bioRxiv - Pathology最新文献

{"title":"Spatial and Single Cell Mapping of Castleman Disease Reveals Key Stromal Cell Types and Cytokine Pathways","authors":"David Smith, Anna Eichinger, Andrew Rech, Julia Wang, Eduardo Esteva, Arta Seyedian, Xiaoxu Yang, Mei Zhang, Dan Martinez, Kai Tan, Minjie Luo, Christopher Park, Boris Reizis, Vinodh Pillai","doi":"10.1101/2024.09.09.609717","DOIUrl":"https://doi.org/10.1101/2024.09.09.609717","url":null,"abstract":"Castleman disease (CD) is inflammatory lymphoproliferative disorder of unclear etiology. To determine the cellular and molecular basis of CD, we analyzed the spatial proteome of 4,485,009 single cells, transcriptome of 50,117 single nuclei, immune repertoire of 8187 single nuclei, and pathogenic mutations in Unicentric CD, idiopathic Multicentric CD, HHV8-associated MCD, and reactive lymph nodes. CD was characterized by increased non-lymphoid and stromal cells that formed unique microenvironments where they interacted with lymphoid cells. Interaction of activated follicular dendritic cell (FDC) cytoplasmic meshworks with mantle zone B cells was associated with B cell activation and differentiation. VEGF, IL-6, MAPK, and extracellular matrix pathways were elevated in stromal cells of CD. CXCL13+ FDCs, PDGFRA+ T-zone reticular cells (TRC), and ACTA2-positive perivascular reticular cells (PRC) were identified as the predominant source of increased VEGF expression and IL-6 signaling in CD. VEGF expression by FDCs was associated with peri-follicular neovascularization. FDC, TRC and PRC of CD activated JAK-STAT, TGF-bet;, and MAPK pathways via ligand-receptor interactions involving collagen, integrins, complement components, and VEGF receptors. T, B and plasma cells were polyclonal but showed class-switched and somatically hypermutated IgG1+ plasma cells consistent with stromal cell-driven germinal center activation. In conclusion, our findings show that stromal cell activation and associated B-cell activation and differentiation, neovascularization and stromal remodeling underlie CD and suggest new targets for treatment.","PeriodicalId":501471,"journal":{"name":"bioRxiv - Pathology","volume":"30 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-09-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142182950","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Geographic characterization of RPE structure and lipid changes in the PEX1-p.Gly844Asp mouse model for Zellweger spectrum disorder.","authors":"Samy Omri, Catherine Argyriou, Rachel Pryce, Erminia Di Pietro, Pierre Chaurand, Nancy E Braverman","doi":"10.1101/2024.09.05.611330","DOIUrl":"https://doi.org/10.1101/2024.09.05.611330","url":null,"abstract":"Peroxisome Biogenesis Disorders-Zellweger Spectrum (PBD-ZSD) are a heterogenous group of autosomal recessive disorders caused by defects in PEX genes whose proteins are required for peroxisome assembly and function. Peroxisomes are ubiquitous organelles that play a critical role in complex lipid metabolism. Dysfunctional peroxisomes in ZSD cause multisystem effects, with progressive retinal degeneration (RD) leading to childhood blindness being one of the most frequent clinical findings. Despite progress in understanding the role of peroxisomes in normal cellular functions, much remains unknown about how their deficiency causes RD, and there is no treatment. To study RD pathophysiology in this disease, we used the knock-in PEX1-p.GlyG844Asp (G844D) mouse model of milder ZSD, which represents the common human PEX1-p.Gly843Asp allele. We previously reported diminished retinal function, functional vision, and neural retina structural defects in this model. Beyond the neural retina, structural defects in retinal pigment epithelium (RPE) have been reported in ZSD patients and murine models with single peroxisome enzyme deficiency, suggesting that RPE degeneration may contribute to overall RD progression in this disease. Here, we investigate the RPE phenotype in our PEX1-G844D mouse model, observing morphological, inflammatory, and lipid changes at 1, 3, and 6 months of age. We report that RPE cell degeneration appears at 3 months of age and worsens with time, starts in the dorsal pole, and is accompanied by subretinal inflammatory cell infiltration. We match these events with lipid remodelling using imaging mass spectrometry which allowed regional analysis specific to the RPE cell layer. We identified 47 lipid alterations that precede structural changes, 10 of which are localized to the dorsal pole. 32 of these lipid alterations persist to 3 months, with remodelling of the lipid signature at the dorsal pole. 14 new alterations occur concurrent with histological changes. Changes in peroxisome-dependent lipids detected by liquid chromatography tandem mass spectrometry (reduced docosahexanoic acid and increased very long chain lysophosphatidylcholines) are exacerbated over time. This study represents the first characterization of RPE in any animal model of ZSD, and the first in situ lipid analysis in any peroxisome-deficient tissue. Our findings reveal candidate lipid drivers that could be targeted to alleviate RD progression in ZSD, as well as candidate biomarkers that could be used to evaluate retinopathy progression and response to therapy.","PeriodicalId":501471,"journal":{"name":"bioRxiv - Pathology","volume":"1 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-09-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142223821","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Ablation of LRP6 in alpha-smooth muscle actin-expressing cells abrogates lung inflammation and fibrosis upon bleomycin-induced lung injury","authors":"Eun-Ah Sung, Mikhali G. Dozmorov, SuJeong Song, Theingi Aung, Min Hee Park, Patricia J. Sime, Wook-Jin Chae","doi":"10.1101/2024.09.05.611327","DOIUrl":"https://doi.org/10.1101/2024.09.05.611327","url":null,"abstract":"Low-density lipoprotein receptor-related protein 6 (LRP6) is a receptor for Wnt ligands. Tissue fibrosis is a progressive pathological process with excessive extracellular matrix proteins (ECM) deposition. Myofibroblasts, identified by alpha-smooth muscle actin (alphaSMA) expression, play an important role in tissue fibrosis by producing ECM production. Here we found that Wnt antagonist Dickkopf1 (DKK1) induced gene expressions associated with inflammation and fibrosis in lung fibroblasts. We demonstrated that genetic deletion of LRP6 in alphaSMA-expressing cells using Acta2-cre Lrp6 fl/fl (Lrp6 AKO) mice abrogated bleomycin (BLM)-induced lung inflammation and fibrosis phenotype, suggesting an important role of LRP6 in modulating inflammation and fibrotic processes in the lung. Our results highlight the crucial role of LRP6 in fibroblasts in regulating inflammation and fibrosis upon BLM-induced lung injury.","PeriodicalId":501471,"journal":{"name":"bioRxiv - Pathology","volume":"7 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-09-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142227742","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"The golden Syrian hamster (Mesocricetus auratus) as a model to decipher relevant pathogenic aspects of sheep-associated malignant catarrhal fever","authors":"Rosalie Fabian, Eleanor G Bentley, Adam Kirby, Parul Sharma, James P Stewart, Anja Kipar","doi":"10.1101/2024.09.06.611580","DOIUrl":"https://doi.org/10.1101/2024.09.06.611580","url":null,"abstract":"Malignant catarrhal fever (MCF) is an often fatal sporadic gammaherpesvirus-induced disease of ruminants with global relevance. Ovine gammaherpesvirus-2 (OvHV-2), with sheep as reservoir host, is a major cause of MCF in susceptible species. Despite extensive research on the molecular aspects of the disease, its pathogenesis is not yet fully understood. The present study re-established the Syrian golden hamster (Mesocricetus auratus) as amenable animal model of MCF and applied complementary in situ approaches to confirm recent findings in natural disease that could shed new light on pathogenetic aspects of MCF. These showed that systemic OvHV-2 infection is associated with T cell and macrophage dominated mononuclear infiltrates and vasculitis in various organs. Both T cells and monocytes/macrophages harbor the virus, and infected leukocytes are abundant in the infiltrates. The results also indicate that OvHV-2 has a broader target cell spectrum, including vascular endothelial cells and selected squamous epithelia. The former supports the interpretation that the inflammatory processes develop due to circulating activated, infected T cells and monocytes that home to tissues and emigrate from vessels prone to leukocyte emigration, possibly with direct interaction between virus infected leukocytes and endothelial cells. The latter supports the hypothesis of a graft versus host disease scenario, without viral cytopathic effect on epithelial cells but infiltration of the mucosa by infected T cells and macrophages. The disease processes are accompanied by evidence of expansion of the T cell compartments and the monocyte/macrophage pool in lymphatic tissues and bone marrow.","PeriodicalId":501471,"journal":{"name":"bioRxiv - Pathology","volume":"36 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-09-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142182998","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Molecular and spatiotemporal characterization of cells in murine atherosclerotic plaques","authors":"Pengbo Hou, Zhanhong Liu, Jiankai Fang, Ziyi Wang, Shisong Liu, Shiqing Wang, Peishan Li, Gerry Melino, Yufang Shi, Changshun Shao","doi":"10.1101/2024.09.04.611323","DOIUrl":"https://doi.org/10.1101/2024.09.04.611323","url":null,"abstract":"Objective: Single-cell technologies have revolutionized our understanding of the phenotypic and transcriptional diversity of aortic leukocytes in atherosclerotic humans and mice. However, enzymatically dissociated tissues lose the spatial context of plaque cells in situ. Here we utilized imaging mass cytometry (IMC) combining with single-cell RNA sequencing (scRNA-seq) to characterize the spatial distribution dynamics, phenotypic transitions, metabolic and functional phenotypes, and the intercellular interaction networks of plaque cells during atherosclerotic progression. Additionally, the dynamic immune landscape of circulating leukocytes associated with atherosclerosis was characterized using cytometry of time of flight (CyTOF).\u0000Approach and Results: A highly multiplexed IMC panel with 33 metal-conjugated antibodies was designed to generate 11 highly multiplexed histology images of aortic root tissues from ApoE-/- mice on high-fat diet at different stage of atherosclerosis. Using histoCAT, we identified 8 principal cell subtypes with distinct phenotypic and geographic dynamics. Furthermore, IMC-defined cell subsets partially corresponded to scRNA-seq-annotated aortic cell subtypes, including 4 macrophage subsets, neutrophils, smooth muscle cells (SMCs) and SMC-derived SEMs (Stem cell, endothelial cell and macrophage-like cell). Activation of inflammatory pathways, increased oxidative phosphorylation and augmented osteoclast differentiation were observed in macrophage populations, SMCs and SEMs from an early stage to advanced stage of atherosclerosis. Notably, cell neighborhood analysis by IMC uncovered multifaceted cell-cell interactions within the plaque, in particular in neutrophil-mediated interactions with smooth muscle cells and macrophages, which were confirmed by ligand-receptor interactions based on scRNA-seq data. Additionally, characterization of the peripheral immune cells by CyTOF revealed an increased ratio of myeloid cells to lymphocytes, and certain neutrophil and monocyte subpopulations also exhibited enhanced lipid metabolism and glycolysis as well as activated inflammatory signaling.\u0000Conclusion: This study provides a dynamic spatiotemporal landscape of atherosclerotic lesions and peripheral leukocytes. The new information based on IMC may help understand atherosclerotic pathology and develop novel therapeutic strategies.","PeriodicalId":501471,"journal":{"name":"bioRxiv - Pathology","volume":"23 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-09-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142183036","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Natural Language Processing-like Deep Learning Aided in Identification and Validation of Thiosulfinate Tolerance Clusters in Diverse Bacteria","authors":"Brendon K Myers, Anuj Lamichhane, Brian H Kvitko, Bhabesh Dutta","doi":"10.1101/2024.09.03.611110","DOIUrl":"https://doi.org/10.1101/2024.09.03.611110","url":null,"abstract":"Allicin tolerance (alt) clusters in phytopathogenic bacteria, which provide resistance to thiosulfinates like allicin, are challenging to find using conventional approaches due to their varied architecture and the paradox of being vertically maintained within genera despite likely being horizontally transferred. This results in significant sequential diversity that further complicates their identification. Natural language processing (NLP) - like techniques, such as those used in DeepBGC, offers a promising solution by treating gene clusters like a language, allowing for identifying and collecting gene clusters based on patterns and relationships within the sequences. We curated and validated alt-like clusters in Pantoea ananatis 97-1R (PA), Burkholderia gladioli pv. gladioli FDAARGOS 389 (BG), and Pseudomonas syringae pv. tomato DC3000 (PTO). Leveraging sequences from the RefSeq bacterial database, we conducted comparative analyses of gene synteny, gene/protein sequences, protein structures, and predicted protein interactions. This approach enabled the discovery of several novel alt-like clusters previously undetectable by other methods, which were further validated experimentally. Our work highlights the effectiveness of NLP-like techniques for identifying underrepresented gene clusters and expands our understanding of the diversity and utility of alt-like clusters in diverse bacterial genera. This work demonstrates the potential of these techniques to simplify the identification process and enhance the applicability of biological data in real-world scenarios.","PeriodicalId":501471,"journal":{"name":"bioRxiv - Pathology","volume":"34 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-09-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142183037","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Harmonizing terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) with multiplexed iterative immunofluorescence enriches spatial contextualization of cell death","authors":"Marc Samuel Sherman, Thomas McMahon-Skates, Lindsey Sara Gaston, Wolfram Goessling, Joseph A Majzoub","doi":"10.1101/2024.09.04.611218","DOIUrl":"https://doi.org/10.1101/2024.09.04.611218","url":null,"abstract":"Terminal deoxynucleotidyl transferase dUTP Nick End Labeling (TUNEL) is an essential tool for the detection of cell death in tissues. Although TUNEL is not known to be compatible with multiplexed spatial proteomic methods, harmonizing TUNEL with such methods offers the opportunity to delineate cell-type specific cell death labeling and precise spatial contextualization of cell death in complex tissues. Here we evaluated variations of the TUNEL assay for their compatibility with a multiplexed immunofluorescence method, multiple iterative labeling by antibody neodeposition (MILAN), in two different tissues and injury models for cell death, acetaminophen-induced hepatocyte necrosis and dexamethasone-induced adrenocortical apoptosis. Using a commercial Click-iT-based assay as a standard, TUNEL signal could be reliably produced independent of antigen-retrieval method, with tissue-specific minor differences in signal-to-noise. In contrast, proteinase K treatment consistently reduced or even abrogated protein antigenicity, while pressure cooker treatment consistently enhanced protein antigenicity for the targets tested. Antibody-based TUNEL protocols using pressure-cooker antigen retrieval were MILAN erasure-compatible thus enabling harmonization of TUNEL with MILAN. As many as four staining cycles could be performed without loss of subsequent TUNEL signal, while first-round TUNEL did not influence protein antigenicity in subsequent rounds. We conclude this harmonized assay performs comparably to an established commercial assay, but preserves protein antigenicity, thus enabling versatile integration with multiplexed immunofluorescence using MILAN. We anticipate this harmonized protocol will enable broad and flexible integration of TUNEL into multiplexed spatial proteomic assays, thus vastly enhancing the spatial contextualization of cell death in complex tissues.","PeriodicalId":501471,"journal":{"name":"bioRxiv - Pathology","volume":"23 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-09-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142183045","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Vascular HIF2 signaling prevents cardiomegaly, alveolar congestion and capillary remodeling during chronic hypoxia","authors":"Teresa Albendea-Gomez, Susana Mendoza-Tamajon, Rosana Castro-Mecinas, Beatriz Escobar, Susana Ferreira Rocha, Sonia Urra-Balduz, Jose Angel Nicolas-Avila, Eduardo Oliver, Maria Villalba-Orero, Silvia Martin-Puig","doi":"10.1101/2024.09.03.610947","DOIUrl":"https://doi.org/10.1101/2024.09.03.610947","url":null,"abstract":"Hypoxia is associated with the onset of cardiovascular diseases including cardiac hypertrophy and pulmonary arterial hypertension (PAH). Endothelial HIF2 signaling mediates pulmonary arterial remodeling and subsequent right ventricular systolic pressure (RVSP) elevation during chronic hypoxia, encouraging novel therapeutic opportunities for PAH based on specific HIF2 inhibitors. Nevertheless, HIF2 relevance beyond the pulmonary endothelium or in the cardiac adaptation to hypoxia remains elusive. Wilms tumor 1 lineage contributes to heart and lung vascular compartments including pericytes, endothelial and smooth muscle cells. Here we describe the response to chronic hypoxia of a novel HIF2 mutant mouse model in the Wt1 lineage (Hif2/Wt1 cKO). Hif2/Wt1 cKO is protected against pulmonary remodeling and increased RVSP induced by hypoxia, but displays alveolar congestion, inflammation and hemorrhages associated with microvascular instability. Furthermore, lack of HIF2 in the Wt1 lineage leads to cardiomegaly, capillary remodeling, right and left ventricular hypertrophy, systolic dysfunction and left ventricular dilation, suggesting pulmonary-independent cardiac direct roles of HIF2 in hypoxia. These structural defects are partially restored upon reoxygenation, while functional parameters remain altered. Our results suggest that cardiopulmonary HIF2 signaling prevents excessive vascular proliferation during chronic hypoxia and define novel protective roles of HIF2 to warrant stable microvasculature and organ function.","PeriodicalId":501471,"journal":{"name":"bioRxiv - Pathology","volume":"7 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-09-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142183043","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Unsupervised learning for labeling global glomerulosclerosis","authors":"Hrafn Weishaupt, Justinas Besusparis, Cleo-Aron Weis, Stefan Porubsky, Arvydas Laurinavicius, Sabine Leh","doi":"10.1101/2024.09.01.610244","DOIUrl":"https://doi.org/10.1101/2024.09.01.610244","url":null,"abstract":"Current deep learning models for classifying glomeruli in nephropathology are trained almost exclusively in a supervised manner, requiring expert-labeled images. Very little is known about the potential for unsupervised learning to overcome this bottleneck. To address this open question in a proof-of-concept, the project focused on the most fundamental classification task: globally sclerosed versus non-globally sclerosed glomeruli. The performance of clustering between the two classes was extensively studied across a variety of labeled datasets with diverse compositions and histological stains, and across the feature embeddings produced by 34 different pre-trained CNN models. As demonstrated by the study, clustering of globally and non-globally sclerosed glomeruli is generally highly feasible, yielding accuracies of over 95% in most datasets. Further work will be required to expand these experiments towards the clustering of additional glomerular lesion categories. We are convinced that these efforts (i) will open up opportunities for semi-automatic labeling approaches, thus alleviating the need for labor-intensive manual labeling, and (ii) illustrate that glomerular classification models can potentially be trained even in the absence of expert-derived class labels.","PeriodicalId":501471,"journal":{"name":"bioRxiv - Pathology","volume":"19 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-09-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142183039","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Weakly Supervised Vector Quantization for Whole Slide Images Classification","authors":"Dawei Shen, Yao-zhong Zhang, Seiya Imoto","doi":"10.1101/2024.08.31.610626","DOIUrl":"https://doi.org/10.1101/2024.08.31.610626","url":null,"abstract":"Whole Slide Images (WSIs) are high-resolution digital scans of entire microscope slides, extensively used in pathology to enable detailed examination of tissue samples. WSI tumor classification is a classic application of Multiple Instance Learning (MIL). In this process, a WSI is first divided into image tiles, and each tile is encoded into an embedding vector using a pretrained vision encoder. A lightweight MIL model then aggregates all the embeddings in a WSI for classification. A key factor affecting the performance of this classification is the quality of the embedding vectors. However, the embedding vectors generated by the pretrained vision encoder are continuous and not task-specific, causing them to contain significant noise and resulting in low distinguishability between tumor tiles and normal tiles. This weakens the model's capability. In this work, inspired by VQ-VAE, we propose VQ-MIL, where each continuous embedding vector is mapped to a discrete, task-specific space using weakly supervised vector quantization. This approach effectively separates tumor instances from normal instances and reduces the noise associated with each instance. Our experiments demonstrate that our method achieves state-of-the-art classification results on two benchmark datasets.","PeriodicalId":501471,"journal":{"name":"bioRxiv - Pathology","volume":"2019 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-09-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142183040","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}