital image analysis data

Digital image analysis (DIA) enables higher accuracy, reproducibility, and capacity to enumerate cell populations by immunohistochemistry; however, the most unique benefits may be obtained by evaluating the spatial distribution and intra-tissue variance of markers. The proliferative activity of breast cancer tissue, estimated by the Ki67 labeling index (Ki67 LI), is a prognostic and predictive biomarker requiring robust measurement methodologies. We performed DIA on whole-slide images (WSI) of 302 surgically removed Ki67-stained breast cancer specimens; the tumour classifier algorithm was used to automatically detect tumour tissue but was not trained to distinguish between invasive and non-invasive carcinoma cells. The WSI DIA-generated data were subsampled by hexagonal tiling (HexT). Distribution and texture parameters were compared to conventional WSI DIA and pathology report data. Factor analysis of the data set, including total numbers of tumor cells, the Ki67 LI and Ki67 distribution, and texture indicators, extracted 4 factors, identified as entropy, proliferation, bimodality, and cellularity. The factor scores were further utilized in cluster analysis, outlining subcategories of heterogeneous tumors with predominant entropy, bimodality, or both at different levels of proliferative activity. The methodology also allowed the visualization of Ki67 LI heterogeneity in tumors and the automated detection and quantitative evaluation of Ki67 hotspots, based on the upper quintile of the HexT data, conceptualized as the "Pareto hotspot". We conclude that systematic subsampling of DIA-generated data into HexT enables comprehensive Ki67 LI analysis that reflects aspects of intra-tumor heterogeneity and may serve as a methodology to improve digital immunohistochemistry in general.

A methodology to ensure and improve accuracy of Ki67 labelling index estimation by automated digital image analysis in breast cancer tissue

Introduction

Immunohistochemical Ki67 labelling index (Ki67 LI) reflects proliferative activity and is a potential prognostic/predictive marker of breast cancer. However, its clinical utility is hindered by the lack of standardized measurement methodologies. Besides tissue heterogeneity aspects, the key element of methodology remains accurate estimation of Ki67-stained/counterstained tumour cell profiles. We aimed to develop a methodology to ensure and improve accuracy of the digital image analysis (DIA) approach.

Methods

Tissue microarrays (one 1-mm spot per patient, n = 164) from invasive ductal breast carcinoma were stained for Ki67 and scanned. Criterion standard (Ki67-Count) was obtained by counting positive and negative tumour cell profiles using a stereology grid overlaid on a spot image. DIA was performed with Aperio Genie/Nuclear algorithms. A bias was estimated by ANOVA, correlation and regression analyses. Calibration steps of the DIA by adjusting the algorithm settings were performed: first, by subjective DIA quality assessment (DIA-1), and second, to compensate the bias established (DIA-2). Visual estimate (Ki67-VE) on the same images was performed by five pathologists independently.

Results

ANOVA revealed significant underestimation bias (P < 0.05) for DIA-0, DIA-1 and two pathologists’ VE, while DIA-2, VE-median and three other VEs were within the same range. Regression analyses revealed best accuracy for the DIA-2 (R-square = 0.90) exceeding that of VE-median, individual VEs and other DIA settings. Bidirectional bias for the DIA-2 with overestimation at low, and underestimation at high ends of the scale was detected. Measurement error correction by inverse regression was applied to improve DIA-2-based prediction of the Ki67-Count, in particular for the clinically relevant interval of Ki67-Count < 40%. Potential clinical impact of the prediction was tested by dichotomising the cases at the cut-off values of 10, 15, and 20%. Misclassification rate of 5-7% was achieved, compared to that of 11-18% for the VE-median-based prediction.

Conclusions

Our experiments provide methodology to achieve accurate Ki67-LI estimation by DIA, based on proper validation, calibration, and measurement error correction procedures, guided by quantified bias from reference values obtained by stereology grid count. This basic validation step is an important prerequisite for high-throughput automated DIA applications to investigate tissue heterogeneity and clinical utility aspects of Ki67 and other immunohistochemistry (IHC) biomarkers.

Digital immunohistochemistry wizard: image analysis-assisted stereology tool to produce reference data set for calibration and quality control

Background

Digital image analysis (DIA) enables better reproducibility of immunohistochemistry (IHC) studies. Nevertheless, accuracy of the DIA methods needs to be ensured, demanding production of reference data sets. We have reported on methodology to calibrate DIA for Ki67 IHC in breast cancer tissue based on reference data obtained by stereology grid count. To produce the reference data more efficiently, we propose digital IHC wizard generating initial cell marks to be verified by experts.

Methods

Digital images of proliferation marker Ki67 IHC from 158 patients (one tissue microarray spot per patient) with an invasive ductal carcinoma of the breast were used. Manual data (mD) were obtained by marking Ki67-positive and negative tumour cells, using a stereological method for 2D object enumeration. DIA was used as an initial step in stereology grid count to generate the digital data (dD) marks by Aperio Genie and Nuclear algorithms. The dD were collected into XML files from the DIA markup images and overlaid on the original spots along with the stereology grid. The expert correction of the dD marks resulted in corrected data (cD). The percentages of Ki67 positive tumour cells per spot in the mD, dD, and cD sets were compared by single linear regression analysis. Efficiency of cD production was estimated based on manual editing effort.

Results

The percentage of Ki67-positive tumor cells was in very good agreement in the mD, dD, and cD sets: regression of cD from dD (R²=0.92) reflects the impact of the expert editing the dD as well as accuracy of the DIA used; regression of the cD from the mD (R²=0.94) represents the consistency of the DIA-assisted ground truth (cD) with the manual procedure. Nevertheless, the accuracy of detection of individual tumour cells was much lower: in average, 18 and 219 marks per spot were edited due to the Genie and Nuclear algorithm errors, respectively. The DIA-assisted cD production in our experiment saved approximately 2/3 of manual marking.

Conclusions

Digital IHC wizard enabled DIA-assisted stereology to produce reference data in a consistent and efficient way. It can provide quality control measure for appraising accuracy of the DIA steps.

Histomorphometric analysis of collagen and elastic fibres in the cranial and caudal fold of the porcine glottis

The porcine glottis differs from the human glottis in its cranial and caudal vocal folds (CraF, CauF). The fibre apparatus of these folds was studied histomorphometrically in adult minipigs. For object definition and quantification, the colour-selection tools of the Adobe-Photoshop program were used. Another key feature was the subdivision of the cross-sections of the folds into proportional subunits. This allowed a statistical analysis irrespective of differences in thickness of the folds. Both folds had a distinct, dense subepithelial layer equivalent to the basement membrane zone in humans. The subsequent, loose layer was interpreted - in principle - as being equivalent to Reinke's space of the human vocal fold. The next two layers were not clearly separated. Due to this, the concept of a true vocal ligament did not appear applicable to neither CauF nor CraF. Instead, the body-cover model was emphasized by our findings. The missing vocalis muscle in the CraF is substituted by large collagen fibre bundles in a proportional depth corresponding to the position of the muscle of the CauF. The distribution of elastic fibres made the CraF rather than the CauF more similar to the human vocal fold. We suggest that these data are useful for those wishing to use the porcine glottis as a model for studying oscillatory properties during phonation.

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.

Pearls and pitfalls in neural CGRP immunohistochemistry

This review outlines the pearls and pitfalls of calcitonin-gene related protein (CGRP) immunohistochemistry of the brain.

Pearls

In 1985, CGRP was first described in cerebral arteries using immunohistochemistry. Since then, cerebral CGRP (and, using novel antibodies, its receptor components) has been widely scrutinized. Here, we describe the distribution of cerebral CGRP and pay special attention to the surprising reliability of results over time.

Pitfalls

Pitfalls might include a fixation procedure, antibody clone and dilution, and interpretation of results. Standardization of staining protocols and true quantitative methods are lacking. The use of computerized image analysis has led us to believe that our examination is objective. However, in the steps of performing such an analysis, we make subjective choices. By pointing out these pitfalls, we aim to further improve immunohistochemical quality.

Recommendations

Having a clear picture of the tissue/cell morphology is a necessity. A primary morphological evaluation with, for example, hematoxylin-eosin, helps to ensure that small changes are not missed and that background and artifactual changes, which may include vacuoles, pigments, and dark neurons, are not over-interpreted as compound-related changes. The antigen-antibody reaction appears simple and clear in theory, but many steps might go wrong. Remember that methods including the antigen-antibody complex rely on handling/fixation of tissues or cells, antibody shipping/storing issues, antibody titration, temperature/duration of antibody incubation, visualization of the antibody and interpretation of the results. Optimize staining protocols to the material you are using.

Automatic nonsubjective estimation of antigen content visualized by immunohistochemistry using color deconvolution

We describe a method for the automatic, nonsubjective estimation of 3,3' diaminobenzidine (DAB) in digital images obtained from routine central nervous system immunohistochemistry using freely available, platform-independent public domain image processing software. This technique estimates the amount of antigen visualized but does not measure antigen content directly. Combined with whole brain section high-resolution scanning, a "virtual dissection" (extracting the region of interest) makes it possible to estimate relative antigen content in either subcellular structures, specific brain regions, or in whole tissue sections at magnifications up to 40×. The digital image is processed using Ruifrok and Johnston's color deconvolution method to separate the brown DAB chromogen from the hematoxylin counterstain on a microscope slide. A monochrome image representing the DAB content is then subjected to frequency analysis using NIH-ImageJ and a weighting calculation to estimate the amount of DAB (antigen) as a dimensionless index. The method described produces results that agree with enzyme-linked immunosorbent assays, and is automatic and nonsubjective. The method could easily be adapted to other types of tissue or cell cultures.

Immunohistochemistry profiles of breast ductal carcinoma: factor analysis of digital image analysis data

BACKGROUND

Molecular studies of breast cancer revealed biological heterogeneity of the disease and opened new perspectives for personalized therapy. While multiple gene expression-based systems have been developed, current clinical practice is largely based upon conventional clinical and pathologic criteria. This gap may be filled by development of combined multi-IHC indices to characterize biological and clinical behaviour of the tumours. Digital image analysis (DA) with multivariate statistics of the data opens new opportunities in this field.

METHODS

Tissue microarrays of 109 patients with breast ductal carcinoma were stained for a set of 10 IHC markers (ER, PR, HER2, Ki67, AR, BCL2, HIF-1α, SATB1, p53, and p16). Aperio imaging platform with the Genie, Nuclear and Membrane algorithms were used for the DA. Factor analysis of the DA data was performed in the whole group and hormone receptor (HR) positive subgroup of the patients (n = 85).

RESULTS

Major factor potentially reflecting aggressive disease behaviour (i-Grade) was extracted, characterized by opposite loadings of ER/PR/AR/BCL2 and Ki67/HIF-1α. The i-Grade factor scores revealed bimodal distribution and were strongly associated with higher Nottingham histological grade (G) and more aggressive intrinsic subtypes. In HR-positive tumours, the aggressiveness of the tumour was best defined by positive Ki67 and negative ER loadings. High Ki67/ER factor scores were strongly associated with the higher G and Luminal B types, but also were detected in a set of G1 and Luminal A cases, potentially indicating high risk patients in these categories. Inverse relation between HER2 and PR expression was found in the HR-positive tumours pointing at differential information conveyed by the ER and PR expression. SATB1 along with HIF-1α reflected the second major factor of variation in our patients; in the HR-positive group they were inversely associated with the HR and BCL2 expression and represented the major factor of variation. Finally, we confirmed high expression levels of p16 in Triple-negative tumours.

CONCLUSION

Factor analysis of multiple IHC biomarkers measured by automated DA is an efficient exploratory tool clarifying complex interdependencies in the breast ductal carcinoma IHC profiles and informative value of single IHC markers. Integrated IHC indices may provide additional risk stratifications for the currently used grading systems and prove to be useful in clinical outcome studies.

VIRTUAL SLIDES

The virtual slide(s) for this article can be found here: http://www.diagnosticpathology.diagnomx.eu/vs/1512077125668949.

Embedding topic discovery in conditional random fields model for segmenting nuclei using multispectral data

Segmentation of cells/nuclei is a challenging problem in 2-D histological and cytological images. Although a large number of algorithms have been proposed, newer efforts continue to be devoted to investigate robust models that could have high level of adaptability with regard to considerable amount of image variability. In this paper, we propose a multiclassification conditional random fields (CRFs) model using a combination of low-level cues (bottom-up) and high-level contextual information (top-down) for separating nuclei from the background. In our approach, the contextual information is extracted by an unsupervised topic discovery process, which efficiently helps to suppress segmentation errors caused by intensity inhomogeneity and variable chromatin texture. In addition, we propose a multilayer CRF, an extension of the traditional single-layer CRF, to handle high-dimensional dataset obtained through spectral microscopy, which provides combined benefits of spectroscopy and imaging microscopy, resulting in the ability to acquire spectral images of microscopic specimen. The approach is evaluated with color images, as well as spectral images. The overall accuracy of the proposed segmentation algorithm reaches 95% when applying multilayer CRF model to the spectral microscopy dataset. Experiments also show that our method outperforms seeded watershed, a widely used algorithm for cell segmentation.

Digital image analysis: a review of reproducibility, stability and basic requirements for optimal results

Digital image analysis (DIA) is increasingly implemented in histopathological research to facilitate truly quantitative measurements, decrease inter-observer variation and reduce hands-on time. Originally, efforts were made to enable DIA to reproduce manually obtained results on histological slides optimized for light microscopy and the human eye. With improved technical methods and the acknowledgement that computerized readings are different from analysis by human eye, recognition has been achieved that to really empower DIA, histological slides must be optimized for the digital 'eye', with reproducible results correlating with clinical findings. In this review, we focus on the basic expectations and requirements for DIA to gain wider use in histopathological research and diagnostics. With a reference to studies that specifically compare DIA with conventional methods, this review discusses reproducibility, application of stereology-based quantitative measurements, time consumption, optimization of histological slides, regions of interest selection and recent developments in staining and imaging techniques.

Membrane connectivity estimated by digital image analysis of HER2 immunohistochemistry is concordant with visual scoring and fluorescence in situ hybridization results: algorithm evaluation on breast cancer tissue microarrays

Introduction

The human epidermal growth factor receptor 2 (HER2) is an established biomarker for management of patients with breast cancer. While conventional testing of HER2 protein expression is based on semi-quantitative visual scoring of the immunohistochemistry (IHC) result, efforts to reduce inter-observer variation and to produce continuous estimates of the IHC data are potentiated by digital image analysis technologies.

Methods

HER2 IHC was performed on the tissue microarrays (TMAs) of 195 patients with an early ductal carcinoma of the breast. Digital images of the IHC slides were obtained by Aperio ScanScope GL Slide Scanner. Membrane connectivity algorithm (HER2-CONNECT™, Visiopharm) was used for digital image analysis (DA). A pathologist evaluated the images on the screen twice (visual evaluations: VE1 and VE2). HER2 fluorescence in situ hybridization (FISH) was performed on the corresponding sections of the TMAs. The agreement between the IHC HER2 scores, obtained by VE1, VE2, and DA was tested for individual TMA spots and patient's maximum TMA spot values (VE1max, VE2max, DAmax). The latter were compared with the FISH data. Correlation of the continuous variable of the membrane connectivity estimate with the FISH data was tested.

Results

The pathologist intra-observer agreement (VE1 and VE2) on HER2 IHC score was almost perfect: kappa 0.91 (by spot) and 0.88 (by patient). The agreement between visual evaluation and digital image analysis was almost perfect at the spot level (kappa 0.86 and 0.87, with VE1 and VE2 respectively) and at the patient level (kappa 0.80 and 0.86, with VE1max and VE2max, respectively). The DA was more accurate than VE in detection of FISH-positive patients by recruiting 3 or 2 additional FISH-positive patients to the IHC score 2+ category from the IHC 0/1+ category by VE1max or VE2max, respectively. The DA continuous variable of the membrane connectivity correlated with the FISH data (HER2 and CEP17 copy numbers, and HER2/CEP17 ratio).

Conclusion

HER2 IHC digital image analysis based on membrane connectivity estimate was in almost perfect agreement with the visual evaluation of the pathologist and more accurate in detection of HER2 FISH-positive patients. Most immediate benefit of integrating the DA algorithm into the routine pathology HER2 testing may be obtained by alerting/reassuring pathologists of potentially misinterpreted IHC 0/1+ versus 2+ cases.

Immunohistology--past, present, and future

The rapid development of immunohistochemistry, a morphology-based technique, has come about through refinements in detection systems and an increasing range of sensitive and specific antibodies that have allowed application of the technique to formalin-fixed, paraffin-embedded tissues. The introduction of heat-induced antigen retrieval has been a significant milestone to compliment these developments so that the immunohistochemistry is firmly entrenched as an indispensable adjunct to morphologic diagnosis. Although this ancillary stain was initially used in a qualitative manner, problems surrounding the many variables that influence antigen preservation in formalin-fixed, paraffin-embedded tissues were not a major issue and laboratories strived to optimize their staining protocols to the material they accessioned and processed. The advent of personalized medicine and targeted cancer treatment has imposed the need to quantitate the stain reaction product and has resulted in calls to standardize the process of immunostaining. A closer examination of the variables that influence the ability to show antigens in formalin-fixed, paraffin-embedded tissues revealed many important variables, particularly in the preanalytical phase of the assay, that are beyond the control of the accessioning laboratory. Although analytical factors have the potential to be standardized, the actions of many pivotal procedures including fixation and antigen retrieval are not completely understood. Postanalytical processes including threshold and cut-off values require consensus and standardization and it is clear that some of these goals can be achieved through the direction of national and international organizations associated with cancer diagnosis and treatment. With the ability to serve as a surrogate marker of many genetic abnormalities, immunohistochemistry enters a new era and the need to better understand some of the mechanisms fundamental to the technique become more pressing and the development of true quantitative assays is imperative. There is also an increasing appreciation that the technique highlights patterns of staining that reflect exquisite localization to organelles and tissue structures that are not appreciable in routine stains, adding a further dimension to morphologic diagnosis.