Fig 4 - available via license: Creative Commons Attribution 4.0 International
Content may be subject to copyright.
| Patch sizes and magnifications employed by studies. Studies were taken into account if at least one of the two pieces of information was specified. Other patch sizes and magnifications employed by single studies (e.g., 150 × 150 pixels, 4x magnification) are not shown.
Source publication
In recent years, the diagnosis of gliomas has become increasingly complex. Analysis of glioma histopathology images using artificial intelligence (AI) offers new opportunities to support diagnosis and outcome prediction. To give an overview of the current state of research, this review examines 83 publicly available research studies that have propo...
Context in source publication
Context 1
... patches at 40x magnification 14 ; and highest performance for survival risk group classification consistent across multiple CNN architectures by utilizing 256 × 256 pixels patches at 20x magnification 85 . Apart from these comparisons, smaller patches, i.e., 224 × 224 and 256 × 256 pixels, and 20x magnification were most frequently employed (Fig. ...
Citations
... Further, in scenarios with limited labeled data, few-shot learning techniques have made significant progress. The best methods have achieved accuracies exceeding 70%, 80%, and 85% in 5-way 1-shot, 5-way 5-shot, and 5-way 10-shot cases, respectively [174]. ...
Computer vision and artificial intelligence have revolutionized the field of pathological image analysis, enabling faster and more accurate diagnostic classification. Deep learning architectures like convolutional neural networks (CNNs), have shown superior performance in tasks such as image classification, segmentation, and object detection in pathology. Computer vision has significantly improved the accuracy of disease diagnosis in healthcare. By leveraging advanced algorithms and machine learning techniques, computer vision systems can analyze medical images with high precision, often matching or even surpassing human expert performance. In pathology, deep learning models have been trained on large datasets of annotated pathology images to perform tasks such as cancer diagnosis, grading, and prognostication. While deep learning approaches show great promise in diagnostic classification, challenges remain, including issues related to model interpretability, reliability, and generalization across diverse patient populations and imaging settings.
... Against this background, AI applications in pathological image analysis of brain tumors have gained attention 7 . While molecular diagnostic techniques provide crucial information for treatment selection, their time and resource requirements pose practical challenges, particularly for initial treatment planning. ...
Foundation models pretrained on large-scale pathology datasets have shown promising results across various diagnostic tasks. Here, we present a systematic evaluation of transfer learning strategies for brain tumor classification using these models. We analyzed 252 cases comprising five major tumor types: glioblastoma, astrocytoma, oligodendroglioma, primary central nervous system lymphoma, and metastatic tumors. Comparing state-of-the-art foundation models with conventional approaches, we found that foundation models demonstrated robust classification performance with as few as 10 patches per case, challenging the traditional assumption that extensive per-case image sampling is necessary. Furthermore, our evaluation revealed that simple transfer learning strategies like linear probing were sufficient, while fine-tuning often degraded model performance. These findings suggest a paradigm shift from extensive data collection to efficient utilization of pretrained features, providing practical implications for implementing AI-assisted diagnosis in clinical pathology.
... One of the broadest applications of AI in neuro-oncology concerns its use by neuropathologists to analyze histological images. This application encompasses the automated measurement of features, assistance in tumor classification and grading, enhancement of tumor detection, and comprehensive analysis of cellular and tissue structures through histomolecular classification [10]. Bioinformatics also represents a significant tool for brain tumor research [11,12]. ...
... Diffuse adult-type tumors encompass the majority of primary brain tumors encountered in adult neuro-oncology practice, including astrocytoma (IDH-mutant) and glioblastoma (IDH-wild type). Adult-type diffuse gliomas are stratified according to histological features (increased mitoses, necrosis and/or microvascular proliferation) and molecular alterations, including mutations in isocitrate dehydrogenase ½ (IDH1/2) genes and the whole-arm codeletion of chromosomes 1p and 19q [10]. From a diagnostic perspective, the intraoperative delineation of glial tumors poses a significant challenge, yet Photonics 2025, 12, 37 3 of 28 this can be addressed through the use of spectroscopic techniques. ...
... tant) and glioblastoma (IDH-wild type). Adult-type diffuse gliomas are stratified according to histological features (increased mitoses, necrosis and/or microvascular proliferation) and molecular alterations, including mutations in isocitrate dehydrogenase ½ (IDH1/2) genes and the whole-arm codeletion of chromosomes 1p and 19q [10]. From a diagnostic perspective, the intraoperative delineation of glial tumors poses a significant challenge, yet this can be addressed through the use of spectroscopic techniques. ...
Decision support systems based on machine learning (ML) techniques are already empowering neuro-oncologists. These systems provide comprehensive diagnostics, offer a deeper understanding of diseases, predict outcomes, and assist in customizing treatment plans to individual patient needs. Collectively, these elements represent artificial intelligence (AI) in neuro-oncology. This paper reviews recent studies which apply machine learning algorithms to optical spectroscopy data from central nervous system (CNS) tumors, both ex vivo and in vivo. We first cover general issues such as the physical basis of the optical-spectral methods used in neuro-oncology, and the basic algorithms used in spectral signal preprocessing, feature extraction, data clustering, and supervised classification methods. Then, we review in more detail the methodology and results of applying ML techniques to fluorescence, elastic and inelastic scattering, and IR spectroscopy.
... In addition, [15] developed a rapid intraoperative molecular diagnostic method for glioma using ultrasound radiofrequency signals, showcasing the unique advantages of ultrasound in this context. Moreover [16] examined a review of 70 research studies that used AI for analyzing histopathology images of human gliomas, almost of them have applied Convolutional Neural Network (CNN) models on The Cancer Genome Atlas (TCGA) dataset, in order to diagnosis various tasks such as subtyping, grading, molecular marker prediction, and survival prediction. The current research focus claimed by the author were the assessment of haematoxylin and eosin-stained tissue sections of adult-type diffuse gliomas. ...
Brain tumors are among the most serious and fatal cancers, claiming thousands of lives annually. Early diagnosis, which typically relies on MRI, is crucial for saving lives. Recently, many Artificial Intelligence (AI) techniques have been proposed to alleviate radiologists' workload and expedite diagnosis. Most of these methods utilize deep learning (DL), which, while powerful, can be less reliable and resource-intensive. A significant challenge is the labeling of large and frequently updated biomedical datasets for training DL models. This paper proposes an algorithm that addresses these challenges by leveraging weak annotation and eXplainable AI (XAI). The algorithm also incorporates various image processing steps, such as feature extraction and clustering, to determine both the type and degree of brain tumors. Our study demonstrates state-of-the-art performance, achieving 96% accuracy in the binary classification of Glioma and Meningioma tumors using Random Forest (RF). Additionally, the algorithm provides transparent clustering of Glioma into Low-Grade Glioma (LGG) and High-Grade Glioma (HGG) after type classification. The algorithm is also designed to offer XAI results for explaining DL outcomes and will continue to be developed in this area.
... Recently, it has been shown that the assessment of histopathology and radiology imaging data with automated algorithms (eg, the assessment of allograft rejection status and other pathology in kidney biopsies, and alternatively also in cancer diagnostics) that computational algorithms and in particular AI-enhanced analysis/case diagnosis of highly complex imaging data can be even faster and more accurate, as opposed to interpretation/diagnosis by human individual pathologists and teams. [103][104][105][106][107][108][109][110][111][112][113] Here, the assessment of kidney histopathology or allograft rejection status, even when conducted by a team of highly experienced histopathologists, can still deviate within a team of specialists for some individual cases, which can lead to reclassification by pathologists in as much as 30%-50% of rejection cases in some extreme settings. 108 This variability in human interpretation may be a crucial challenge to be addressed in the coming years by increasing the integration of AI-enhanced analysis/interpretation of data and to overcome the increasing shortages of pathologists. ...
... 104,105 Similarly, AI-assisted identification of cancer diagnosis and etiology according to WHO classification can be of advantage to human capability in imaging analysis. [109][110][111][112] Replacing one pathologist with AI led to a 4% higher noninferior cancer detection rate compared with radiologist double reading in screening mammograms. 112 Similar applies to SC research, where humans can be overwhelmed by the large amount of data that are generated in both research and clinical trial contexts. ...
Advanced bioinformatics analysis, such as systems biology (SysBio) and artificial intelligence (AI) approaches, including machine learning (ML) and deep learning (DL), is increasingly present in stem cell (SC) research. An approximate timeline on these developments and their global impact is still lacking. We conducted a scoping review on the contribution of SysBio and AI analysis to SC research and therapy development based on literature published in PubMed between 2000 and 2024. We identified an 8-10-fold increase in research output related to all three search terms between 2000 and 2021, with a 10-fold increase in AI-related production since 2010. Use of SysBio and AI still predominates in preclinical basic research with increasing use in clinically oriented translational medicine since 2010. SysBio- and AI-related research was found all over the globe, with SysBio output led by the United States (US, n=1487), United Kingdom (UK, n=1094), Germany (n=355), The Netherlands (n=339), Russia (n=215), and France (n=149), while for AI-related research the US (n=853) and UK (n=258) take a strong lead, followed by Switzerland (n=69), The Netherlands (n=37), and Germany (n=19). The US and UK are most active in SCs publications related to AI/ML and AI/DL. The prominent use of SysBio in ESC research was recently overtaken by prominent use of AI in iPSC and MSC research. This study reveals the global evolution and growing intersection between AI, SysBio, and SC research over the past two decades, with substantial growth in all three fields and exponential increases in AI-related research in the past decade.
... CPath methods for brain tumors have primarily focused on the adult population [21], with few studies exclusively investigating pediatric tumors for diagnosis or survival prediction [22,23,24,25] mainly due to the unavailability of large datasets. Among these, Steyaer et at., proposed a deep learning-based approach for survival prediction of pediatric brain tumors (low-grade glioma, high-grade astrocytoma, high-grade ependymoma, and high-grade medulloblastoma) and adult (low-grade glioma, glioblastoma) using H&E WSIs and genetic data [25]. ...
Brain tumors are the most common solid tumors in children and young adults, but the scarcity of large histopathology datasets has limited the application of computational pathology in this group. This study implements two weakly supervised multiple-instance learning (MIL) approaches on patch-features obtained from state-of-the-art histology-specific foundation models to classify pediatric brain tumors in hematoxylin and eosin whole slide images (WSIs) from a multi-center Swedish cohort. WSIs from 540 subjects (age 8.54.9 years) diagnosed with brain tumor were gathered from the six Swedish university hospitals. Instance (patch)-level features were obtained from WSIs using three pre-trained feature extractors: ResNet50, UNI and CONCH. Instances were aggregated using attention-based MIL (ABMIL) or clustering-constrained attention MIL (CLAM) for patient-level classification. Models were evaluated on three classification tasks based on the hierarchical classification of pediatric brain tumors: tumor category, family and type. Model generalization was assessed by training on data from two of the centers and testing on data from four other centers. Model interpretability was evaluated through attention-mapping. The highest classification performance was achieved using UNI features and AMBIL aggregation, with Matthew's correlation coefficient of 0.860.04, 0.630.04, and 0.530.05, for tumor category, family and type classification, respectively. When evaluating generalization, models utilizing UNI and CONCH features outperformed those using ResNet50. However, the drop in performance from the in-site to out-of-site testing was similar across feature extractors. These results show the potential of state-of-the-art computational pathology methods in diagnosing pediatric brain tumors at different hierarchical levels with fair generalizability on a multi-center national dataset.
... This research is essential for understanding the biological mechanisms that cause cancer and for selecting targeted therapies. Furthermore, studies have been conducted to estimate gene mutations and protein expression levels from pathology images to reduce the delay in patient treatment initiation due to the weeks-long time required for clinicians to obtain genetic and immunological test results(11) , (12). Such research is progressing in the development of AI that integrates multimodal information and presents it in an interpretable form for pathologists and clinicians. ...
Pathology has played a crucial role in the diagnosis and evaluation of patient tissue samples obtained from surgeries and biopsies for many years. The advent of Whole Slide Scanners and the development of deep learning technologies have significantly advanced the field, leading to extensive research and development in pathology AI (Artificial Intelligence). These advancements have contributed to reducing the workload of pathologists and supporting decision-making in treatment plans. Recently, large-scale AI models known as Foundation Models (FMs), which are more accurate and applicable to a wide range of tasks compared to traditional AI, have emerged, and expanded their application scope in the healthcare field. Numerous FMs have been developed in pathology, and there are reported cases of their application in various tasks, such as disease diagnosis, rare cancer diagnosis, patient survival prognosis prediction, biomarker expression prediction, and the scoring of immunohistochemical expression intensity. However, several challenges remain for the clinical application of FMs, which healthcare professionals, as users, must be aware of. Research is ongoing to address these challenges. In the future, it is expected that the development of Generalist Medical AI, which integrates pathology FMs with FMs from other medical domains, will progress, leading to the effective utilization of AI in real clinical settings to promote precision and personalized medicine.
... Glioblastoma is the most aggressive type of brain tumor, with a median survival of 15 months. Its treatment efficacy remains low despite the combination of complete surgical resection, radiotherapy, and temozolomide (TMZ) chemotherapy, as well as significant advances in glioma diagnosis, including the development of deep learning-based approaches (Rui et al., 2023;Redlich et al., 2024). Complications of glioblastoma treatment include tumor infiltration into the brain tissue, high recurrence rates, and resistance to irradiation and TMZ (Rodríguez-Camacho et al., 2022). ...
Soloxolone amides are semisynthetic triterpenoids that can cross the blood-brain barrier and inhibit glioblastoma growth both in vitro and in vivo. Here we investigate the impact of these compounds on processes associated with glioblastoma invasiveness and therapy resistance. Screening of soloxolone amides against glioblastoma cells revealed the ability of compound 7 (soloxolone para-methylanilide) to inhibit transforming growth factor-beta 1 (TGF-β1)-induced glial-mesenchymal transition Compound 7 inhibited morphological changes, wound healing, transwell migration, and expression of mesenchymal markers (N-cadherin, fibronectin, Slug) in TGF-β1-induced U87 and U118 glioblastoma cells, while restoring their adhesiveness. Confocal microscopy and molecular docking showed that 7 reduced SMAD2/3 nuclear translocation probably by direct interaction with the TGF-β type I and type II receptors (TβRI/II). In addition, 7 suppressed stemness of glioblastoma cells as evidenced by inhibition of colony forming ability, spheroid growth, and aldehyde dehydrogenase (ALDH) activity. Furthermore, 7 exhibited a synergistic effect with temozolomide (TMZ) on glioblastoma cell viability. Using N-acetyl-L-cysteine (NAC) and flow cytometry analysis of Annexin V-FITC-, propidium iodide-, and DCFDA-stained cells, 7 was found to synergize the cytotoxicity of TMZ by inducing ROS-dependent apoptosis. Further in vivo studies showed that 7, alone or in combination with TMZ, effectively suppressed the growth of U87 xenograft tumors in mice. Thus, 7 demonstrated promising potential as a component of combination therapy for glioblastoma, reducing its invasiveness and increasing its sensitivity to chemotherapy.
Rapid advances in computer vision (CV) and artificial intelligence have opened new avenues for digital pathology, including the diagnosis and treatment of central nervous system (CNS) tumors. In addition to reviewing the state-of-the-art in CV-based digital pathology and highlighting its potential to revolutionize the field, this chapter also provides a general introduction to digital pathology and Machine Learning (ML) for neuropathologists. Although currently limited to research, the integration of CV tools into digital pathology already offers significant advantages, such as automating tissue analysis and providing quantitative assessments. The transition from research to clinical application is slowly gaining momentum. To provide neuropa-thologists with the necessary skills to succeed in digital pathology and ML, the chapter also discusses how physicians and researchers can create custom models and tools tailored to specific needs using tools such as nnU-Net, deepflash , and PathML. Emphasis is placed on the importance of interdisciplinary collaboration and continued research to fully realize the potential of CV in digital pathology for CNS tumors, to address the challenges of workforce shortages and increased workloads in neuropathology.
