SEARCH · Search bioRxiv

Results for “Pathology”

Search indexed bioRxiv preprints in genomics, neuroscience, cell biology and bioinformatics. Read source abstracts and check manuscript versions; preprints are not peer reviewed.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 1,171 records · Page 65Linked to original sources

bioRxiv2026

Hierarchical Temporal Transformer for Cancer Grade Prediction and Cross-Cancer Transfer Learning from Pathology Reports

Background: Language models for cancer reports usually read one document at a time and are tested on tumour types seen in training. Continuous-time sequence models address the first gap, but whether they use visit timing, not just order, is seldom tested. Methods: We developed the Hierarchical Temporal Transformer (HTT), which encodes reports with adapted BiomedBERT and reads the sequence with a cancer-type-conditioned temporal transformer. We evaluated it against best-possible AUROC ceilings on two synthetic benchmarks, one with the label stated in the text and one recoverable only from tumour sizes and visit intervals (ten seeds). Diagnostics led to an interval-aware time code (HTT-IA). On 4,786 TCGA pathology reports in 14 types, three types were withheld, and a text-only HTT was compared with a survival model. Results: On the phrase benchmark HTT reached AUROC 0.931 +/- 0.012 against 0.913 +/- 0.007 for a last-report baseline (+0.018, 95% CI -0.009 to +0.050); dates cannot raise the ceiling there. On the growth benchmark the initial time code scored 0.790, equal to a no-time model (0.787), whereas HTT-IA reached 0.843 (+0.053, 0.026 to 0.082) against ceilings of 0.894 without and 0.979 with dates. On unseen tumour types HTT reached 0.925 +/- 0.007, equal to in-distribution performance and to a single-report baseline. Text-only macro C-index was 0.592 (CALM 0.626-0.652). Conclusion: Cross-type transfer is carried by the pre-trained encoder. Visit timing is used only when the data make it identifiable and the time code is on a clinical scale. Ceilings, order-only ablations and re-dating probes are inexpensive checks for such models.

bioinformatics↗

bioRxiv2016

Stratifying risk of prostate cancer recurrence following external beam radiation therapy: Comparing prostate MRI with prostate biopsy pathology.

Abbreviations Abbreviations INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSION Conflict of Interests Statement References (DIPL), Dominant intraprostatic lesion\n\n(DN), Dominant nodule\n\n(EBRT), External Beam Radiation Therapy\n\n(ECE), Extracapsular Extension\n\n(NVB), Neurovascular Bundle Invasion\n\n(SVI), Seminal Vesicle Invasion\n\n(MIBC), Maximum involvement of any biopsy core\n\n(PCV), Percentage of cancer volume\n\n(PPBC), Percentage of positive biopsy cores\n\n(NCCN), National Comprehensive Cancer Network\n\nINTRODUCTION Abbreviations INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSION ...

Pathology↗

bioRxiv2020

MicroRNA-574 Regulates FAM210A Expression and Influences Pathological Cardiac Remodeling

Aberrant synthesis of mitochondrial proteins impairs cardiac function and causes heart disease. However, the mechanism of regulation of mitochondria encoded protein expression during cardiac disease remains underexplored. Here, we have shown that multiple pathogenic cardiac stressors induce the expression of miR-574 guide and passenger strands (miR-574-5p/3p) in both humans and mice. miR-574 knockout mice exhibit severe cardiac disorder under heart disease-triggering stresses. miR-574-5p/3p mimics that are delivered systematically using nanoparticles reduce cardiac pathogenesis under disease insults. Transcriptome analysis of miR-574-null hearts uncovers FAM210A as a common target mRNA for both strands of miR-574. The interactome capture and translational state analyses suggest that FAM210A interacts with mitochondrial translation factors and regulates the protein expression of mitochondrial encoded electron transport chain genes. Using a human cardiomyocyte cell culture system, we discover that miR-574 regulates FAM210A expression and modulates mitochondrial encoded protein expression, which influences cardiac remodeling in heart failure.

pathology↗

bioRxiv2020

Whole Slide Preview Image Segmentation and Setup for Digital Pathology Scanners

The problem of tissue finding is of special interest in automating WSI scanners where it decomposes the preview image of tissue glass slides into a simplified and abstract level of localization and identification to setup WSI scanner for high-resolution scan. Prior to such scanning, a preview image is captured to calibrate the scanners parameters. Scan parameters such as focus depth and scan region are determined using a tissue finding software package. This paper introduces a series of pipelines (e.g. binary mask segmentation, tissue/artifact classification, region-of-interest allocation) to automate tissue preview segmentation in both brightfield and darkfield microscopy.

pathology↗

bioRxiv2020

Multiscale three-dimensional pathology findings of COVID-19 diseased lung using high-resolution cleared tissue microscopy

The study of pulmonary samples from individuals who have died as a direct result of COVID-19 infection is vital to our understanding of the pathogenesis of this disease. Histopathologic studies of lung tissue from autopsy of patients with COVID-19 specific mortality are only just emerging. All existing reports have relied on traditional 2-dimensional slide-based histological methods for specimen preparation. However, emerging methods for high-resolution, massively multiscale imaging of tissue microstructure using fluorescence labeling and tissue clearing methods enable the acquisition of tissue histology in 3-dimensions, that could open new insights into the nature of SARS-Cov-2 infection and COVID-19 disease processes. In this article, we present the first 3-dimensional images of lung autopsy tissues taken from a COVID-19 patient, including 3D "virtual histology" of cubic-millimeter volumes of the diseased lung, providing unique insights into disease processes contributing to mortality that could inform frontline treatment decisions.

pathology↗

bioRxiv2020

Repurposing Nintedanib for Pathological Cardiac Remodeling and Dysfunction

BackgroundHeart Failure (HF) is the leading cause of death worldwide. Myocardial fibrosis, one of the clinical manifestations implicated in almost every form of heart disease, contributes significantly to HF development. However, there is no approved drug specifically designed to target cardiac fibrosis. Nintedanib (NTB) is an FDA approved tyrosine kinase inhibitor for idiopathic pulmonary fibrosis (IPF) and chronic fibrosing interstitial lung diseases (ILD). The favorable clinical outcome of NTB in IPF patients is well established. Furthermore, NTB is well tolerated in IPF patients irrespective of cardiovascular comorbidities. However, there is a lack of direct evidence to support the therapeutic efficacy and safety of NTB in cardiac diseases. Methods and ResultsWe examined the effects of NTB treatment on cardiac fibrosis and dysfunction using a murine model of HF. Specifically, 10 weeks old C57BL/6J male mice were subjected to Transverse Aortic Constriction (TAC) surgery. NTB was administered once daily by oral gavage (50mg/kg) till 16 weeks post-TAC. Cardiac function was monitored by serial echocardiography. Histological analysis and morphometric studies were performed at 16 weeks post-TAC. In the control group, systolic dysfunction started developing from 4 weeks post-surgery and progressed till 16 weeks. However, NTB treatment prevented TAC-induced cardiac functional decline. In another experiment, NTB treatment was stopped at 8 weeks, and animals were followed till 16 weeks post-TAC. Surprisingly, NTBs beneficial effect on cardiac function was maintained even after treatment interruption. NTB treatment remarkably reduced cardiac fibrosis as confirmed by Massons trichome staining and decreased expression of collagen genes (COL1A1, COL3A1). Compared to TAC group, NTB treated mice showed lower HW/TL ratio and cardiomyocyte cross-sectional area. Our in vitro studies demonstrated that NTB prevents myofibroblast transformation, TGF{beta}1-induced SMAD3 phosphorylation, and production of fibrogenic proteins (Fibronectin-1). However, NTB significantly altered vital signaling pathways in both, isolated fibroblast and cardiomyocytes, suggesting that its biological effect and underlying cardiac protection mechanisms are not limited to fibroblast and fibrosis alone. ConclusionOur findings provide a proof of concept for repurposing NTB to combat adverse myocardial fibrosis and encourage the need for further validation in large animal models and subsequent clinical development for HF patients.

pathology↗

bioRxiv2021

Dual mTORC1/mTORC2 inhibition as a Host-Directed Therapeutic Target in Pathologically Distinct Mouse Models of Tuberculosis

Efforts to develop more effective and shorter-course therapies for tuberculosis have included a focus on host-directed therapy (HDT). The goal of HDT is to modulate the host response to infection, thereby improving immune defenses to reduce the duration of antibacterial therapy and/or the amount of lung damage. As a mediator of innate and adaptive immune responses involved in eliminating intracellular pathogens, autophagy is a potential target for HDT in tuberculosis. Because Mycobacterium tuberculosis modulates mammalian target of rapamycin (mTOR) signaling to impede autophagy, pharmacologic mTOR inhibition could provide effective HDT. mTOR exists within two distinct multiprotein complexes, mTOR complex-1 (mTORC1) and mTOR complex-2 (mTORC2). Rapamycin and its analogs only partially inhibit mTORC1. We hypothesized that novel mTOR kinase inhibitors blocking both complexes would have expanded therapeutic potential. We compared the effects of two mTOR inhibitors: rapamycin and the orally available mTOR kinase domain inhibitor CC214-2, which blocks both mTORC1 and mTORC2, as adjunctive therapies against murine TB, when added to the first-line regimen (RHZE) or the novel bedaquiline-pretomanid-linezolid (BPaL) regimen. Neither mTOR inhibitor affected lung CFU counts after 4-8 weeks of treatment when combined with BPaL or RHZE. However, addition of CC214-2 to BPaL and RHZE was associated with significantly fewer relapses in C3HeB/FeJ compared to addition of rapamycin and, in RHZE-treated mice, resulted in fewer relapses compared to RHZE alone. Therefore, CC214-2 and related mTOR kinase inhibitors may be more effective candidates for HDT than rapamycin analogs and may have the potential to shorten the duration of TB treatment.

pathology↗

bioRxiv2021

Interleukin-6 Drives Key Pathologic Outcomes in Experimental Acetaminophen-induced Liver Failure

Background and AimsIn severe cases of acetaminophen (APAP) overdose, acute liver injury rapidly progresses to acute liver failure (ALF), producing life-threatening complications including, hepatic encephalopathy (HE) and multi-organ failure (MOF). Systemic levels of interleukin-6 (IL-6) and IL-10 are highest in ALF patients with the most severe complications and the poorest prognosis. The mechanistic basis for dysregulation of these cytokines, and their association with outcome in ALF, remain poorly defined. MethodsTo investigate the impact of IL-6 and IL-10 in ALF, we used an experimental setting of failed liver repair after APAP overdose in which a high dose of APAP is administered (i.e., 500-600 mg/kg). Mice were treated with neutralizing antibodies to block IL-6 and IL-10. ResultsIn mice with APAP-induced ALF, high levels of IL-10 reduced monocyte recruitment and trafficking in the liver resulting in impaired clearance of dead cell debris. Kupffer cells in these mice, displayed features of myeloid-derived suppressor cells, including high level expression of IL-10 and PD-L1, which were increased in an IL-6-dependent manner. Similar to ALF patients with HE, cerebral blood flow was reduced in mice with APAP-induced ALF. Remarkably, although IL-6 is hepatoprotective in mice treated with low doses of APAP (i.e., 300 mg/kg), IL-6 neutralization in mice with APAP-induced ALF fully restored cerebral blood flow and reduced mortality. ConclusionCollectively, these studies demonstrate that exaggerated production of IL-6 in APAP-induced ALF triggers immune suppression (i.e., high levels of IL-10 and PD-L1), reduces cerebral blood flow (a feature of hepatic encephalopathy), disrupts liver repair (i.e., failed clearance of dead cells), and increases mortality.

pathology↗

bioRxiv2022

Assessing Mouse Kidney Parvovirus' Ability to Confound Research by Examining its Effects on Renally Excreted Chemotherapeutics and its Impact on Pathologic Endpoints in the Adenine Model of Chronic Kidney Disease

Mouse kidney parvovirus (MKPV) causes inclusion body nephropathy in severely immunocompromised mice and renal interstitial inflammation in immunocompetent mice. The purpose of this 2-part study was to determine the impact that MKPV may have on preclinical models as it relates to the pharmacokinetics of chemotherapeutics as well as its impact on the adenine diet model of chronic kidney disease. To assess the impact of MKPV on pharmacokinetics of 2 renally excreted chemotherapeutics commonly used in preclinical oncology studies, methotrexate and lenalidomide, blood and urine drug concentrations were measured in MKPV-infected or uninfected immunodeficient NOD.Cg-PrkdcscidIl2rgtm1Wjl/SzJ (NSG) and immunocompetent C57BL/6NCrl (B6) female mice. Differences in plasma pharmacokinetics were observed for methotrexate, but not for lenalidomide. Differences were most profound between uninfected NSG and B6 mice. The area under the curve (AUC) of methotrexate was 1.5-fold higher in uninfected NSG mice compared to infected NSG mice, 1.9-fold higher in infected B6 mice compared to uninfected B6 mice, and 4.3-fold higher in uninfected NSG mice compared to uninfected B6 mice. Renal clearance of both drugs was not impacted by MKPV infection but was generally lower in NSG mice. To assess the impact of MKPV on the adenine diet model of chronic kidney disease, MKPV-infected and uninfected B6 female mice were fed a 0.2% adenine diet and clinical and histopathologic features of disease were assessed over 8 weeks. Infection with MKPV did not have a significant impact on serum biomarkers of renal function such as BUN, creatinine, and SDMA; urine chemistry; or hemogram. However, infection did impact select histologic outcomes. MKPV-infected mice had significantly more foci of interstitial lymphoplasmacytic infiltrates than uninfected mice after 4 and 8 weeks of diet consumption, and significantly less interstitial fibrosis at week 8. Macrophage infiltrates and renal tubular injury, assessed using various immunohistochemical stains, were similar between groups. Together, these findings indicate that MKPV infection had minimal impact on the renal excretion of 2 chemotherapeutics and serum biomarkers of renal function. However, infection significantly impacted select histologic features of renal disease in the adenine diet model. While MKPV-free mice should be used in biomedical research, it is of the utmost importance in studies evaluating renal histology as an experimental outcome.

pathology↗

bioRxiv2023

Pathology of natural infection with highly pathogenic avian influenza virus (H5N1) clade 2.3.4.4b in wild terrestrial mammals in the United States in 2022

This article describes the first detections of disease due to natural infection with highly pathogenic avian influenza virus (HPAIv) H5N1 of the Eurasian lineage goose/Guangdong clade 2.3.4.4b in wild terrestrial mammals throughout the United States during 2021-2022. Affected mammalian species include 50 red foxes (Vulpes vulpes), 6 striped skunks (Mephitis mephitis), 4 raccoons (Procyon lotor), 2 bobcats (Lynx rufus), 2 Virginia opossums (Didelphis virginiana), 1 coyote (Canis latrans), 1 fisher (Pekania pennanti), and 1 gray fox (Urocyon cinereoargenteus). Infected mammals primarily exhibited neurological signs. Necrotizing meningoencephalitis, interstitial pneumonia, and myocardial necrosis were the most common lesions; however, species variations in lesion distribution were observed. Genotype analysis of sequences from 48 animals indicates that these cases represent spillover infections from wild birds.

pathology↗

bioRxiv2023

PAH DEFICIENT PATHOLOGY IN HUMANIZED c.1066-11G>A PHENYLKETONURIA MICE

We have generated using CRISPR/Cas9 technology a partially humanized mouse model of the neurometabolic disease phenylketonuria (PKU), carrying the highly prevalent PAH variant c.1066-11G>A. This variant creates an alternative 3 splice site, leading to the inclusion of 9 nucleotides coding for 3 extra amino acids between Q355 and Y356 of the protein. Homozygous Pah c.1066-11A mice, with a partially humanized intron 10 sequence with the variant, accurately recapitulate the splicing defect and present almost undetectable hepatic PAH activity. They exhibit fur hypopigmentation, lower brain and body weight and reduced survival. Blood and brain phenylalanine levels are elevated, along with decreased tyrosine, tryptophan and monoamine neurotransmitter levels. They present behavioral deficits, mainly hypoactivity and diminished social interaction, locomotor deficiencies and an abnormal hind-limb clasping reflex. Changes in the morphology of glial cells, increased GFAP and Iba1 staining signals and decreased myelinization are observed. Hepatic tissue exhibits nearly absent PAH protein, reduced levels of chaperones DNAJC12 and HSP70 and increased autophagy markers LAMP1 and LC3BII, suggesting possible coaggregation of mutant PAH with chaperones and subsequent autophagy processing. This PKU mouse model with a prevalent human variant represents a useful tool for pathophysiology research and for novel therapies development.

pathology↗

bioRxiv2024

UNRAVELING THE LUNG VASCULAR REMODELING IN PULMONARY HYPERTENSION USING A QUANTITATIVE DIGITAL PATHOLOGY SOFTWARE

Pulmonary arterial hypertension (PAH) is a rare chronic life-threatening disorder, characterized by the elevation of the mean pulmonary arterial pressure above 20 mmHg at rest. Histologically, PAH induces lung vascular remodeling, with the thickening of vessel wall. The conventional histological analysis commonly used in non-clinical models to assess lung vascular remodeling relies on manual measurements of representative lung vessels and is time-consuming. We have developed a fully automated reader-independent software (MorphoQuant-Lung) to both specifically detect vessels and measure vascular wall components from a-SMA rat lung sections. Analysis was performed on monocrotaline-and Sugen/hypoxia-induced PH rat models, treated or not with Sildenafil. The software requires 3-5 minutes to detect up to 1500 vessels per section, classify them per size, quantify intima, media and wall thicknesses, and calculate their level of occlusion. A comparison of our digital analysis results with those of the pathologists conventional visual analysis was performed for wall thickness and lumen radius showing a strong correlation between the two techniques (r: 0.80 and r: 0.88) regardless of the rat model. In addition, the occlusion estimated by automated analysis also strongly correlated with the mean pulmonary arterial pressure and the pulmonary vascular resistance (r ranging from 0.71 to 0.83) in both rat models. The added value of the present digital analysis paves the way for a more in-depth understanding of the PAH physiopathology in preclinical research and provides a robust and reliable tool for efficient therapeutic drug development.

pathology↗

bioRxiv2024

Moderate-term dimethyl fumarate treatment reduces pathology of dystrophic skeletal and cardiac muscle in a mouse model

In Duchenne muscular dystrophy (DMD), corticosteroids significantly slow disease progression and have been used as a standard of care tool for more than 30 years. However, corticosteroids also impart side effects severe enough to preclude use in some patients. There remains an unmet need for new therapeutics that target the flow-on pathogenic mechanisms of DMD with a more favourable side-effect profile. We have previously demonstrated that short-term treatment with dual-purpose anti-inflammatory, anti-oxidative dimethyl fumarate (DMF), a drug with indication and established safety data in Multiple Sclerosis, more selectively modulates Duchenne (mdx) immunology than the frequently used corticosteroid, prednisone (PRED). Here, we assess the effect of moderate-term DMF treatment over 5 weeks in the typically mild mdx mouse model that we aggravated using exercise. We show that like PRED, DMF maintains anti-inflammatory action but with additional anti-fibrotic and anti-lipogenic effects on muscle with moderate-term use. This study supports our previous work highlighting DMF as a possible repurposing candidate for DMD, especially for patients who cannot tolerate chronic corticosteroid treatment.

pathology↗

bioRxiv2025

Morpho-functional timeline of progressive cystic fibrosis pancreatic exocrine and endocrine pathology derived from semi-quantitative scoring and AI-driven quantitativeimage analysis

Cystic fibrosis (CF) is associated with pancreatic exocrine insufficiency (PEI) early in life and diabetes in up to 50% of adults. The underlying CF-related sequential changes within the pancreas associated with exocrine and endocrine insufficiency remain incompletely understood due to scarcity of available human tissue, protracted disease course and absence of established robust and reproducible analytical approaches. This study aimed to develop and apply a systematic analysis cross-sectionally to CF pancreatic tissue samples from donors over a wide age range to construct a timeline related to the main exocrine and endocrine changes underlying progressive disease. Based on a histopathological semi-quantitative scoring system and AI-driven quantitative image analysis pancreatic changes were individually evaluated and classified according to three patterns: fibrotic; fibrotic and lipotic; and lipoatrophic. This systematic evaluation was applied to 29 CF and 58 control donors without pancreatic disease. Rapid loss of acinar tissue with virtually complete absence by the age of 7 years was confirmed, mirrored by fatty tissue replacement - changes underlying PEI and likely preceding progression towards diabetes. Ductal blockage by thickened secretions was associated with increasing ductal dilatation accompanied by peri-ductal fibrosis, followed by ductal loss with involution of associated fibrosis in parallel with increasing adipocyte proportional area (PA). Remaining ducts were relatively small surrounded by residual fibrosis. Islets became increasingly clustered initially surrounded by activated pancreatic stellate cells (PSCs) and fibrosis and then disorganised by interposing fibrotic tissue between endocrine cell regions and surrounded by residual collagen stranding in a lipoatrophic pancreas. Overall islet mass was not significantly reduced but {beta}-cell PA was significantly reduced from birth without further loss over time. We concluded that the natural history of pancreatic CF progresses inexorably from peri-ductal fibrosis to global fat replacement with relatively well-maintained islet mass but PSC-associated fibrotic islet remodelling circumstantially implicated in {beta}-cell failure.

pathology↗

bioRxiv2025

LRP2 is a potential molecular target for pathological myopia

High myopia (HM) and posterior staphyloma (PS) are major causes of vision loss worldwide. Genetic and environmental factors, especially light exposure, contribute to myopia. Mutations in low-density lipoprotein-related receptor type 2, LRP2 cause syndromic myopia, and the Foxg1-Cre-Lrp2lox/lox mouse is a model for myopia and PS but the involvement of LRP2 in non-syndromic HM (NSHM) was unknown. We showed that LRP2 levels were decreased in the vitreous of 25 patients with NSHM and PS, and that in human donor eyes affected by PS, LRP2 expression was reduced in the neural retina and retinal pigment epithelium (RPE). The morphologic changes of these RPE were similar to those observed in the RPE of the Foxg1- Cre-Lrp2lox/lox mouse. In iPSc-derived human RPE cells (iRPE), LRP2 was expressed at a functional location, and LRP2 silencing by a specific siRNA regulated genes belonging to pathways involved in eye and neuronal development, visual perception, tissue remodeling, hormone metabolism and RPE structure, demonstrating that LRP2 orchestrates in RPE, functions that are essential for eye growth. Exposure of iRPE to light with LEDs of different wavelengths upregulated LRP2 expression, with higher efficacy for red light. Conversely, LRP2 expression was downregulated after cortisol exposure. Our findings link LRP2 to myopization and environmental factors and highlight its role in NSHM and PS in humans. LRP2 appears to be a viable target for interventional strategies in the treatment of NSHM. One Sentence SummaryLRP2, pivotal in the regulation of eye growth, exhibits a decrease in high myopic eyes. Its expression is augmented by light exposure in retinal pigment epithelial cells.

pathology↗

bioRxiv2025

Quantification of collagen and associated features from H&E-stained whole slide pathology images across cancer types using a physics-based deep learning model

BackgroundCollagen is the major component of the extracellular matrix (ECM). Collagen structural organization undergoes significant transformation during tumorigenesis. The visualization of collagen in histological tissue sections would aid in the study of tumor growth, encapsulation, and invasion. However, such visualization requires the use of special stains such as Picrosirius Red (PSR) or Massons Trichrome (MT), or more recently, second-harmonic generation imaging (SHG) in unstained tissue sections. However, PSR and MT both suffer from significant inter- (and intra-) lab stain variabilities, and SHG, while considered a ground truth by many, suffers from issues of system complexity/reliability, cost, and speed/throughput. These technical hurdles limit more widespread assessment of collagen in tissue samples. MethodsUsing high-contrast, high-throughput polarization imaging on PSR-stained slides to generate ground truth training polarization images, we developed a deep learning model (iQMAI) to infer the presence of collagen directly from hematoxylin and eosin (H&E)-stained whole-slide images (WSIs) with high specificity. After iQMAI inference across WSIs, individual collagen fibers were extracted, and features describing overall collagen intensity and fiber morphology were computed. iQMAI pixel-and feature-wise outputs were compared to ground truth polarization imaging to assess model performance. The trained iQMAI model was deployed on H&E-stained WSI from the TCGA LUAD, LUSC, LIHC, and PAAD datasets for evaluation. iQMAI-derived collagen features were compared to tissue composition, gene expression, and overall survival. ResultsThe iQMAI model shows significant generalization across multiple indications. iQMAI collagen predictions were similar to polarization imaging measurements of the same sample, with a mean structural similarity index (SSIM) of 0.84 (95% CI 0.69-0.93), a mean patch-wise RMSE of 0.04 (95% CI 0.02-0.08), and a linear correlation (R2=0.93). Comparing features of the collagen fibers extracted from iQMAI vs. polarization images yielded similar linear correlations between computed fiber tortuosity, length, width, and relative angle. The relationship between collagen fiber density and fibroblast density was distinct in non-small cell lung cancer (LUAD and LUSC), hepatocellular carcinoma (LIHC), and pancreatic ductal adenocarcinoma (PAAD). In PAAD, fiber density and fiber width were both negatively associated with the LRRC-15 gene expression signature, and increased fiber width was associated with longer overall survival. ConclusionsiQMAI is a deep learning model that accurately predicts collagen from an H&E-stained WSI, allowing for spatially resolved quantification of collagen morphology and enabling investigation of the interplay between collagen and other TME components. We demonstrate an example of the utility of iQMAI-based collagen assessment in PAAD, where collagen features are associated with immunosuppressive cancer-associated fibroblasts and overall survival. Understanding the relationship between collagen, the tumor microenvironment composition, and disease progression may aid the development of effective immunotherapies in PAAD and other cancer types.

pathology↗

bioRxiv2025

AI-driven computational pathology for recurrence risk assessment in early-stage breast cancer

Recurrence related to poor prognosis is a leading cause of mortality in patients with breast cancer (BC). The MammaPrint (MP) genomic assay is designed to stratify recurrence risk and evaluate chemotherapy benefits for early-stage HR+/HER2- BC patients. However, MP fails to reveal spatial tumor morphology and is limited by high costs. In this study, we establish a BC MP cohort and develop CPMP, a weakly supervised agent-attention transformer model, to predict MP recurrence risk from annotation-free BC histopathological slides. CPMP achieves an AUROC of 0.824 {+/-} 0.03 in predicting MP risk groups. We further leveraged CPMP for spatial and morphological analyses to explore histological patterns associated with MP risk groups. The model reveals tumor spatial localization at the whole-slide level and highlights distinct intercellular interaction patterns of MP groups. It also characterizes the diversity in tumor morphology and uncovers MP High-specific, Low-specific, and colocalized morphological phenotypes that differ in quantitative cellular composition. Prognostic evaluation in the external cohort exhibits significant stratification of distant metastasis risk (HR: 3.14, p-value = 0.0014), underscoring the prognostic power of CPMP. These findings demonstrate the capability of CPMP in MP risk prediction, offering a flexible supplement to genomic risk assessment in early-stage BC.

pathology↗

bioRxiv2025

IRP1 deficiency alters mitochondrial metabolism and protects against metabolic syndrome pathologies

Iron regulatory protein 1 (IRP1) is a post-transcriptional regulator of cellular iron metabolism. In mice, loss of IRP1 causes polycythemia through translational de-repression of hypoxia-inducible factor 2 (HIF2) mRNA, which increases renal erythropoietin production. Here we show that Irp1-/- mice develop fasting hypoglycemia and are protected against high-fat diet-induced hyperglycemia and hepatic steatosis. Discovery-based proteomics of Irp1-/- livers revealed a mitochondrial dysfunction signature. Seahorse flux analysis in primary hepatocytes and differentiated skeletal muscle myotubes confirmed impaired respiratory capacity, with a shift from oxidative phosphorylation to glycolytic ATP production. This metabolic rewiring was associated with enhanced insulin sensitivity and increased glucose uptake in skeletal muscle. Under metabolic stress, IRP1 deficiency altered the redox balance of mitochondrial iron, resulting in inefficient energy production and accumulation of amino acids and metabolites in skeletal muscle, rendering them unavailable for hepatic gluconeogenesis. These findings identify IRP1 as a critical regulator of systemic energy homeostasis.

pathology↗