Immunohistochemistry Image Research Articles

ASER: Animal Sex Reversal Database

Sex reversal, representing extraordinary sexual plasticity during the life cycle, not only triggers reproduction in animals but also affects reproductive and endocrine system-related diseases and cancers in humans. Sex reversal has been broadly reported in animals; however, an integrated resource hub of sex reversal information is still lacking. Here, we constructed a comprehensive database named ASER (Animal Sex Reversal) by integrating sex reversal-related data of 18 species from teleostei to mammalia. We systematically collected 40,018 published papers and mined the sex reversal-associated genes (SRGs), including their regulatory networks, from 1611 core papers. We annotated homologous genes and computed conservation scores for whole genomes across the 18 species. Furthermore, we collected available RNA-seq datasets and investigated the expression dynamics of SRGs during sex reversal or sex determination processes. In addition, we manually annotated 550 in situ hybridization (ISH), fluorescence in situ hybridization (FISH), and immunohistochemistry (IHC) images of SRGs from the literature and described their spatial expression in the gonads. Collectively, ASER provides a unique and integrated resource for researchers to query and reuse organized data to explore the mechanisms and applications of SRGs in animal breeding and human health. The ASER database is publicly available at http://aser.ihb.ac.cn/.

Application of computer-generated images to train pattern recognition used in semiquantitative immunohistochemistry scoring.

This study aimed to clarify whether the pattern recognition involved in scoring proliferation fractions can be trained by abstract computerized images of virtual tissues. Twenty computer-generated images with randomly distributed blue or red dots were scored by 12 probands (all co-workers or collaborators of the Institute of Pathology, University of Bonn). Afterward, the probands underwent a training phase during which they received an immediate feedback on the actual rate of positivity after each image. Finally, the initial testing series was rescored. In a second round with 15 different probands, 20 Ki-67 immunohistochemistry images of tonsil tissue were scored, followed by the same training phase with computer-generated images, before the immunohistochemistry slides were scored again. Paired t-tests were used to compare the differences in mean rates pre- and post-training. Concerning computerized images, untrained probands scored the percentages of positive dots with a mean deviation from the true rates of 8.2%. Following training, the same testing series was scored significantly better with a mean deviation of 4.9% (mean improvement 3.3%, p < 0.001). Scoring real immunohistochemistry slides, the training with computerized images also improved correct estimations, albeit to a lesser degree (mean improvement 1%, p = 0.03). Abstract computerized images of virtual tissues may be a useful tool to train and improve the accuracy of pattern recognition involved in semiquantitative scoring of immunohistochemistry slides. As a side results, this study highlights the value of computer-generated images to verify the performance of image-analysis software.

Bimodal expression of Wnt5a in the tooth germ: A comparative study using in situ hybridization and immunohistochemistry

BackgroundDuring tooth development, Wnt5a, a member of the noncanonical Wnt ligand, is expressed prominently in the dental mesenchyme. However, the spatiotemporal profiles of Wnt5a protein production and distribution in tooth germs are largely unknown, which impairs elucidation of the Wnt5a-mediated regulatory mechanism of tooth development. MethodsWe performed analyzes of the spatiotemporal expression of Wnt5a in embryonic tooth germs (E11.5–E18.5) by using in situ hybridization and immunohistochemistry in parallel. The developmental stages of the embryonic tooth germs were determined by HE staining. In order to compare the spatiotemporal distribution patterns of Wnt5a mRNA-expressing cells and those of Wnt5a protein-expressing cells, serial frontal sections of paraffinized mouse embryo heads were used for the analyzes. When needed, the immunohistochemistry images were subjected to digital detection analysis of Wnt5a immunostaining signal using the WinROOF 2018 Ver. 4.19.0 image processing software program. ResultsThroughout the developmental process, cells expressing Wnt5a mRNA were found in various tissues including the dental follicle, dental papilla, inner and outer enamel epithelium, stratum intermediate, and stellate reticulum. However, odontoblasts differentiating and polarizing at E18.5 were the only cells representing an accumulation of Wnt5a protein in the apical region of the odontoblast process. The Wnt5a protein was undetectable in undifferentiated mesenchymal cells as well as any other cells positive for Wnt5a mRNA. ConclusionDifferentiating odontoblasts execute Wnt5a expression, the mode of which is distinct from that executed by the other cells constituting tooth germ. Change of the mode of Wnt5a expression begins to take place in the mesenchymal cells by E18.5, starting the elongation of the cytoplasmic process.

Formation of organoid-like structures in the decellularized rat testis.

Objective(s):In testis, the extracellular matrix (ECM) in addition to the supportive role for cells in the seminiferous epithelium, is also essential for the accurate functioning of these cells. Thus, using a decellularized testicular ECM (DTECM), as a scaffold for three-dimensional (3D) culture of testicular cells can mimic native ECM for studying in vitro spermatogenesis. Materials and Methods:The rat testis was decellularized via perfusion of 0.5% sodium dodecyl sulfate (SDS) for 48 hr, followed by 1% Triton X-100 for 6 hr, and then 1% DNase I for 1 hr. The efficiency of decellularization was evaluated by histology, immunohistochemistry (IHC), scanning electron microscopy (SEM), and MTT test. The prepared scaffolds were recellularized with testicular cells and cultured and assessed with hematoxylin-eosin (H&E) staining after two weeks. Results:Based on the H&E image, no trace of cell components could be observed in DTECM. IHC images demonstrated collagen types I and IV, laminin, and fibronectin were preserved. Masson’s trichrome and alcian blue staining revealed that collagen and glycosaminoglycans (GAGs) were retained, and the SEM image indicated that 3D testicular architecture remained after the decellularization process. Based on the results of the MTT test, DTECM was cytocompatible, and H&E images represented that DTECM supports testicular cell arrangements in seminiferous tubule-like structures (STLSs) and organoid-like structures (OLSs). Conclusion:The results showed that the applied protocol successfully decellularized the testis tissue of the rat. Therefore, these scaffolds may provide an appropriate vehicle for in vitro reconstruction of the seminiferous tubule.

A deep learning quantified stroma-immune score to predict survival of patients with stage II\u2013III colorectal cancer

BackgroundProfound heterogeneity in prognosis has been observed in colorectal cancer (CRC) patients with intermediate levels of disease (stage II–III), advocating the identification of valuable biomarkers that could improve the prognostic stratification. This study aims to develop a deep learning-based pipeline for fully automatic quantification of immune infiltration within the stroma region on immunohistochemical (IHC) whole-slide images (WSIs) and further analyze its prognostic value in CRC.MethodsPatients from two independent cohorts were divided into three groups: the development group (N = 200), the internal (N = 134), and the external validation group (N = 90). We trained a convolutional neural network for tissue classification of CD3 and CD8 stained WSIs. A scoring system, named stroma-immune score, was established by quantifying the density of CD3+ and CD8+ T-cells infiltration in the stroma region.ResultsPatients with higher stroma-immune scores had much longer survival. In the development group, 5-year survival rates of the low and high scores were 55.7% and 80.8% (hazard ratio [HR] for high vs. low 0.39, 95% confidence interval [CI] 0.24–0.63, P < 0.001). These results were confirmed in the internal and external validation groups with 5-year survival rates of low and high scores were 57.1% and 78.8%, 63.9% and 88.9%, respectively (internal: HR for high vs. low 0.49, 95% CI 0.28–0.88, P = 0.017; external: HR for high vs. low 0.35, 95% CI 0.15–0.83, P = 0.018). The combination of stroma-immune score and tumor-node-metastasis (TNM) stage showed better discrimination ability for survival prediction than using the TNM stage alone.ConclusionsWe proposed a stroma-immune score via a deep learning-based pipeline to quantify CD3+ and CD8+ T-cells densities within the stroma region on WSIs of CRC and further predict survival.

Spatial Profiles of Intratumoral PD-1+ Helper T Cells Predict Prognosis in Head and Neck Squamous Cell Carcinoma.

BackgroundFunctional interactions between immune cells and neoplastic cells in the tumor immune microenvironment have been actively pursued for both biomarker discovery for patient stratification, as well as therapeutic anti-cancer targets to improve clinical outcomes. Although accumulating evidence indicates that intratumoral infiltration of immune cells has prognostic significance, limited information is available on the spatial infiltration patterns of immune cells within intratumoral regions. This study aimed to understand the intratumoral heterogeneity and spatial distribution of immune cell infiltrates associated with cell phenotypes and prognosis in head and neck squamous cell carcinoma (HNSCC).MethodsA total of 88 specimens of oropharyngeal squamous cell carcinoma, categorized into discovery (n = 38) and validation cohorts (n = 51), were analyzed for immune contexture by multiplexed immunohistochemistry (IHC) and image cytometry-based quantification. Tissue segmentation was performed according to a mathematical morphological approach using neoplastic cell IHC images to dissect intratumoral regions into tumor cell nests versus intratumoral stroma.ResultsTissue segmentation revealed heterogeneity in intratumoral T cells, varying from tumor cell nest-polarized to intratumoral stroma-polarized distributions. Leukocyte composition analysis revealed higher ratios of TH1/TH2 in tumor cell nests with higher percentages of helper T cells, B cells, and CD66b+ granulocytes within intratumoral stroma. A discovery and validation approach revealed a high density of programmed death receptor-1 (PD-1)+ helper T cells in tumor cell nests as a negative prognostic factor for short overall survival. CD163+ tumor-associated macrophages (TAM) provided the strongest correlation with PD-1+ helper T cells, and cases with a high density of PD-1+ helper T cells and CD163+ TAM had a significantly shorter overall survival than other cases.ConclusionThis study reveals the significance of analyzing intratumoral cell nests and reports that an immune microenvironment with a high density of PD-1+ helper T cells in tumoral cell nests is a poor prognostic factor for HNSCC.

Incorporating label correlations into deep neural networks to classify protein subcellular location patterns in immunohistochemistry images.

Analysis of protein subcellular localization is a critical part of proteomics. In recent years, as both the number and quality of microscopic images are increasing rapidly, many automated methods, especially convolutional neural networks (CNN), have been developed to predict protein subcellular location(s) based on bioimages, but their performance always suffers from some inherent properties of the problem. First, many microscopic images have non-informative or noisy sections, like unstained stroma and unspecific background, which affect the extraction of protein expression information. Second, the patterns of protein subcellular localization are very complex, as a lot of proteins locate in more than one compartment. In this study, we propose a new label-correlation enhanced deep neural network, laceDNN, to classify the subcellular locations of multi-label proteins from immunohistochemistry images. The model uses small representative patches as input to alleviate the image noise issue, and its backbone is a hybrid architecture of CNN and recurrent neural network, where the former network extracts representative image features and the latter learns the organelle dependency relationships. Our experimental results indicate that the proposed model can improve the performance of multi-label protein subcellular classification.

An Image Pre-Processing Improvement Procedure for Standardizing Multiple Colour Variations in the Astrocytoma Immunohistochemical Staining Images

Colour variations in the histopathological images can occur when the acquired images came from different laboratories. This issue happens may be due to different protocols between each laboratory. Besides that, the usage of dye or staining is also different since it arrived from different manufacturers. Hence, it will affect the performance of a system, especially for an automated image diagnosis system. If this issue is taken lightly, the whole process of image analysis can produce a false diagnosis outcome, and thus, it will affect the patient’s treatment options and endanger the patient’s life. This paper proposed an automated image pre-processing procedure for standardizing the colour variations in immunohistochemical staining images. The proposed procedure was done by modifying the reference image, which will be used later for the stain standardization process. Based on 50 astrocytoma histopathological images, the proposed procedure showed a promising result, where the quality of the output images has been enhanced and standardized.

Pixelwise H-score: A novel digital image analysis-based metric to quantify membrane biomarker expression from immunohistochemistry images.

Immunohistochemistry (IHC) assays play a central role in evaluating biomarker expression in tissue sections for diagnostic and research applications. Manual scoring of IHC images, which is the current standard of practice, is known to have several shortcomings in terms of reproducibility and scalability to large scale studies. Here, by using a digital image analysis-based approach, we introduce a new metric called the pixelwise H-score (pix H-score) that quantifies biomarker expression from whole-slide scanned IHC images. The pix H-score is an unsupervised algorithm that only requires the specification of intensity thresholds for the biomarker and the nuclear-counterstain channels. We present the detailed implementation of the pix H-score in two different whole-slide image analysis software packages Visiopharm and HALO. We consider three biomarkers P-cadherin, PD-L1, and 5T4, and show how the pix H-score exhibits tight concordance to multiple orthogonal measurements of biomarker abundance such as the biomarker mRNA transcript and the pathologist H-score. We also compare the pix H-score to existing automated image analysis algorithms and demonstrate that the pix H-score provides either comparable or significantly better performance over these methodologies. We also present results of an empirical resampling approach to assess the performance of the pix H-score in estimating biomarker abundance from select regions within the tumor tissue relative to the whole tumor resection. We anticipate that the new metric will be broadly applicable to quantify biomarker expression from a wide variety of IHC images. Moreover, these results underscore the benefit of digital image analysis-based approaches which offer an objective, reproducible, and highly scalable strategy to quantitatively analyze IHC images.

Oncobox molecular profiling for a patient with recurrent endometrial cancer

Relevance. Endometrial cancer (EC) is the most common gynecological tumor. As a rule, it has a good prognosis, but in case of relapse, it worsens significantly. The effectiveness of cytotoxic chemotherapy for the treatment of such patients remains low.Purpose. To present a clinical case demonstrating the possibilities of therapy performed according to the results of molecular profiling of the tumor by RNA sequencing methods.Methods. Analysis of the anamnesis data, the results of histological and immunohistochemical studies, PET-CT, ultrasound images, and RNA sequencing data was carried out. The results of therapy prescribed in accordance with the rating of targeted drugs obtained after processing the transcriptomic profile by the Oncobox diagnostic platform were evaluated.Results. After experimental second-line therapy, a partial response was recorded.Conclusion. Early molecular profiling and therapy assignment in accordance with its results can change the course of the disease and improve the quality of life of patients.

Human myocardial mitochondrial oxidative capacity is impaired in mild acute heart transplant rejection.

AimsAcute cellular rejection (ACR) following heart transplantation (HTX) is associated with long‐term graft loss and increased mortality. Disturbed mitochondrial bioenergetics have been identified as pathophysiological drivers in heart failure, but their role in ACR remains unclear. We aimed to prove functional disturbances of myocardial bioenergetics in human heart transplant recipients with mild ACR by assessing myocardial mitochondrial respiration using high‐resolution respirometry, digital image analysis of myocardial inflammatory cell infiltration, and clinical assessment of HTX patients. We hypothesized that (i) mild ACR is associated with impaired myocardial mitochondrial respiration and (ii) myocardial inflammation, systemic oxidative stress, and myocardial oedema relate to impaired mitochondrial respiration and myocardial dysfunction.Methods and resultsWe classified 35 HTX recipients undergoing endomyocardial biopsy according International Society for Heart and Lung Transplantation criteria to have no (0R) or mild (1R) ACR. Additionally, we quantified immune cell infiltration by immunohistochemistry and digital image analysis. We analysed mitochondrial substrate utilization in myocardial fibres by high‐resolution respirometry and performed cardiovascular magnetic resonance (CMR). ACR (1R) was diagnosed in 12 patients (34%), while the remaining 23 patients revealed no signs of ACR (0R). Underlying cardiomyopathies (dilated cardiomyopathy 50% vs. 65%; P = 0.77), comorbidities (type 2 diabetes mellitus: 50% vs. 35%, P = 0.57; chronic kidney disease stage 5: 8% vs. 9%, P > 0.99; arterial hypertension: 59% vs. 30%, P = 0.35), medications (tacrolimus: 100% vs. 91%, P = 0.54; mycophenolate mofetil: 92% vs. 91%, P > 0.99; prednisolone: 92% vs. 96%, P > 0.99) and time post‐transplantation (21.5 ± 26.0 months vs. 29.4 ± 26.4 months, P = 0.40) were similar between groups. Mitochondrial respiration was reduced by 40% in ACR (1R) compared with ACR (0R) (77.8 ± 23.0 vs. 128.0 ± 33.0; P < 0.0001). Quantitative assessment of myocardial CD3+‐lymphocyte infiltration identified ACR (1R) with a cut‐off of >14 CD3+‐lymphocytes/mm2 (100% sensitivity, 82% specificity; P < 0.0001). Myocardial CD3+ infiltration (r = −0.41, P < 0.05), systemic oxidative stress (thiobarbituric acid reactive substances; r = −0.42, P < 0.01) and myocardial oedema depicted by global CMR derived T2 time (r = −0.62, P < 0.01) correlated with lower oxidative capacity and overt cardiac dysfunction (global longitudinal strain; r = −0.63, P < 0.01).ConclusionsMild ACR with inflammatory cell infiltration associates with impaired mitochondrial bioenergetics in cardiomyocytes. Our findings may help to identify novel checkpoints in cardiac immune metabolism as potential therapeutic targets in post‐transplant care.

Shifting Gears in Precision Oncology-Challenges and Opportunities of Integrative Data Analysis.

For decades, research relating to modification of host immunity towards antitumor response activation has been ongoing, with the breakthrough discovery of immune-checkpoint blockers. Several biomarkers with potential predictive value have been reported in recent studies for these novel therapies. However, with the plethora of therapeutic options existing for a given cancer entity, modern oncology is now being confronted with multifactorial interpretation to devise “the best therapy” for the individual patient. Into the bargain come the multiverse guidelines for established and emerging diagnostic biomarkers, as well as the complex interplay between cancer cells and tumor microenvironment, provoking immense challenges in the therapy decision-making process. Through this review, we present various molecular diagnostic modalities and techniques, such as genomics, immunohistochemistry and quantitative image analysis, which have the potential of becoming powerful tools in the development of an optimal treatment regime when analogized with patient characteristics. We will summarize the underlying complexities of these methods and shed light upon the necessary considerations and requirements for data integration. It is our hope to provide compelling evidence to emphasize on the need for inclusion of integrative data analysis in modern cancer therapy, and thereupon paving a path towards precision medicine and better patient outcomes.

Tumour-infiltrating CD4-, CD8- and FOXP3-positive immune cells as predictive markers of mortality in BRCA1- and BRCA2-associated breast cancer.

The prognostic value of tumour-infiltrating lymphocytes (TILs) in breast cancer is well-established. However, the investigation of specific T-cell subsets exclusively in BRCA-associated breast cancer is sparse. Tumour tissues from 414 BRCA-mutated breast cancer patients were analysed by immunohistochemistry and digital image analysis for expression of CD4, CD8 and FOXP3 immune markers. Distribution of CD4-, CD8- and FOXP3-positive cells and clinicopathological characteristics were assessed according to groups of low or high expression. The prognostic value was evaluated as continuous variables in univariate and multivariate analyses of overall survival and disease-free survival. Both CD4 and CD8 expression are associated with histological diagnosis, tumour grade and oestrogen and progesterone receptor expression status. CD4 expression is associated with BRCA gene status. A high percentage of tumour-infiltrating CD4-, CD8- or FOXP3-positive cells is significantly associated with lower mortality in BRCA1- and BRCA2-associated breast cancer and CD8-positive cells areassociated with disease-free survival. No heterogeneity according to BRCA gene status was found for the prognostic value of the immune markers. The results support a prognostic role of specific T-cell subsets in BRCA-associated breast cancer and the promising potential of targeting the immune system in the treatment of these patients.

Quantitative assessment of inflammatory infiltrates in kidney transplant biopsies using multiplex tyramide signal amplification and deep learning

Delayed graft function (DGF) is a strong risk factor for development of interstitial fibrosis and tubular atrophy (IFTA) in kidney transplants. Quantitative assessment of inflammatory infiltrates in kidney biopsies of DGF patients can reveal predictive markers for IFTA development. In this study, we combined multiplex tyramide signal amplification (mTSA) and convolutional neural networks (CNNs) to assess the inflammatory microenvironment in kidney biopsies of DGF patients (n = 22) taken at 6 weeks post-transplantation. Patients were stratified for IFTA development (<10% versus ≥10%) from 6 weeks to 6 months post-transplantation, based on histopathological assessment by three kidney pathologists. One mTSA panel was developed for visualization of capillaries, T- and B-lymphocytes and macrophages and a second mTSA panel for T-helper cell and macrophage subsets. The slides were multi spectrally imaged and custom-made python scripts enabled conversion to artificial brightfield whole-slide images (WSI). We used an existing CNN for the detection of lymphocytes with cytoplasmatic staining patterns in immunohistochemistry and developed two new CNNs for the detection of macrophages and nuclear-stained lymphocytes. F1-scores were 0.77 (nuclear-stained lymphocytes), 0.81 (cytoplasmatic-stained lymphocytes), and 0.82 (macrophages) on a test set of artificial brightfield WSI. The CNNs were used to detect inflammatory cells, after which we assessed the peritubular capillary extent, cell density, cell ratios, and cell distance in the two patient groups. In this cohort, distance of macrophages to other immune cells and peritubular capillary extent did not vary significantly at 6 weeks post-transplantation between patient groups. CD163+ cell density was higher in patients with ≥10% IFTA development 6 months post-transplantation (p < 0.05). CD3+CD8−/CD3+CD8+ ratios were higher in patients with <10% IFTA development (p < 0.05). We observed a high correlation between CD163+ and CD4+GATA3+ cell density (R = 0.74, p < 0.001). Our study demonstrates that CNNs can be used to leverage reliable, quantitative results from mTSA-stained, multi spectrally imaged slides of kidney transplant biopsies.This study describes a methodology to assess the microenvironment in sparse tissue samples. Deep learning, multiplex immunohistochemistry, and mathematical image processing techniques were incorporated to quantify lymphocytes, macrophages, and capillaries in kidney transplant biopsies of delayed graft function patients. The quantitative results were used to assess correlations with development of interstitial fibrosis and tubular atrophy.

Tumor microenvironment-adjusted prognostic implications of the KRAS mutation subtype in patients with stage III colorectal cancer treated with adjuvant FOLFOX

Several studies have reported that the prognostic effect of KRAS mutations on colorectal cancers (CRCs) varies depending on the type of mutation. Considering the effect of KRAS mutations on tumor microenvironment, we analyzed the prognostic significance of KRAS mutation types after adjusting for the tumor-infiltrating lymphocytes (TIL) and tumor-stromal percentage (TSP) statuses. In two independent cohorts, KRAS mutations were analyzed by Sanger sequencing and/or next-generation sequencing. TIL density and the TSP were quantified from whole-slide immunohistochemical images. KRAS-mutant CRCs were divided into three subgroups (G12D/V, other codon 12 mutations and codon 13 mutations) to examine their differential effect on TIL density, the TSP and recurrence-free survival (RFS). Among the KRAS mutations, only the G12D/V subgroups showed significantly less TIL infiltration than the wild-type CRCs. According to survival analysis, G12D/V mutations were associated with short RFS; codon 13 mutations showed discordant trends in the two cohorts, and other codon 12 mutations showed no significant association. Multivariate analysis further supported the prognostic value of G12D/V mutations. This result is not only consistent with a recent study suggesting the immunosuppressive effect of mutant KRAS but also provides insight into the type-specific prognostic effect of KRAS mutations.

Sensitive and Specific Immunohistochemistry Protocol for Nucleocapsid Protein from All Common SARS-CoV-2 Virus Strains in Formalin-Fixed, Paraffin Embedded Tissues.

Human coronavirus disease 2019 (COVID-19) is a life-threatening and highly contagious disease caused by coronavirus SARS-CoV-2. Sensitive and specific detection of SARS-CoV-2 viral proteins in tissues and cells of COVID-19 patients will support investigations of the biologic behavior and tissue and cell tropism of this virus. We identified commercially available affinity-purified polyclonal antibodies raised against nucleocapsid and spike proteins of SARS-CoV-2 that provide sensitive and specific detection of the virus by immunohistochemistry in formalin-fixed, paraffin-embedded tissue. Two immunohistochemistry protocols are presented that are mutually validated by the matched detection patterns of the two distinct viral antigens in virus-infected cells within autopsy lung tissue of COVID-19 deceased patients. Levels of nucleocapsid protein in the lungs of COVID-19 decedents, as measured by quantitative histo-cytometry of immunohistochemistry images, showed an excellent log–linear relationship with levels of viral nucleocapsid RNA levels, as measured by qRT-PCR. Importantly, since the nucleocapsid protein sequence is conserved across all known viral strains, the nucleocapsid immunohistochemistry protocol is expected to recognize all common variants of SARS-CoV-2. Negative controls include autopsy lung tissues from patients who died from non-COVID-19 respiratory disease and control rabbit immunoglobulin. Sensitive detection of SARS-CoV-2 in human tissues will provide insights into viral tissue and cell distribution and load in patients with active infection, as well as provide insight into the clearance rate of virus in later COVID-19 disease stages. The protocols are also expected to be readily transferable to detect SARS-CoV-2 proteins in tissues of experimental animal models or animals suspected to serve as viral reservoirs.

Abstract 193: A deep learning workflow to quantify the H-Score on immunohistochemistry images

Abstract Quantification of protein expression in IHC is typically assessed with H-scores, calculated by a pathologist on a scale from 0 to 300 based on the intensity of carcinoma cells and their relative proportion.The subjectivity of the methodology is an inherent risk of inconsistency for inter- intra- assay. Here, we developed a semi-automated workflow based on deep learning to calculate the H-score on IHC slides for the CDH17 protein, a component of the gastrointestinal (GI) system. CDH17 membranous expression makes its quantification particularly challenging and prone to variability. This workflow was validated on 112 samples of GI tumors (primary as well as metastatic), reaching an H-score correlation score R2=0.92 (p-value below 10-5) and an overall median absolute error=17.5 compared to a pathologist's evaluation. In our method, tumor regions on each slide image were first annotated by a pathologist, containing clusters of epithelial carcinoma mixed with non-tumor tissue. A segmentation model was trained to outline contours of marked carcinoma clusters, later classified as a low, moderate or strong expression based on optical density distribution. Finally, all cluster results were aggregated at slide level to get a final H-score. The method was evaluated on 112 tumor sections including the following cancer histotypes: pancreatic (N=20), gastric (N=18), esophageal (N=20), cholangiocarcinoma (N=20) and colorectal (N=34). A U-Net segmentation model with a ResNet101 encoder was trained on 303 annotated regions of interest (ROIs) of size 240x240 µm2 taken from 6 images, scanned at 20x. The model was taught to segment CDH17- and CDH17+ carcinoma clusters with 80% of the annotated ROIs. Once trained, the model reached a mIoU=0.74 and mDice=0.85 on a previously unseen evaluation set compound of the remaining 20% ROIs. The automatic method was then validated to attribute a score to each segmented CDH17+ cluster. DAB expressions were extracted for each cluster using color deconvolution and the best classification performances were obtained using the 90th percentile of the DAB distribution (Kappa score=0.74) compared to pathologist scoring. Finally, the whole workflow was launched on the entire image dataset to get predicted H-scores and compare them to the pathologist ones. The workflow reached a correlation R2=0.92 and a median absolute error=17.5 over all tumor locations. Our workflow reproduces the pathologist H-Score and shows a consistent performance on low- and high- range H-Scores (i.e. close to 0 and superior to 200), across all tumor locations of the study. This methodology fosters robustness and reproducibility of H-score quantification results for the CDH17 biomarker. Our workflow can straightforwardly be applied to other histological biomarkers, aiming to increase the rigorousness of its validation for future application in patient selection assays. Citation Format: Paul Klein, Thomas Le Meur, Roberto Fiorelli, Claudia Buehnemann, Aurelie Auguste, Sylvain Berlemont. A deep learning workflow to quantify the H-Score on immunohistochemistry images [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 193.

Abstract 2776: Sequential immunohistochemistry and computational image analysis for a deeper characterization of tumor-infiltrating myeloid cells

Abstract Tumor-infiltrating macrophages (TIMs) often outnumber other immune cells and across different tumor histologies have been linked to resistance to therapies, including immunotherapy. Despite the huge steps headed in the characterization of these cells, their phenotyping and quantification within the tumor micro-environment (TME) still remain challenging. Multiplex Immunohistochemistry/Immunofluorescence (mIHC/IF) technologies, which allow the simultaneous detection of multiple markers on a single tissue section have empowered the approach to study the TME, allowing the analysis of co-expressing molecules and the spatial relationship between distinct immune and non-immune cells. However, this technique well performed for lymphoid cells and nuclear markers, needs to be optimized for TIMs, often characterized by an elongated shape and an enlarged cytoplasm. mIHC by iterative immunostaining on a single paraffin-embedded (FFPE) tissue slide followed by whole slide digital pathology analysis was performed. Different TIM-related antibodies were used, including: CD68, CD206, CD163, CD14, MS4A4A, HLA-DR. Moreover, an array of checkpoint molecules (B7-H3, VISTA, TIM-3, PD-L1), an inflammation-related marker (STING), a pan-leukocyte marker (CD45) and lymphoid related markers (CD56, CD8, CD3, CD4) have been also validated. Together with the staining procedure, we developed computational pipelines for the processing of the whole-slide images. To perform a precise ascription of the staining intensity, we used a customized CellProfiler pipeline for segmenting single cells after using the Color Deconvolution ImageJ plugin for the generation of combined pseudo-fluorescent images. This segmentation pipeline permits the conversion of microscopy data into CSV files reporting the single-cell feature values (e.g. staining intensity and reciprocal position). With this approach IHC images can be visualized in user-friendly multi-parameter cytometric-like outputs allowing a comprehensive in situ TIM quantification and phenotyping. Nonetheless, our approach does not require specialized equipment and makes usage of commonly reagents, allowing its easy translation also in real-life clinical practice for the tailoring of therapeutic choices. In conclusion, in this study we discuss critical aspects and computational tools for data visualization and analysis of mIHC TIM data together with results obtained in different clinical setting (melanoma, breast cancer and lymphoma). We believe that this approach can be useful for a deeper characterization of the TME. Citation Format: Maria Maddalena Tumedei, Filippo Piccinini, Jenny Bulgarelli, Massimo Guidoboni, Francesco Limarzi, Barbara Vergani, Biagio Eugenio Leone, Maurizio Puccetti, Sara Bravaccini, Giovanni Martinelli, Toni Ibrahim, Marcella Tazzari. Sequential immunohistochemistry and computational image analysis for a deeper characterization of tumor-infiltrating myeloid cells [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 2776.

Utilizing Digital Pathology Informatics Algorithms for Diffuse Large B-Cell Lymphoma Subtyping

Introduction:Diffuse large B-cell lymphoma (DLBCL) is the most commonly occurring lymphoid malignancy in the United States. Through gene expression profiling (GEP), two biologically distinct molecular subgroups of DLBCL have been identified - germinal center B-cell (GCB) and activated B-cell - with disparate prognostic implications. Because GEP is not yet easily implemented in clinical practice, immunohistochemical (IHC) algorithms have been developed and validated to segment DLBCL into cell-of-origin subtypes. The most commonly employed algorithm uses CD10, BCL6 and MUM1 to distinguish GCB and non-GCB subtypes, but has been shown to misclassify up to 20% of cases. Advancements in digital pathology technology have led to the development of image analysis algorithms that allow for the extraction of quantitative descriptions of pathology image features such as level of IHC staining. We explore the feasibility of developing a precise, repeatable, and objective image analysis-based scoring algorithm for DLBCL whole slide images to replicate DLBCL subtype classification and prognostic assessment.Methods:Tissue slides of the immunostains included in the Hans algorithm (CD10, BCL6, MUM1) for 40 DLBCL patients were digitized at 40x objective resolution using Hamamatsu Nanozoomer 2.0 HT or Aperio AT2. We developed and trained an image analysis algorithm on the IHC whole slide images. The algorithm first masks the tissue area of the whole slide image, then performs superpixel segmentation, dividing tissue area into compact regions (60 microns square). Each region is normalized to a reference IHC image and has a color deconvolution algorithm applied to separate the hematoxylin and diaminobenzidine (DAB) stains. Features describing the distribution of both the hematoxylin and DAB stains are extracted using convolutional autoencoders. We then trained a classification algorithm using these features to classify the staining of these regions as positive or negative using machine learning techniques. Whereas pathologists examine specific regions in a high-power field and use visual estimation to determine if >30% of tumor cells are positive, our algorithm measures the percentage of positive regions over the entire DLBCL tissue. Receiver operating characteristics (ROC) curves were calculated to assess ability of image analysis-based measurements to predict pathologist classifications as positive or negative for CD10, BCL6, and MUM1. Pathologist classification of DLBCL subtype and positive or negative for each immunostain were obtained from pathology reports. Using thresholds for percent positive regions established by the ROC curves, patients were classified into GCB/non-GCB by the Hans algorithm using sequential application of the identified thresholds as shown in the Figure. The regions in the figure define the algorithm classification and the lines correspond to the pathologist classification displaying GCB in blue and non-GCB in red. Concordance with pathologist subtyping was calculated by agreement rate and Cohen's κ statistic.Results:Of the 40 patients included in our study, pathologists classified CD10, BCL6, and MUM1 as positive in 27.5%, 90%, and 84.9% respectively. Area under the ROC curve (AUC) for predicting pathologist classifications of positivity from image analysis measurements for CD10, BCL6, and MUM1 were 0.92, 1.0, and 0.95 respectively. Thresholds from the ROC curves for CD10, BCL6, and MUM1 were 13%, 15%, and 20% respectively. Using these thresholds, classification by image analysis algorithm was concordant with pathologist classification in 82.5% (κ = 0.65) of cases.Discussion:Pathologist IHC classification is limited by interobserver and interlaboratory variability, human error, difficulties with human eye distinguishing between low-intensity stains, and ordinal outputs. These limitations have led to inconsistent prognostic significance with IHC cell-of-origin subtyping. We developed a novel, objective image analysis scoring for predicting DLBCL subtype that uses whole slide images. Our image analysis algorithm may provide an effective decision-support tool for pathologist subtype classification and address known limitations. Future work will compare subtyping by the image analysis algorithm to gene expression profiling and inter-rater reliability of tissue microarrays and whole slide images performed by expert hematopathologists. [Display omitted] DisclosuresFlowers:National Institutes Of Health: Research Funding; Burroughs Welcome Fund: Research Funding; Celgene: Consultancy, Research Funding; Gilead: Consultancy; Research to Practice: Research Funding; Infinity: Research Funding; Onyx: Research Funding; Seattle Genetics: Consultancy; V Foundation: Research Funding; Genentech/Roche: Consultancy, Research Funding; OptumRx: Consultancy; Bayer: Consultancy; Millennium/Takeda: Research Funding; Eastern Cooperative Oncology Group: Research Funding; Clinical Care Options: Research Funding; National Cancer Institute: Research Funding; TG Therapeutics: Research Funding; Acerta: Research Funding; Prime Oncology: Research Funding; Educational Concepts: Research Funding; Spectrum: Consultancy; Pharmacyclics LLC, an AbbVie Company: Research Funding; Janssen Pharmaceutical: Research Funding; Abbvie: Consultancy, Research Funding. Cooper:Varian Medical Systems: Honoraria.

Pixelwise H-score: a novel digital image analysis based-metric to quantify membrane biomarker expression from immunohistochemistry images

Pixelwise H-score: a novel digital image analysis based-metric to quantify membrane biomarker expression from immunohistochemistry images