Abstract: Background/Objectives: Variations in staining styles—arising from differences in tissue preparation, scanners, and laboratory protocols—severely compromise the robustness of automated cell segmentation algorithms in digital pathology. Moreover, manual nucleus annotation is extremely labor-intensive, leading to a scarcity of large-scale, fully anno-tated datasets for supervised nucleus segmentation. This study proposes a novel framework that simultaneously mitigates staining variability and achieves high-accuracy nucleus segmentation using only minimal annotations. Methods: We present 3SGAN, a multitask dual-branch generative adversarial network (GAN) that jointly performs stain normalization and nucleus segmentation in a semi-supervised manner. The framework adopts a teacher-student paradigm: a lightweight teacher model (AttCycle) equipped with attention gates generates reliable pseudo-labels, while a high-capacity student model (TransCycle) leveraging a hybrid CNN-Transformer architecture further refines performance. 3SGAN was trained and evaluated on a large dataset of 1,408 Whole-Slide Images (WSIs) from two medical institutions, encompassing 101 distinct staining styles, with nucleus-level annotations required for only 5% of the data. Results: 3SGAN sig-nificantly outperformed state-of-the-art methods, achieving substantial improvements in stain normalization quality (RMSE, PSNR, SSIM) and marked gains in nucleus seg-mentation performance (F1 score, mean IoU, and AJI). External validation on independent MoNuSeg and PanNuke datasets, as well as on previously untested tumor-rich non-ROI regions from our in-house WSIs, confirmed strong generalizability with excellent stain normalization and top-tier segmentation accuracy across diverse staining protocols, tissue types, and pathological patterns. Conclusions: The proposed 3SGAN framework demonstrates that high-performance nucleus segmentation and stain normalization can be achieved with minimal annotation requirements, offering a practical and scalable solution for digital pathology applications across diverse clinical settings and staining protocols.

Abstract: This review article attempts to provide a unifying hypothesis to explain the myriad of symptoms and predispositions underlying the development of PASC (Postacute Sequelae of COVID), often referred to as Long COVID. The hypothesis described here proposes that Long COVID is best understood as a disorder of persistent immune dysregulation, with chronic macrophage activation representing the fundamental underlying pathophysiology. Unlike transient post-viral syndromes, long COVID involves a sustained innate immune response, particularly within monocyte-derived macrophages, driven by persistent spike protein (peripherally in MAIT cells and centrally in Microglial cells), epigenetic imprinting, and gut-related viral reservoirs. These macrophages are not merely activated temporarily, but become epigenetically “trained” into a prolonged inflammatory state, as demonstrated by enduring histone acetylation markers such as H3K27acDNA Reprogramming. It is proposed that recognizing macrophage activation as the central axis of Long COVID pathology offers a framework for personalized risk assessment, targeted intervention, and therapeutic recalibration.

Abstract: Human papillomavirus (HPV) plays a crucial role in the pathogenesis of oropharyngeal squamous cell carcinomas (OPSCC). Accurate HPV status classification is essential for therapeutic stratification. While p16 immunohistochemistry (IHC) is the clinical surrogate marker, it has limited specificity. In this study, we implemented a weakly supervised deep learning approach using the Clustering-constrained Attention Multiple Instance Learning (CLAM) framework to directly predict HPV status from hematoxylin and eosin (H&E)-stained whole-slide images (WSIs) of OPSCC. A total of 123 WSIs from two cohorts (TCGA and OPSCC-UNINA) were used. Attention heatmaps revealed that the model predominantly focused on tumor-rich regions. Errors were primarily observed in slides with conflicting p16/ISH status or suboptimal quality. Morphological analysis of high-attention patches confirmed that cellular features extracted from correctly classified slides align with HPV status, with a Random Forest classifier achieving 83% accuracy at the cell level. This work supports the feasibility of deep learning-based HPV prediction from routine H&E slides, with potential clinical implications for streamlined, cost-effective diagnostics.

Abstract: Background: Histologic healing is increasingly recognized as a sensitive marker of disease remission in ulcerative colitis (UC). However, the dynamics of mucosal T lymphocytes and pro-inflammatory cytokines during healing remain incompletely understood. Methods: In this observational, cross-sectional study, paired colonic biopsies from 20 adult UC patients were analyzed during active inflammation and subsequent histologic healing. Immunohistochemistry was performed for CD3, CD4, CD8, and IL-6. Lymphocyte densities were quantified in intraepithelial and lamina propria compartments, while IL-6 expression was scored semi-quantitatively. Histologic activity was assessed using the Geboes score. Results: Intraepithelial CD4⁺ T cells significantly decreased during histologic healing (mean 6.8 → 3.75 cells/HPF, p < 0.05), whereas lamina propria CD4⁺ cells remained variably persistent, suggesting ongoing immune regulation. Intraepithelial CD8⁺ cells increased during remission, indicating a potential reparative or surveillance role. IL-6 expression markedly declined in epithelial and stromal compartments during healing, reflecting resolution of mucosal inflammation. Correlation analyses revealed enhanced coordination between CD4⁺ and CD8⁺ cells in the healing phase, consistent with immune homeostasis. Conclusions: Histologic healing in UC involves compartment-specific shifts in T lymphocyte populations and a marked reduction of IL-6 expression, reflecting coordinated immune regulation beyond clinical remission. These findings highlight the potential of combined cellular and cytokine biomarkers to monitor mucosal healing and guide immunomodulatory therapies.

Abstract: To contrast the COVID-19 pandemic brought by the corona virus SARS-CoV-2, two mRNA-based anti-COVID-19 vaccines (by Pfizer-BioNTech and Moderna) were rela-tively soon made available and deployed worldwide based on an emergency approval. Being considered vulnerable and at risk of infection, cancer patients have been priori-tized for COVID-19 vaccination and vaccinated repeatedly because of the short-time protection provided by these vaccines. During the pandemic, because of the large number of infected patients requiring assistance, many hospitals opted for giving pri-ority to these patients while postponing the specialist treatment for other pathologies, including cancer. Recently, a surge in the incidence and rapid progression of cancers has been observed in many countries, which could (at least partially) represent those cancers undiagnosed or untreated during the pandemic. It has also been suggested that the SARS-CoV-2 itself or even the anti-COVID-19 mRNA vaccines could have contrib-uted to the recurrence and worse clinical outcome in cancer patients, given the high incidence of COVID-19 in hospitalized patients and that these patients have been vac-cinated with priority several times and in a short period. Although it appears ex-tremely unlikely that SARS-CoV-2 and anti-COVID-19 mRNA vaccines elicit genotoxic events and cause neo-cancerogenesis in a short time, they could still cause non-genotoxic pro-carcinogenic effects by triggering an exaggerated inflammatory re-action, compromising immune homeostasis, stimulating cell proliferation, and nega-tively affecting cellular stress response and damage repair machinery. This could re-sult in the promotion of regrowth of dormant micrometastases or relapse of stable minimal residual disease. Such a harmful outcome may likely result from a synergy between the virus and the vaccine, especially in multi-vaccinated and multi-infected individuals. Here, I bring the cell pathologist's point of view and discuss the multiple possible mechanisms by which the virus and the anti-COVID-19 mRNA vaccine might favor tumorigenesis. While a causal link cannot be established at this stage, knowledge of potential carcinogenic risks could help doctors and health policymakers take the best actions to protect vulnerable patients and convince the vaccine developer to de-sign a vaccine free from such harm.

Abstract:

Background and Objectives: Differentiating between subtypes of renal cell carcinoma (RCC) can be challenging due to overlapping immunohistochemical and histomorphological features. Furthermore, despite the evaluation of numerous histopathologic parameters, predicting metastasis and prognosis remains unclear. In this study, we aimed to evaluate the potential of microRNA (miRNA) expression levels in establishing definitive diagnoses, assessing metastatic potential, and determining patient prognosis. Materials and Methods: A total of 35 clear cell RCC (cc-RCC), 11 chromophobe RCC (ch-RCC), 9 papillary RCC (p-RCC) cases, and 9 cc-RCC metastases were retrospectively analyzed. MiRNA-21 and miRNA-210 expression profiles were assessed using non-parametric tests, receiver operating characteristic (ROC) analysis, and Cox regression models. Results: The expression levels of both miRNA-21 and miRNA-210 were significantly elevated in cc-RCC and p-RCC cases. In contrast, miRNA-210 expression levels were significantly reduced in all ch-RCC cases, whereas miRNA-21 expression showed no significant change. Although miRNA-21 expression levels were higher in metastases compared to primary cc-RCC tumors, the difference was not statistically significant (p=0.053). No significant difference was observed in miRNA-210 expression between metastatic and primary cc-RCC tumors (p=0.237). Regardless of histological subtype, an increase in miRNA-210 expression was associated with a significant reduction in overall survival (p=0.03, HR=1.729, 95% CI: 1.043-2.864). Conclusion: The findings suggest that miRNA-21 and miRNA-210 expression levels can help distinguish ch-RCC from cc-RCC and p-RCC. Additionally, miRNA-210, but not miRNA-21, may differentiate cc-RCC from p-RCC. Higher miRNA-210 expression levels were associated with worse overall survival in RCC patients. While miRNA-21 levels were increased in cc-RCC metastases, further studies with larger sample sizes are needed to establish its role in predicting metastasis.

Abstract: Parkinson’s disease (PD) is a multisystem disorder, with early changes extending beyond basal ganglia circuitries and involving non-dopaminergic pathways, including cerebellar networks. Whether cerebellar dysfunction reflects a compensatory mechanism or an intrinsic hallmark of disease progression remains unresolved. In this cross-sectional study, we examined how cerebellar γ-aminobutyric acid (GABA) and glutamate/glutamine (Glx) systems, as well as their excitatory/inhibitory (E/I) balance, are modulated along the disease course. As to ascertain how these mechanisms contribute to motor and non-motor features in the premotor and early stages of PD, 18 individuals with isolated REM sleep behaviour disorder (iRBD) 20 de novo, drug-naïve PD (dnPD), and 18 matched healthy controls underwent clinical, cognitive, and neuropsychiatric assessments alongside cerebellar magnetic resonance spectroscopy (MRS, MEGA-PRESS, 3T). While cerebellar neurotransmitter levels did not differ significantly across groups, dnPD patients exhibited a shift toward hyperexcitability in the E/I ratio, without correlation to clinical or cognitive measures. In contrast, in iRBD an inverse relationship between heightened GABAergic activity and neuropsychiatric symptoms emerged. These findings suggest an early, dynamic cerebellar involvement, potentially reflecting compensatory modulation of altered basal ganglia output. Our results support cerebellar GABA MRS as a promising biomarker and open perspectives for targeting non-dopaminergic pathways in PD.

Abstract:

Background: Several markers have been shown to define subpopulations enriched for cancer stem cells (CSCs) and correlate with poor clinical outcome in oral squamous cell carcinoma (OSCC). Objective: To investigate the pattern of expression and correlation with clinical parameters of two CSC markers, namely p75NTR and ALDH1A1, in both patient samples and cell lines. Methods: Archival formalin-fixed paraffin embedded samples from normal human oral mucosa (NHOM, n=31), oral dysplasia (OD, n=10) or OSCC (n=177) were subjected to multiple immunohistochemistry and some to qRT-PCR for expression of CSC and proliferation-related markers, BMI1 and Ki67. Correlations between CSC marker expression and clinical parameters were investigated. Primary cells and cell lines derived from NHOM, OD or OSCC were FACS- analyzed for the same markers. Results: A higher frequency of cells positive for CSC markers was detected in OD and OSCC compared to NHOM. Co-localization of the two markers was a rare finding in OSCC as compared to NHOM or OD and was more heterogeneous in OSCC cell lines than in OD and NHOM cells. Cells positive for p75NTR exhibited higher expression of proliferative and self-renewal markers in comparison to ALDH1A1+ or double ALDH1A1+/p75NTR+ cells. Cells positive for p75NTR were more frequent in small size tumors, poorly to moderately differentiated tumors, and correlated with poor survival of patients otherwise (clinically) deemed as of better prognosis. Higher frequency of ALDH1A1+ cells was found to be associated with lymph node metastasis. Both p75NTR+ cells and ALDH1A1+ cells could emerge de novo from the respective negative sub-population after FACS sorting and in vitro growth, but with different kinetics. Conclusion: Here we show that several stem cell sub-populations with distinct phenotypes co-exist in a tumor, each having impact on different clinical parameters. The cell subpopulations identified by use of different CSC markers were found to be dynamic populations, able to switch between phenotypes. In addition, our data suggest also that the stem cell heterogeneity is acquired and evolve parallel with carcinoma progression.

Abstract: Pancreatic ductal adenocarcinoma (PDAC) is an aggressive cancer that frequently presents at an advanced stage with limited effective treatment options and a dismal prognosis. It is a highly heterogenous disease driven by various genetic and epigenetic alterations. Recent advances in sequencing modalities have significantly improved our understanding of the genetics of PDAC, which could lead to promising and novel therapeutic strategies. In this review, we summarize the most up-to-date literature on the molecular landscape of PDAC and its precursor lesions, and recent advances in targeted therapy.

Abstract: Background: Vaes et al. (2023)[@vaes2023] identified 13 symptom clusters in a large cohort of ME/CFS patients. Symptom intensity is broadly correlated with post-exertional malaise (PEM) severity, with variation across clusters that seems disorganized. Despite this research, no broadly accepted organizing principle has emerged from this paper or other attempts at phenotyping ME/CFS. Objective: To identify and characterize potential subgroups defined by symptom domain severity relative to PEM within the original Vaes symptom clusters. Our analysis identified two groups of patient clusters with distinct symptom-domain profiles. Anchoring symptom domains to PEM collapses the 13 Vaes clusters into two reproducible families: one characterized by parallel offsets (autonomic, neuroendocrine, other) and one by amplified slopes (pain, neurocognitive), with high-end convergence after accounting for a single influential cluster. This subgroup's symptom trends could align with those of fibromyalgia. Further exploration with individual patient data is needed to validate these findings.

Abstract: Accessory breasts denote the formation of extra breast tissue along the milk line, being more prevalent among Black and Asian populations, affecting both genders. This study aimed to determine the genetic aetiology of accessory breasts in a multiplex family, where all female siblings presented with bilateral accessory breasts. The study also ascertained secondary findings (SFs) that may impact the health of the family members. Clinical data and saliva samples were obtained from all family members. Ultrasound and histopathology confirmed the diagnosis. Whole-exome sequencing was conducted on DNA samples obtained from the saliva, with variant calling conducted utilizing Sentieon workflow. Variant classification was based on American College of Medical Genetics and Genomics guidelines. After segregation analysis, 12 candidate genes emerged. Among these, PRSS50 and FANCC emerged as top candidates, being implicated in breast diseases. However, two variants in FANCC (c.360del; p.His120GlnfsTer24 and c.355_358del; p.Ser119IlefsTer24) were selected as the most probable causal variants because of the role of this gene in familial hereditary breast and ovarian cancer syndromes. The remaining ten genes were reported as potentially accounting for co-occurring conditions segregating with accessory breasts. Reported SFs involve TTR and RYR1. In conclusion, pathogenic variants in FANCC are causal for familial accessory breasts.

Abstract: Severely ill ME/CFS patients are bedridden and suffer from hypersensitivities against light and noise, severe orthostatic intolerance reducing cerebral blood flow, and skeletal muscle symptoms including loss of force, fatigue, pain, fasciculations and cramps. Since neurological investigations exclude neuronal causes for myasthenia, we assume a muscular pathomechanism. In previous papers we considered insufficient activity of the Na+/K+-ATPase as the main cause of mitochondrial damage via high intracellular sodium which reverses the transport mode of the sodium-calcium-exchanger to import calcium causing calcium-overload. Low Na+/K+-ATPase-activity also causes sarcolemmal depolarization leading to less effective action potential propagation and loss of force. Depolarization brings membrane potential closer to the firing threshold causing hyperexcitability explaining fasciculations and cramps. These raise sodium influx during excitation to further increase the workload of Na+/K+-ATPase. Thereby, depolarization causes further depolarization. Higher intracellular sodium favors calcium-overload and mitochondrial damage to lower energy supply of Na+/K+-ATPase and to increase reactive-oxygen species that further inhibit it. Even at rest, muscle is in a state of depolarization. Depolarization and mitochondrial damage reinforce each other. Thus, dysfunction of Na+/K+-ATPase as a single mechanism can explain the different skeletal muscle symptoms of the severely ill ME/CFS patient comprising loss of force, fatigue and fasciculations.

Abstract: High-grade serous carcinoma (HGSC) of the ovary is characterized by two major histological patterns: a classic papillary/micropapillary architecture and a solid pseudo-endometrioid transitional (SET) variant. We investigated whether the distinct morphological subtypes are underpinned by transcriptomic differences in the tumor microenvironment (TME). We profiled 21 HGSC tumors (7 SET, 14 classic) using a 770-gene NanoString PanCancer Progression panel. Differential expression analysis revealed ~20 genes with significantly different expression (>4-fold, adjusted p < 0.01) between SET and classic tu-mors. Unsupervised clustering partially separated SET and classic tumors, suggesting that global gene expression patterns correlate with histologic subtype. SET tumors exhibited upregulation of cell-cycle and epithelial genes (e.g., PTTG1, TRAIL, HER3) and down-regulation of genes involved in epithelial–mesenchymal transition (EMT), extracellular matrix (ECM) organization, and angiogenesis (e.g., TWIST2, FGF2, decorin) relative to classic tumors. Notably, PTTG1 and TRAIL were upregulated ~6–9-fold in SET tumors, whereas TWIST2 was ~7-fold downregulated, consistent with reduced EMT in SET tumors. Pathway analysis indicated that SET tumors appear to have an immune-active, stroma-poor microenvironment, potentially consistent with an "immunoreactive" phenotype, whereas classic tumors showed a mesenchymal, stroma-rich profile. These molecular distinctions could have diagnostic utility and may inform therapeutic stratification, with key dysregulated genes (e.g., HER3, TRAIL, FGF2) representing potential prognostic or predictive biomarkers. For example, high HER3 expression in SET tumors might predict sensitivity to ERBB3/PI3K inhibitors, whereas stromal factors (e.g., FGF2) enriched in classic HGSC could be targeted with microenvironment-modulating therapies. These preliminary findings require validation before translation into pathology practice via immunohistochemical (IHC) assays (e.g., for HER3 or TRAIL), potentially enabling improved classification and personalized treatment of HGSC.

Abstract: Coeliac disease (CeD) is a gastrointestinal enteropathy triggered by the consumption of gluten in predisposed individuals. Only around 50% of CeD genetic risk is understood, with the majority of risk attributed to the HLA loci. We investigated the butyrophilin family of immunomodulators as novel CeD risk loci. We sequenced the butyrophilin loci of 48 CeD and 46 control patients and carried out gene-based burden testing on the captured single nucleotide polymorphisms (SNPs). We found significantly increased BTN2A1 gene burden in CeD patients. To validate these results, the SNP data of 3094 CeD patients and 29 762 control participants from the UK Biobank database were subjected to single variant analyses. Fourteen BTN2A1, 10 BTN3A1, and 13 BTN3A2 SNPs were significantly associated with CeD status. Twenty of the 37 SNPs above were associated with CeD status independent of the risk associated HLA genotypes. All twenty of these SNPs, alongside a novel SNP not included in the above SNPs were associated with CeD in HLA-DQ2.5-matched case-control groups. This study reaffirmed the association of the BTN3A2 locus with CeD risk, and identified BTN2A1 and BTN3A1 as putative novel CeD risk loci.

Abstract:

Endometrial carcinoma is the most frequent gynecologic malignancy in western countries. In recent years, mutations in CTNNB1 have been associated with worse prognosis in low-risk carcinomas. However, there is a lack of understanding of the proteomic implications of CTNNB1 mutations in these tumors. In this study, we performed shotgun proteomics using Formalin Fixed Paraffin Embedded (FFPE) tissue samples of CTNNB1 mutated and wild-type low-risk endometrial carcinomas. A publicly available proteomic and transcriptomic database was used to validate results. Differential protein expression and Gene Set Enrichment Analysis revealed dysregulation of pathways associated with cell keratinization, immune response modulation, and intracellular calcium regulation. CTNNB1 mutated tumors showed immune dysregulation at multiple levels including cytokine secretion, cell adhesion, and lymphocyte activation. These results were supported by tissue multiplex immunofluorescence analysis, demonstrating reduced CD8 tumor infiltrating lymphocytes and different immune spatial interaction patterns. Intracellular calcium dysfunction was associated with key transcript dysregulation. We found an increased expression of CAMK2A and ROR2, suggesting a potential role for non-canonical Wnt pathway activation in CTNNB1 mutated tumors.

Abstract: Cutaneous leiomyosarcoma is a rare, malignant tumor that arises from smooth muscle cells, accounting for less than 3% of all cutaneous sarcomas. Our case report details a 63-year-old male patient who presented with a rapidly growing, painful nodule in the popliteal region. The patient underwent initial surgical excision in September 2021, followed by three subsequent resections until March 2022 due to local recurrence. Histopathological analysis of all specimens revealed a dermal neoplasm composed of spindle cells arranged in intersecting fascicles with storiform patterns. The immunohistochemistry profile showed strong positivity for markers SMA and desmin, confirming the diagnosis. Despite early interventions, the deep surgical margins were positive, and further surgeries were required until tumor-free margins were achieved. This case emphasizes the morphological characteristics, clinical behavior, and therapeutic challenges in managing cutaneous leiomyosarcoma. A favorable prognosis is achieved with long-term follow-up and a multidisciplinary approach.

Abstract: Digital pathology (DP) has become an integral part of modern diagnostic practice, enabling efficient workflows, reproducible assessments, and the integration of computational tools. While many institutions are transitioning toward digital diagnostics, deployment, validation, and routine usage of DP tools in clinical settings remains a major challenge. Existing recommendations often remain high-level, leaving laboratories without detailed, practical guidance on how to structure and document validation processes in real-world settings. To address this gap, we present a framework derived from the experience of Geneva University Hospitals (HUG), which since 2019 has operated a fully digital primary diagnostic workflow processing more than 230,000 slides annually. Drawing on this implementation, we outline best practices and lessons learned for the development, deployment, validation, and accreditation of DP tools. Document templates are also provided, which may serve as adaptable blueprints for other laboratories. This practical guide is intended to support reproducible, transparent, and audit-ready validation processes, helping institutions bridge the gap between general recommendations and operational deployment of DP tools in accredited diagnostic environments.

Abstract: Background: Myalgic Encephalomyelitis/Chronic Fatigue Syndrome (ME/CFS) is marked by heterogeneous symptom clusters and only partial or short-lived responses to interventions. While gut involvement has been implicated in ME/CFS, it has not been proven to be the root cause, nor have symptoms been consistently responsive to treatment.Methods: We analyze prototype illnesses with known gut barrier dysfunctions from the duodenum to the colon to identify similarities and symptoms that are direct results of barrier dysfunction. In particular, we assess what can be learned from these prototype diseases that depends not on pathogenesis, but on barrier breakdown and metabolite leakage.Results: Prototype diseases showed consistent involvement of pore pathways, frequent histamine leakage, and parallel treatment responses, despite differing in pathogenesis.

Abstract:

Background: Celiac disease (CD) is a gluten-sensitive immune-related enteropathy of the small intestine characterized by villus atrophy, crypt hyperplasia, and increased intraepithelial lymphocytes (IELs). Objectives: To characterize the phenotype of IELs and immune cells of the lamina propria of small intestine control using immuno-oncology and immune-phenotype markers and test the most relevant marker, an immune checkpoint co-inhibitory receptor, leukocyte associated immunoglobulin like receptor 1 (LAIR1) in CD. Methods: Immunohistochemical analysis of CD3, CD4, CD8, CD103 (ITGAE), Granzyme B, TCR beta (β), TCR delta (δ), CD56 (NCAM), CD16, LAIR1 (CD305), PD-L1, PD1 (CD279), BTLA (CD272), TOX2, HVEM (TNFRSF14), CD163, HLA-DP-DQ, IL4I1, and FOXP3 was performed using histological analysis. Gene expression analysis was performed using an independent dataset to expand and confirm the findings. Results: IELs exhibited a cytotoxic T-cell phenotype and were positive for CD3, CD8, CD103, TCRβ, and LAIR1. The lamina propria was abundant in CD163, HLA-DP-DQ, BTLA, PD-L1, CD103, CD56, and LAIR1-positive cells corresponding to macrophages and T- and B-lymphocytes. In CD, IELs and part of the inflammatory cells of the lamina propria cells were LAIR1-positive. CD was characterized by higher LAIR1-positive cell expression than the small intestine control (P = 0.004). Higher intestinal lesions evaluated by Marsh scoring were correlated with higher LAIR1 (P < 0.001). Gene expression analysis confirmed the overexpression of the LAIR1 pathway in CD and highlighted BTLA. At the protein level, BTLA overexpression was confirmed in CD. Finally, as a proof-of-concept AI analysis, a convolutional neural network classified LAIR1-stained image-patches between the 3 diagnoses of small intestine control, CD, and reactive tonsils with high accuracy (99.6%). Conclusions: IELs exhibit cytotoxic T-cell phenotype and are CD3, CD8, CD103, TCRβ, and LAIR1 positive in small intestine control. Increased numbers of LAIR1-positive IELs and lamina propria immune cells characterize CD.

Abstract: Backgrounds. Programmed death-ligand 1 (PD-L1) immunohistochemistry (IHC) is a critical predictive biomarker for immune checkpoint inhibitor (ICI) therapy in triple-negative breast cancer (TNBC). However, prolonged storage of formalin-fixed paraffin-embedded (FFPE) tissue may reduce antigenicity, potentially leading to false-negative results. We investigated the effect of FFPE storage duration on PD-L1 immunoreactivity. Methods. We retrospectively analyzed 63 TNBC cases with PD-L1 testing using the 22C3 pharmDx assay at diagnosis and repeated IHC on the same FFPE blocks after varying storage durations (&lt;1, 1–2, 2–3, ≥3 years). PD-L1 positivity was defined as Combined Positive Score (CPS) ≥10. Associations with clinicopathologic features, pathologic complete response (pCR) after neoadjuvant chemotherapy (NAC), and survival were evaluated. Results. At diagnosis, 41 patients (65.1%) were PD-L1–positive. In the PD-L1–positive group, decreased staining was observed in 0%, 11%, 13%, and 50% of cases for &lt;1, 1–2, 2–3, and ≥3 years storage, respectively (p=0.015). No increased staining occurred. PD-L1 positivity correlated with higher Ki67 and nuclear grade but not with p53 status. pCR was achieved in 33% of PD-L1–positive vs 0% of PD-L1–negative NAC patients (p=0.0527). Survival analysis showed a non-significant trend toward shorter recurrence-free and overall survival in PD-L1–positive patients. Conclusions. Prolonged FFPE storage, particularly beyond three years, significantly reduces PD-L1 immunoreactivity. Testing on recent specimens is recommended to avoid false-negative results that may impact ICI eligibility.