Hasil untuk "Pathology"

Zhuo Chen, Shawn Young, Lijian Xu

The application of large vision-language models to computational pathology holds great promise for diagnostic assistants but faces a critical computational bottleneck: the gigapixel scale of Whole Slide Images (WSIs). A single WSI typically contains over 105 patches, creating sequence lengths that exceed the constraints of standard Transformer architectures. Existing solutions often resort to spatial sampling, which risks discarding diagnostically critical evidence. To address this, we propose TC-SSA (Token Compression via Semantic Slot Aggregation), a learnable token compression framework that aggregates patch features into a fixed number of semantic slots. A gated routing module assigns patches to slots using sparse Top-2 routing, followed by weighted aggregation, enabling global slide coverage under a strict token budget. The resulting representation retains diagnostically relevant information while reducing the number of visual tokens to 1.7% of the original sequence. On the SlideBench(TCGA), our model achieves 78.34% overall accuracy and 77.14% on the diagnosis subset, outperforming sampling-based baselines under comparable token budgets. The method also generalizes to MIL classification, reaching AUC of 95.83% on TCGA-BRCA, 98.27% on TCGA-NSCLC and 79.80% on PANDA. These results suggest that learnable semantic aggregation provides an effective trade-off between efficiency and diagnostic performance for gigapixel pathology reasoning.

Mingchi Hou, Enno Hermann, Ina Kodrasi

The performance of state-of-the-art speech enhancement (SE) models considerably degrades for pathological speech due to atypical acoustic characteristics and limited data availability. This paper systematically investigates data augmentation (DA) strategies to improve SE performance for pathological speakers, evaluating both predictive and generative SE models. We examine three DA categories, i.e., transformative, generative, and noise augmentation, assessing their impact with objective SE metrics. Experimental results show that noise augmentation consistently delivers the largest and most robust gains, transformative augmentations provide moderate improvements, while generative augmentation yields limited benefits and can harm performance as the amount of synthetic data increases. Furthermore, we show that the effectiveness of DA varies depending on the SE model, with DA being more beneficial for predictive SE models. While our results demonstrate that DA improves SE performance for pathological speakers, a performance gap between neurotypical and pathological speech persists, highlighting the need for future research on targeted DA strategies for pathological speech.

Manqing Liu, David Williams-King, Ida Caspary et al.

Chain-of-thought (CoT) reasoning is fundamental to modern LLM architectures and represents a critical intervention point for AI safety. However, CoT reasoning may exhibit failure modes that we note as pathologies, which prevent it from being useful for monitoring. Prior work has identified three distinct pathologies: post-hoc rationalization, where models generate plausible explanations backwards from predetermined answers; encoded reasoning, where intermediate steps conceal information within seemingly interpretable text; and internalized reasoning, where models replace explicit reasoning with meaningless filler tokens while computing internally. To better understand and discriminate between these pathologies, we create a set of concrete metrics that are simple to implement, computationally inexpensive, and task-agnostic. To validate our approach, we develop model organisms deliberately trained to exhibit specific CoT pathologies. Our work provides a practical toolkit for assessing CoT pathologies, with direct implications for training-time monitoring.

Yuan Zhang, Xinfeng Zhang, Xiaoming Qi et al.

Content generation modeling has emerged as a promising direction in computational pathology, offering capabilities such as data-efficient learning, synthetic data augmentation, and task-oriented generation across diverse diagnostic tasks. This review provides a comprehensive synthesis of recent progress in the field, organized into four key domains: image generation, text generation, molecular profile-morphology generation, and other specialized generation applications. By analyzing over 150 representative studies, we trace the evolution of content generation architectures -- from early generative adversarial networks to recent advances in diffusion models and generative vision-language models. We further examine the datasets and evaluation protocols commonly used in this domain and highlight ongoing limitations, including challenges in generating high-fidelity whole slide images, clinical interpretability, and concerns related to the ethical and legal implications of synthetic data. The review concludes with a discussion of open challenges and prospective research directions, with an emphasis on developing integrated and clinically deployable generation systems. This work aims to provide a foundational reference for researchers and practitioners developing content generation models in computational pathology.

Tianyu Liu, Weihao Xuan, Hao Wu et al.

Advances in AI have introduced several strong models in computational pathology to usher it into the era of multi-modal diagnosis, analysis, and interpretation. However, the current pathology-specific visual language models still lack capacities in making the diagnosis with rigorous reasoning paths as well as handling divergent tasks, and thus, challenges of building AI Copilots for real scenarios still exist. Here we introduce TeamPath, an AI system powered by reinforcement learning and router-enhanced solutions based on large-scale histopathology multimodal datasets, to work as a virtual assistant for expert-level disease diagnosis, patch-level information summarization, and cross-modality generation to integrate transcriptomic information for clinical usage. We also collaborate with pathologists from Yale School of Medicine to demonstrate that TeamPath can assist them in working more efficiently by identifying and correcting expert conclusions and reasoning paths. We also discuss the human evaluation results to support the reasoning quality from TeamPath. Overall, TeamPath can flexibly choose the best settings according to the needs, and serve as an innovative and reliable system for information communication across different modalities and experts.

Kei Taguchi, Kazumasa Ohara, Tatsuya Yokota et al.

We propose a method for representing malignant lymphoma pathology images, from high-resolution cell nuclei to low-resolution tissue images, within a single hyperbolic space using self-supervised learning. To capture morphological changes that occur across scales during disease progression, our approach embeds tissue and corresponding nucleus images close to each other based on inclusion relationships. Using the Poincaré ball as the feature space enables effective encoding of this hierarchical structure. The learned representations capture both disease state and cell type variations.

Hassan Keshvarikhojasteh, Mihail Tifrea, Sibylle Hess et al.

Multiple instance learning (MIL) is a promising approach for weakly supervised classification in pathology using whole slide images (WSIs). However, conventional MIL methods such as Attention-Based Deep Multiple Instance Learning (ABMIL) typically disregard spatial interactions among patches that are crucial to pathological diagnosis. Recent advancements, such as Transformer based MIL (TransMIL), have incorporated spatial context and inter-patch relationships. However, it remains unclear whether explicitly modeling patch relationships yields similar performance gains in ABMIL, which relies solely on Multi-Layer Perceptrons (MLPs). In contrast, TransMIL employs Transformer-based layers, introducing a fundamental architectural shift at the cost of substantially increased computational complexity. In this work, we enhance the ABMIL framework by integrating interaction-aware representations to address this question. Our proposed model, Global ABMIL (GABMIL), explicitly captures inter-instance dependencies while preserving computational efficiency. Experimental results on two publicly available datasets for tumor subtyping in breast and lung cancers demonstrate that GABMIL achieves up to a 7 percentage point improvement in AUPRC and a 5 percentage point increase in the Kappa score over ABMIL, with minimal or no additional computational overhead. These findings underscore the importance of incorporating patch interactions within MIL frameworks. Our code is available at \href{https://github.com/tueimage/GABMIL}{\texttt{GABMIL}}.

Xutong She

Alzheimer's disease (AD) persists as a paramount challenge in neurological research, characterized by the pathological hallmarks of amyloid-beta (Abeta) plaques and neurofibrillary tangles composed of hyperphosphorylated tau. This review synthesizes the evolving understanding of AD pathogenesis, moving beyond the linear amyloid cascade hypothesis to conceptualize the disease as a cross-talk of intricately interacting pathologies, encompassing Abeta, tau, and neuroinflammation. This evolving pathophysiological understanding parallels a transformation in diagnostic paradigms, where biomarker-based strategies -- such as the AT(N) framework -- enable early disease detection during preclinical or prodromal stages. Within this new landscape, while anti-Abeta monoclonal antibodies (e.g., lecanemab, donanemab) represent a breakthrough as the first disease-modifying therapies, their modest efficacy underscores the limitation of single-target approaches. Therefore, this review explores the compelling rationale for combination therapies that simultaneously target Abeta pathology, aberrant tau, and neuroinflammation. Looking forward, we emphasize emerging technological platforms -- such as gene editing and biophysical neuromodulation -- n advancing precision medicine. Ultimately, the integration of early biomarker detection, multi-target therapeutic strategies, and AI-driven patient stratification charts a promising roadmap toward fundamentally altering the trajectory of AD. The future of AD management will be defined by preemptive, biomarker-guided, and personalized combination interventions.

Wangbin Ding, Lei Li, Junyi Qiu et al.

Myocardial infarction (MI) is a leading cause of death worldwide. Late gadolinium enhancement (LGE) and T2-weighted cardiac magnetic resonance (CMR) imaging can respectively identify scarring and edema areas, both of which are essential for MI risk stratification and prognosis assessment. Although combining complementary information from multi-sequence CMR is useful, acquiring these sequences can be time-consuming and prohibitive, e.g., due to the administration of contrast agents. Cine CMR is a rapid and contrast-free imaging technique that can visualize both motion and structural abnormalities of the myocardium induced by acute MI. Therefore, we present a new end-to-end deep neural network, referred to as CineMyoPS, to segment myocardial pathologies, \ie scars and edema, solely from cine CMR images. Specifically, CineMyoPS extracts both motion and anatomy features associated with MI. Given the interdependence between these features, we design a consistency loss (resembling the co-training strategy) to facilitate their joint learning. Furthermore, we propose a time-series aggregation strategy to integrate MI-related features across the cardiac cycle, thereby enhancing segmentation accuracy for myocardial pathologies. Experimental results on a multi-center dataset demonstrate that CineMyoPS achieves promising performance in myocardial pathology segmentation, motion estimation, and anatomy segmentation.

Yuki Anraku, Shunsuke Kita, Taishi Onodera et al.

Abstract Medical treatments using potent neutralizing SARS-CoV-2 antibodies have achieved remarkable improvements in clinical symptoms, changing the situation for the severity of COVID-19 patients. We previously reported an antibody, NT-108 with potent neutralizing activity. However, the structural and functional basis for the neutralizing activity of NT-108 has not yet been understood. Here, we demonstrated the therapeutic effects of NT-108 in a hamster model and its protective effects at low doses. Furthermore, we determined the cryo-EM structure of NT-108 in complex with SARS-CoV-2 spike. The single-chain Fv construction of NT-108 improved the cryo-EM maps because of the prevention of preferred orientations induced by Fab orientation. The footprints of NT-108 illuminated how escape mutations such as E484K evade from class 2 antibody recognition without ACE2 affinity attenuation. The functional and structural basis for the potent neutralizing activity of NT-108 provides insights into the rational design of therapeutic antibodies.

Kristina Hasselgren, Caroline Williamsson, Johanna Wennerblom et al.

Abstract Introduction Patients with pancreatic ductal adenocarcinoma (PDAC) have a dismal prognosis. The majority of patients are diagnosed at an advanced stage, and for these patients, the only possible treatment is palliative chemotherapy. There are increasing data from retrospective studies indicating that a subgroup of patients with liver-limited metastases may benefit from surgical treatment of liver metastases. However, there is a need for prospective trials. Objective The aim of this study is to prospectively investigate the safety and feasibility of surgically treating patients who are resectable, including those with borderline venous resectable, histopathologically confirmed PDAC, and histopathologically or radiologically confirmed liver metastases. Methods Five Swedish and one Finnish hepatopancreaticobiliary (HPB) centre will participate. Eligible patients will be identified at regional multidisciplinary conferences (MDTs). Before inclusion, they will undergo computed tomography (CT), magnetic resonance imaging (MRI, ) and (positron emission tomography computed tomography)PET-CT to rule out extrahepatic metastases. To be included, patients will have to have four or fewer liver metastases, which must be no larger than 5 cm for patients planning for resection and no larger than 2 cm for patients planning for ablation. The metastases may be either synchronous or metachronous. Patients will undergo four months of chemotherapy before surgical treatment (either resection or ablation), and postoperatively, they will undergo two months of chemotherapy. For those with synchronous metastases, resection of the pancreatic tumour will be performed. Follow-up will be performed over two years postoperatively with regular CT scans and assessments of quality of life. Conclusions In conclusion, this trial will provide increased knowledge concerning whether surgical treatment of liver metastases from pancreatic cancer can result in improved survival. Clinical Trial Number Clinical.Trials.gov (NCT05271110), registered February 26th 2022

Martyna Poziomska, Marian Dovgialo, Przemysław Olbratowski et al.

This study investigates the impact of quantity and diversity of data on the performance of various machine-learning models for detecting general EEG pathology. We utilized an EEG dataset of 2,993 recordings from Temple University Hospital and a dataset of 55,787 recordings from Elmiko Biosignals sp. z o.o. The latter contains data from 39 hospitals and a diverse patient set with varied conditions. Thus, we introduce the Elmiko dataset - the largest publicly available EEG corpus. Our findings show that small and consistent datasets enable a wide range of models to achieve high accuracy; however, variations in pathological conditions, recording protocols, and labeling standards lead to significant performance degradation. Nonetheless, increasing the number of available recordings improves predictive accuracy and may even compensate for data diversity, particularly in neural networks based on attention mechanism or transformer architecture. A meta-model that combined these networks with a gradient-boosting approach using handcrafted features demonstrated superior performance across varied datasets.

Vijay Sadashivaiah, Pingkun Yan, James A. Hendler

Deep learning models have revolutionized medical imaging and diagnostics, yet their opaque nature poses challenges for clinical adoption and trust. Amongst approaches to improve model interpretability, concept-based explanations aim to provide concise and human-understandable explanations of any arbitrary classifier. However, such methods usually require a large amount of manually collected data with concept annotation, which is often scarce in the medical domain. In this paper, we propose Conceptual Counterfactual Explanations for Chest X-ray (CoCoX), which leverages the joint embedding space of an existing vision-language model (VLM) to explain black-box classifier outcomes without the need for annotated datasets. Specifically, we utilize textual concepts derived from chest radiography reports and a pre-trained chest radiography-based VLM to explain three common cardiothoracic pathologies. We demonstrate that the explanations generated by our method are semantically meaningful and faithful to underlying pathologies.

Wenhao Tang, Fengtao Zhou, Sheng Huang et al.

Multiple instance learning (MIL) is the most widely used framework in computational pathology, encompassing sub-typing, diagnosis, prognosis, and more. However, the existing MIL paradigm typically requires an offline instance feature extractor, such as a pre-trained ResNet or a foundation model. This approach lacks the capability for feature fine-tuning within the specific downstream tasks, limiting its adaptability and performance. To address this issue, we propose a Re-embedded Regional Transformer (R$^2$T) for re-embedding the instance features online, which captures fine-grained local features and establishes connections across different regions. Unlike existing works that focus on pre-training powerful feature extractor or designing sophisticated instance aggregator, R$^2$T is tailored to re-embed instance features online. It serves as a portable module that can seamlessly integrate into mainstream MIL models. Extensive experimental results on common computational pathology tasks validate that: 1) feature re-embedding improves the performance of MIL models based on ResNet-50 features to the level of foundation model features, and further enhances the performance of foundation model features; 2) the R$^2$T can introduce more significant performance improvements to various MIL models; 3) R$^2$T-MIL, as an R$^2$T-enhanced AB-MIL, outperforms other latest methods by a large margin.The code is available at: https://github.com/DearCaat/RRT-MIL.

Cheryl Sarah Philipose, Sinchana KM, Haritha Haridas et al.

Introduction Human immunodeficiency virus (HIV) and tuberculosis (TB) are two main leading global causes of mortality and morbidity. TB and HIV increase progressive deterioration of immunological functions by speeding progression of one another. Material and methods The present 5-year retrospective study was carried out in the Department of Pathology at a tertiary care hospital in South India. Study included clinico-pathological profile of 80 people living with HIV (PLHIV) and subsequently developed TB co-infection; their CD4+ counts done at the time of admission were examined. Results The present study included 80 HIV-TB co-infected cases. The age of the patients ranged from 18 to 65 years. The mean CD4+ T lymphocyte count was 164.7 cells/μl. Pulmonary TB was diagnosed in 59 patients (73.8%), while extra-pulmonary TB was detected in 21 (26.2%) cases. Abdominal TB was the most common site among extra-pulmonary TB cases. Opportunistic infections (OIs) other than TB, included 2 cases with oral candidiasis and 1 case with central nervous system (CNS) toxoplasmosis. Two of the HIV-TB co-infected cases were subsequently diagnosed with primary CNS (n = 1) and retroperitoneal lymphoma (n = 1). Conclusions In the present study, HIV-TB co-infection is more common in 25-50 years age group. Antiretroviral therapy has changed the nature of disease from fatal to chronic condition. OIs other than TB and neoplasms reported in our study included oral candidiasis, CNS toxoplasmosis, and lymphoma. PLHIV with low CD4+ count require close monitoring, adequate counselling, and further evaluation for atypical presentation of TB, OIs, and neoplasms to improve their outcomes.

Mireia Sueca-Comes, Elena Cristina Rusu, Jennifer C. Ashworth et al.

Maryam Alsadat Hashemipour, Parsa Behnam, Fatemeh Ghasemzadeh

Background: The present work deals with examine the awareness and attitudes of specialists of oral diseases, pathology and maxillofacial surgery in 2022 in oral exfoliative cytology.Methods: This study was an analytical and cross-sectional research. The statistical population of the present study was Iranian oral diseases specialists, pathologists and maxillofacial surgeons. A researcher-made questionnaire was given to the specialists by senior student and then they were asked to complete it . The questionnaire was given to people by the final year student. The obtained results were analyzed using the T-test, Chi-Square and SPSS-18 Software. The significance level in data analysis was considered by P <0.05 level.Results: A total of 210 questionnaires were distributed in the study, of which 192 were examined (Response Rate = 91.4%). The study revealed that only 18 participants used exfoliative cytology technique. Moreover, 62% of people had a positive attitude towards the application and performance of cytology. There was a significant relationship between the positive attitude, field of study, graduation year and age. The mean scores of awareness in men and women were 32.38 ± 4.21 and 34.42 ± 3.89 , respectively. The participants had a high level of awareness and attitude towards this technique.Additionally, the mean score of awareness in oral specialists, surgeons and pathologists was 33.12±4.23, 33.65±5.12 and 33.45±5.34, respectively. The study revealed no significant relationship between awareness score, field of study and gender. Nonetheless, there was a significant relationship between graduation year, age and awareness score.Conclusion: The study revealed that only 9.4 percent of participants used exfoliative cytology technique. However, they had a high level of awareness and attitude towards this technique.Keywords: Awareness, Attitude, Exfoliative Cytology, Dentistry

Bassma Mahmoud Elkhairy, Nabil Mohamed Salama, Abdalrahman Mohammad Desouki et al.

Li Li, Hao Li, Jun Zheng et al.

Introduction Blood blister-like aneurysm (BBA) is a special type of intracranial aneurysm with relatively low morbidity and high mortality. Various microsurgical techniques and endovascular approaches have been reported, but the optimal management remains controversial. For a better understanding of the treatment of BBA patients, a network meta-analysis that comprehensively compares the effects of different therapies is necessary.Methods and analysis This protocol has been reported following the Preferred Reporting Items for Systematic Review and Meta-Analysis Protocols. Related studies in the following databases will be searched until November 2022: PubMed, Embase, Scopus, Web of Science, Cochrane Library, China National Knowledge Infrastructure (CNKI), VIP and Wanfang. Randomised controlled trials (RCTs) and non-randomised studies comparing at least two different interventions in BBA patients will be included. Quality assessment will be conducted using Cochrane Collaboration’s tool or Newcastle-Ottawa Scale based on their study designs. The primary outcome is the composite of the incidences of intraoperative bleeding, postoperative bleeding and postoperative recurrence. The secondary outcome is an unfavourable functional outcome. Pairwise and network meta-analyses will be conducted using STATA V.14 (StataCorp, College Station, Texas, USA). Mean ranks and the surface under the cumulative ranking curve will be used to evaluate every intervention. Statistical inconsistency assessment, subgroup analysis, sensitivity analysis and publication bias assessment will be performed.Ethics and dissemination Ethics approval is not necessary because this study will be based on publications. The results of this study will be published in a peer-reviewed journal.PROSPERO registration number CRD42022383699.

Philippe Weitz, Masi Valkonen, Leslie Solorzano et al.

The analysis of FFPE tissue sections stained with haematoxylin and eosin (H&E) or immunohistochemistry (IHC) is an essential part of the pathologic assessment of surgically resected breast cancer specimens. IHC staining has been broadly adopted into diagnostic guidelines and routine workflows to manually assess status and scoring of several established biomarkers, including ER, PGR, HER2 and KI67. However, this is a task that can also be facilitated by computational pathology image analysis methods. The research in computational pathology has recently made numerous substantial advances, often based on publicly available whole slide image (WSI) data sets. However, the field is still considerably limited by the sparsity of public data sets. In particular, there are no large, high quality publicly available data sets with WSIs of matching IHC and H&E-stained tissue sections. Here, we publish the currently largest publicly available data set of WSIs of tissue sections from surgical resection specimens from female primary breast cancer patients with matched WSIs of corresponding H&E and IHC-stained tissue, consisting of 4,212 WSIs from 1,153 patients. The primary purpose of the data set was to facilitate the ACROBAT WSI registration challenge, aiming at accurately aligning H&E and IHC images. For research in the area of image registration, automatic quantitative feedback on registration algorithm performance remains available through the ACROBAT challenge website, based on more than 37,000 manually annotated landmark pairs from 13 annotators. Beyond registration, this data set has the potential to enable many different avenues of computational pathology research, including stain-guided learning, virtual staining, unsupervised pre-training, artefact detection and stain-independent models.