Hasil untuk "Pathology"

M. Dunnill

R. A. Willis

H. Schuknecht

W. Burgdorf

M. Goedert

S. Moncada, E. Higgs

Gexin Huang, Anqi Li, Yusheng Tan et al.

Pathology foundation models (FMs) have become central to computational histopathology, offering strong transfer performance across a wide range of diagnostic and prognostic tasks. The rapid proliferation of pathology foundation models creates a model-selection bottleneck: no single model is uniformly best, yet exhaustively adapting and validating many candidates for each downstream endpoint is prohibitively expensive. We address this challenge with a lightweight and novel model fusion strategy, LogitProd, which treats independently trained FM-based predictors as fixed experts and learns sample-adaptive fusion weights over their slide-level outputs. The fusion operates purely on logits, requiring no encoder retraining and no feature-space alignment across heterogeneous backbones. We further provide a theoretical analysis showing that the optimal weighted product fusion is guaranteed to perform at least as well as the best individual expert under the training objective. We systematically evaluate LogitProd on \textbf{22} benchmarks spanning WSI-level classification, tile-level classification, gene mutation prediction, and discrete-time survival modeling. LogitProd ranks first on 20/22 tasks and improves the average performance across all tasks by ~3% over the strongest single expert. LogitProd enables practitioners to upgrade heterogeneous FM-based pipelines in a plug-and-play manner, achieving multi-expert gains with $\sim$12$\times$ lower training cost than feature-fusion alternatives.

Yousef Kotp, Vincent Quoc-Huy Trinh, Christopher Pal et al.

Computational pathology needs whole-slide image (WSI) foundation models that transfer across diverse clinical tasks, yet current approaches remain largely slide-centric, often depend on private data and expensive paired-report supervision, and do not explicitly model relationships among multiple slides from the same patient. We present MOOZY, a patient-first pathology foundation model in which the patient case, not the individual slide, is the core unit of representation. MOOZY explicitly models dependencies across all slides from the same patient via a case transformer during pretraining, combining multi-stage open self-supervision with scaled low-cost task supervision. In Stage 1, we pretrain a vision-only slide encoder on 77,134 public slide feature grids using masked self-distillation. In Stage 2, we align these representations with clinical semantics using a case transformer and multi-task supervision over 333 tasks from 56 public datasets, including 205 classification and 128 survival tasks across four endpoints. Across eight held-out tasks with five-fold frozen-feature probe evaluation, MOOZY achieves best or tied-best performance on most metrics and improves macro averages over TITAN by +7.37%, +5.50%, and +7.83% and over PRISM by +8.83%, +10.70%, and +9.78% for weighted F1, weighted ROC-AUC, and balanced accuracy, respectively. MOOZY is also parameter efficient with 85.77M parameters, 14x smaller than GigaPath. These results demonstrate that open, reproducible patient-level pretraining yields transferable embeddings, providing a practical path toward scalable patient-first histopathology foundation models.

Utku Böcüoğlu, Esra Ateş Yıldırım, Selma Erdoğan Düzcü et al.

Abstract Objectives This in vitro study aimed to evaluate the effect of laser photobiomodulation on the structural integrity and degradation resistance of two types of platelet-rich fibrin membranes: Leukocyte- and Platelet-Rich Fibrin (L-PRF) and Titanium-Prepared Platelet-Rich Fibrin (T-PRF). Structural changes in the fibrin network were assessed using Scanning Electron Microscopy (SEM) and light microscopy. Materials and methods Our study was performed on 15 systemically healthy individuals and four L-PRF and four T-PRF membranes obtained from each individual, totaling 120 samples. L-PRF was prepared first using standard vacuum glass tubes. Two weeks later, new blood samples were collected from the same individuals, and T-PRF membranes were prepared using sterile titanium tubes to enhance biocompatibility. Both membrane types were obtained by centrifugation at 2700 revolutions per minute (rpm) for 12 min. Two of the four membranes were treated with a diode laser device at a wavelength of 980 nanometers (nm) and a power of 0.5 W (W) in continuous mode for 3 min at a distance of 1–2 milimeters (mm). The other two membranes were not lasered. One of the laser treated L-PRF and T-PRF membranes was cut in half and stored under appropriate conditions for histological examination and SEM analysis. The other membrane was separated for degradation. The same procedures were performed for L-PRF and T-PRF membranes without laser treatment. Result Laser-treated L-PRF and T-PRF membranes showed lower degradation percentages compared to non-laser-treated membranes, but this difference did not reach statistical significance (p > 0.05). However, when laser treated L-PRF and T-PRF membranes were compared, the degradation percentage was significantly higher in L-PRF membrane (p < 0.05). Histologic examination showed that the fibrin network structure of the laser-applied L-PRF and T-PRF membrane groups was significantly denser than the non-laser-applied groups (p < 0.05). SEM analysis revealed that the fibrin network was denser, thicker and more complex in the laser-applied L-PRF and T-PRF membrane groups. Conclusion In this study, the biostimulative effect of laser increased the fibrin network thickness, cross-link structure and density of L-PRF and T-PRF membranes. When the degradation percentages on the membranes were evaluated, no significant difference was observed between the groups. Clinical relevance Understanding how laser photobiomodulation affects the structure and degradation resistance of both L-PRF and T-PRF membranes can guide clinicians in selecting the most suitable autologous biomaterial for enhancing wound healing and regenerative outcomes in dental procedures. The structure of PRF membranes used in dentistry can be improved using the biostimulative effect of the laser. This application may increase the use of these autologous and easily obtainable materials in treatments.

Xiao-Jiang Li, Shihua Li, A. Sharp et al.

Advait Gosai, Arun Kavishwar, Stephanie L. McNamara et al.

Recent work has shown promising performance of frontier large language models (LLMs) and their multimodal counterparts in medical quizzes and diagnostic tasks, highlighting their potential for broad clinical utility given their accessible, general-purpose nature. However, beyond diagnosis, a fundamental aspect of medical image interpretation is the ability to localize pathological findings. Evaluating localization not only has clinical and educational relevance but also provides insight into a model's spatial understanding of anatomy and disease. Here, we systematically assess two general-purpose MLLMs (GPT-4 and GPT-5) and a domain-specific model (MedGemma) in their ability to localize pathologies on chest radiographs, using a prompting pipeline that overlays a spatial grid and elicits coordinate-based predictions. Averaged across nine pathologies in the CheXlocalize dataset, GPT-5 exhibited a localization accuracy of 49.7%, followed by GPT-4 (39.1%) and MedGemma (17.7%), all lower than a task-specific CNN baseline (59.9%) and a radiologist benchmark (80.1%). Despite modest performance, error analysis revealed that GPT-5's predictions were largely in anatomically plausible regions, just not always precisely localized. GPT-4 performed well on pathologies with fixed anatomical locations, but struggled with spatially variable findings and exhibited anatomically implausible predictions more frequently. MedGemma demonstrated the lowest performance on all pathologies, but showed improvements when provided examples through few shot prompting. Our findings highlight both the promise and limitations of current MLLMs in medical imaging and underscore the importance of integrating them with task-specific tools for reliable use.

Renao Yan

Deep learning-based pathological image analysis presents unique challenges due to the practical constraints of network design. Most existing methods apply computer vision models directly to medical tasks, neglecting the distinct characteristics of pathological images. This mismatch often leads to computational inefficiencies, particularly in edge-computing scenarios. To address this, we propose a novel Network Similarity Directed Initialization (NSDI) strategy to improve the stability of neural architecture search (NAS). Furthermore, we introduce domain adaptation into one-shot NAS to better handle variations in staining and semantic scale across pathology datasets. Experiments on the BRACS dataset demonstrate that our method outperforms existing approaches, delivering both superior classification performance and clinically relevant feature localization.

Yunqi Hong, Johnson Kao, Liam Edwards et al.

AI tools in pathology have improved screening throughput, standardized quantification, and revealed prognostic patterns that inform treatment. However, adoption remains limited because most systems still lack the human-readable reasoning needed to audit decisions and prevent errors. We present RECAP-PATH, an interpretable framework that establishes a self-learning paradigm, shifting off-the-shelf multimodal large language models from passive pattern recognition to evidence-linked diagnostic reasoning. At its core is a two-phase learning process that autonomously derives diagnostic criteria: diversification expands pathology-style explanations, while optimization refines them for accuracy. This self-learning approach requires only small labeled sets and no white-box access or weight updates to generate cancer diagnoses. Evaluated on breast and prostate datasets, RECAP-PATH produced rationales aligned with expert assessment and delivered substantial gains in diagnostic accuracy over baselines. By uniting visual understanding with reasoning, RECAP-PATH provides clinically trustworthy AI and demonstrates a generalizable path toward evidence-linked interpretation.

Adam Bajger, Jan Obdržálek, Vojtěch Kůr et al.

We present a clustering-based explainability technique for digital pathology models based on convolutional neural networks. Unlike commonly used methods based on saliency maps, such as occlusion, GradCAM, or relevance propagation, which highlight regions that contribute the most to the prediction for a single slide, our method shows the global behaviour of the model under consideration, while also providing more fine-grained information. The result clusters can be visualised not only to understand the model, but also to increase confidence in its operation, leading to faster adoption in clinical practice. We also evaluate the performance of our technique on an existing model for detecting prostate cancer, demonstrating its usefulness.

Juseung Yun, Sunwoo Yu, Sumin Ha et al.

Cancer progression arises from interactions across multiple biological layers, especially beyond morphological and across molecular layers that remain invisible to image-only models. To capture this broader biological landscape, we present EXAONE Path 2.5, a pathology foundation model that jointly models histologic, genomic, epigenetic and transcriptomic modalities, producing an integrated patient representation that reflects tumor biology more comprehensively. Our approach incorporates three key components: (1) multimodal SigLIP loss enabling all-pairwise contrastive learning across heterogeneous modalities, (2) a fragment-aware rotary positional encoding (F-RoPE) module that preserves spatial structure and tissue-fragment topology in WSI, and (3) domain-specialized internal foundation models for both WSI and RNA-seq to provide biologically grounded embeddings for robust multimodal alignment. We evaluate EXAONE Path 2.5 against six leading pathology foundation models across two complementary benchmarks: an internal real-world clinical dataset and the Patho-Bench benchmark covering 80 tasks. Our framework demonstrates high data and parameter efficiency, achieving on-par performance with state-of-the-art foundation models on Patho-Bench while exhibiting the highest adaptability in the internal clinical setting. These results highlight the value of biologically informed multimodal design and underscore the potential of integrated genotype-to-phenotype modeling for next-generation precision oncology.

Ekaterina Redekop, Mara Pleasure, Vedrana Ivezic et al.

Foundation models in digital pathology use massive datasets to learn useful compact feature representations of complex histology images. However, there is limited transparency into what drives the correlation between dataset size and performance, raising the question of whether simply adding more data to increase performance is always necessary. In this study, we propose a prototype-guided diffusion model to generate high-fidelity synthetic pathology data at scale, enabling large-scale self-supervised learning and reducing reliance on real patient samples while preserving downstream performance. Using guidance from histological prototypes during sampling, our approach ensures biologically and diagnostically meaningful variations in the generated data. We demonstrate that self-supervised features trained on our synthetic dataset achieve competitive performance despite using ~60x-760x less data than models trained on large real-world datasets. Notably, models trained using our synthetic data showed statistically comparable or better performance across multiple evaluation metrics and tasks, even when compared to models trained on orders of magnitude larger datasets. Our hybrid approach, combining synthetic and real data, further enhanced performance, achieving top results in several evaluations. These findings underscore the potential of generative AI to create compelling training data for digital pathology, significantly reducing the reliance on extensive clinical datasets and highlighting the efficiency of our approach.

Deng-Pan Wu, Yan-Su Wei, Li-Xiang Hou et al.

Abstract Background Abnormal microglial polarization phenotypes contribute to the pathogenesis of Alzheimer’s disease (AD). Circular RNAs (circRNAs) have garnered increasing attention due to their significant roles in human diseases. Although research has demonstrated differential expression of circRNAs in AD, their specific functions in AD pathogenesis remain largely unexplored. Methods CircRNA microarray was performed to identify differentially expressed circRNAs in the hippocampus of APP/PS1 and WT mice. The stability of circAPP was assessed via RNase R treatment assay. CircAPP downstream targets miR-1906 and chloride intracellular channel 1 (CLIC1) were identified using bioinformatics and proteomics, respectively. RT-PCR assay was conducted to detect the expression of circAPP, miR-1906 and CLIC1. Morris water maze (MWM) test, passive avoidance test and novel object recognition task were used to detect cognitive function of APP/PS1 mice. Microglial M1/M2 polarization and AD pathology were assessed using Western blot, flow cytometry and Golgi staining assays. CLIC1 expression and channel activity were evaluated using Western blot and functional chloride channel assays, respectively. The subcellular location of circAPP was assessed via FISH and RT-PCR assays. RNA pull-down assay was performed to detect the interaction of miR-1906 with circAPP and 3’ untranslated region (3’UTR) of CLIC1 mRNA. Results In this study, we identified a novel circRNA, named circAPP, that is encoded by amyloid precursor protein (APP) and is implicated in AD. CircAPP is a stable circRNA that was upregulated in Aβ-treated microglial cells and the hippocampus of APP/PS1 mice. Downregulation of circAPP or CLIC1, or overexpression of miR-1906 in microglia modulated microglial M1/M2 polarization in Aβ-treated microglial cells and the hippocampus of APP/PS1 mice, and improved AD pathology and the cognitive function of APP/PS1 mice. Further results revealed that circAPP was mainly distributed in the cytoplasm, and circAPP could regulate CLIC1 expression and channel activity by interacting with miR-1906 and affecting miR-1906 expression, thereby regulating microglial polarization in AD. Conclusions Taken together, our study elucidates the regulatory role of circAPP in AD microglial polarization via miR-1906/CLIC1 axis, and suggests that circAPP may act as a critical player in AD pathogenesis and represent a promising therapeutic target for AD.

Emely Rosbach, Jonathan Ganz, Jonas Ammeling et al.

Artificial intelligence (AI)-based clinical decision support systems (CDSS) promise to enhance diagnostic accuracy and efficiency in computational pathology. However, human-AI collaboration might introduce automation bias, where users uncritically follow automated cues. This bias may worsen when time pressure strains practitioners' cognitive resources. We quantified automation bias by measuring the adoption of negative system consultations and examined the role of time pressure in a web-based experiment, where trained pathology experts (n=28) estimated tumor cell percentages. Our results indicate that while AI integration led to a statistically significant increase in overall performance, it also resulted in a 7% automation bias rate, where initially correct evaluations were overturned by erroneous AI advice. Conversely, time pressure did not exacerbate automation bias occurrence, but appeared to increase its severity, evidenced by heightened reliance on the system's negative consultations and subsequent performance decline. These findings highlight potential risks of AI use in healthcare.

Luzhe Huang, Yuzhu Li, Nir Pillar et al.

Histopathological staining of human tissue is essential for disease diagnosis. Recent advances in virtual tissue staining technologies using artificial intelligence (AI) alleviate some of the costly and tedious steps involved in traditional histochemical staining processes, permitting multiplexed staining and tissue preservation. However, potential hallucinations and artifacts in these virtually stained tissue images pose concerns, especially for the clinical uses of these approaches. Quality assessment of histology images by experts can be subjective. Here, we present an autonomous quality and hallucination assessment method, AQuA, for virtual tissue staining and digital pathology. AQuA autonomously achieves 99.8% accuracy when detecting acceptable and unacceptable virtually stained tissue images without access to histochemically stained ground truth, and presents an agreement of 98.5% with the manual assessments made by board-certified pathologists, including identifying realistic-looking images that could mislead diagnosticians. We demonstrate the wide adaptability of AQuA across various virtually and histochemically stained human tissue images. This framework enhances the reliability of virtual tissue staining and provides autonomous quality assurance for image generation and transformation tasks in digital pathology and computational imaging.

Baraah T. Abu AlSel, Abdelrahman A. Mahmoud, Elham O. Hamed et al.

Metabolic syndrome (MetS) is a worldwide public health challenge. Accumulating evidence implicates elevated serum ferritin and disruptions in iron metabolism as potential elements linked to an increased risk of MetS. This study investigates the relationship between iron homeostasis—including hepcidin levels, serum iron concentration, unsaturated iron-binding capacity (UIBC), and the hepcidin/ferritin (H/F) ratio—and MetS. In this descriptive cross-sectional study, 209 participants aged 24–70 were categorized into two groups: 103 with MetS and 106 without MetS. All participants underwent medical assessment, including anthropometric measures, indices of glycemic control, lipid profiles, and iron-related parameters. Participants were further stratified by the Homeostasis Model Assessment—Insulin Resistance index into three subgroups: insulin-sensitive (IS) (<1.9), early insulin resistance (EIR) (>1.9 to <2.9), and significant insulin resistance (SIR) (>2.9). Notable increments in serum ferritin and hepcidin were observed in the SIR group relative to the IS and EIR groups, with a significant association between metabolic parameters. The UIBC and serum ferritin emerged as significant predictors of MetS, particularly in men, with an area under the curve (AUC) of 0.753 and 0.792, respectively (<i>p</i> ≤ 0.001). In contrast, hepcidin was notably correlated with MetS in women, with an AUC of 0.655 (<i>p</i> = 0.007). The H/F ratio showed superior predictive capability for MetS across both sexes (at cutoff level = 0.67). Among women, this ratio had an AUC of 0.639 (<i>p</i> = 0.015), and for men, it had an AUC of 0.792 (<i>p</i> < 0.001). Hypertension proved an independent risk factor for MetS, affirming its role in metabolic dysregulation. The findings highlight a significant interconnection between iron homeostasis parameters and MetS, with sex-specific variations underscoring the importance of personalized diagnostic criteria. The crucial role of the H/F ratio and the UIBC as emerging predictive markers for MetS indicates their potential utility in identifying at-risk individuals. Further longitudinal research is essential to establish causality and explore the interplay between these biomarkers and MetS.