Table 2 - uploaded by Kate Lillard
Content may be subject to copyright.
Source publication
The advent of whole-slide imaging in digital pathology has brought about the advancement of computer-aided examination of tissue via digital image analysis. Digitized slides can now be easily annotated and analyzed via a variety of algorithms. This study reviews the fundamentals of tissue image analysis and aims to provide pathologists with basic i...
Contexts in source publication
Context 1
... software provides a collaborative option for image analysis, whereas commercial software provides more personalized image analysis choices. Each option presents certain advantages and disadvantages [ Table 2]. Given the number of open-source and commercial software solutions, it is often difficult to choose which tool is appropriate for a given task. ...
Context 2
... software provides a collaborative option for image analysis, whereas commercial software provides more personalized image analysis choices. Each option presents certain advantages and disadvantages [ Table 2]. Given the number of open-source and commercial software solutions, it is often difficult to choose which tool is appropriate for a given task. ...
Similar publications
People generally wish to have a lighter or a darker skin. Based on this wish, they regularly compare their skins regarding darkness while there is no simple and accessible system to perform that. This research proposes an algorithm and implements the algorithm in a simple and accessible computerized system that is capable to compare skins of two pe...
Citations
... Glass slides are scanned with a scanning apparatus to create digital slides, which are high-resolution digital pictures that may be seen on computer screens or mobile devices. With the introduction of digital scanners that can generate virtual slides, the area of digital pathology (DP) has arisen as a critical avenue in diagnostic medicine [1]. This revolutionary technology provides high-resolution images of histology slides, allowing for more accurate, faster, and cost-effective cancer diagnosis, prognosis, and prediction [2]. ...
... Through spatial scanning of the imaging sensor, a wide range of scenes under observation can be reconstructed using a FoV stitch. Whole-slide imaging [16][17][18][19] is the FoV stitch application in the microscopy field. A whole-slide image in high spatial resolution can be reconstructed from a series of magnified images in the micro-objective-limited FoVs via scanning over the specimen slide. ...
The space-bandwidth product (SBP) limitation makes it difficult to obtain an image with both a high spatial resolution and a large field of view (FoV) through commonly used optical imaging systems. Although FoV and spectrum stitch provide solutions for SBP expansion, they rely on spatial and spectral scanning, which lead to massive image captures and a low processing speed. To solve the problem, we previously reported a physics-driven deep SBP-expanded framework (Deep SBP $+$ + ) [J. Opt. Soc. Am. A 40, 833 (2023)JOAOD60740-323210.1364/JOSAA.480920]. Deep SBP $+$ + can reconstruct an image with both high spatial resolution and a large FoV from a low-spatial-resolution image in a large FoV and several high-spatial-resolution images in sub-FoVs. In physics, Deep SBP $+$ + reconstructs the convolution kernel between the low- and high-spatial-resolution images and improves the spatial resolution through deconvolution. But Deep SBP $+$ + needs multiple high-spatial-resolution images in different sub-FoVs, inevitably complicating the operations. To further reduce the image captures, we report an updated version of Deep SBP $+$ + 2.0, which can reconstruct an SBP expanded image from a low-spatial-resolution image in a large FoV and another high-spatial-resolution image in a sub-FoV. Different from Deep SBP $+$ + , the assumption that the convolution kernel is a Gaussian distribution is added to Deep SBP $+$ + 2.0 to make the kernel calculation simple and in line with physics. Moreover, improved deep neural networks have been developed to enhance the generation capability. Proven by simulations and experiments, the receptive field is analyzed to prove that a high-spatial-resolution image in the sub-FoV can also guide the generation of the entire FoV. Furthermore, we also discuss the requirement of the sub-FoV image to obtain an SBP-expanded image of high quality. Considering its SBP expansion capability and convenient operation, the updated Deep SBP $+$ + 2.0 can be a useful tool to pursue images with both high spatial resolution and a large FoV.
... However, the research field of AI-driven salivary gland tumor analysis is still in development mostly due to the limited number of cases in the same center of diagnosis [69]. A reduced number of parotid gland tumors (ranging from 25 to 293) analyzed with AI algorithms have been reported in the recent literature [58,62,[69][70][71][72][73][74][75][76][77][78]. All the data of these studies are examined retrospectively. ...
... For analyzing histopathological slides, the important feature to develop in deep learning is to enable computers to automatically extract features from the images and build an algorithm of diagnosis. For the pathological diagnosis, the digitized imaging of slides (represented by static images of individual fields-of-view) and whole-slide imaging are used by the new branch of pathology-computational pathology [74]. Deep learning methods, mostly of convolutional neural networks, have been used as computational pathology techniques for the analysis of images of bladder, lung, brain, breast, skin, digestive and genital tumors. ...
Salivary glands tumors are uncommon neoplasms with variable incidence, heterogenous histologies and unpredictable biological behaviour. Most tumors are located in the parotid gland. Benign salivary tumors represent 54–79% of cases and pleomorphic adenoma is frequently diagnosed in this group. Salivary glands malignant tumors that are more commonly diagnosed are adenoid cystic carcinomas and mucoepidermoid carcinomas. Because of their diversity and overlapping features, these tumors require complex methods of evaluation. Diagnostic procedures include imaging techniques combined with clinical examination, fine needle aspiration and histopathological investigation of the excised specimens. This narrative review describes the advances in the diagnosis methods of these unusual tumors—from histomorphology to artificial intelligence algorithms.
... In medical image analysis, a Whole Slide Image (WSI) [36] is challenging to fit entirely into a network due to its gigapixel size (more than 10 8 pixels). Existing medical image analysis typically segments them into multiple patches as images [1,16,20,46,49], yet these multiple patches still represent the same sample. Therefore, applying MLLMs to multimodal tasks with richer visual inputs holds much practical significance. ...
Multimodal Large Language Models (MLLMs) have achieved SOTA performance in various visual language tasks by fusing the visual representations with LLMs leveraging some visual adapters. In this paper, we first establish that adapters using query-based Transformers such as Q-former is a simplified Multi-instance Learning method without considering instance heterogeneity/correlation. We then propose a general component termed Multi-instance Visual Prompt Generator (MIVPG) to incorporate enriched visual representations into LLMs by taking advantage of instance correlation between images or patches for the same sample. Quantatitive evaluation on three public vision-language (VL) datasets from different scenarios shows that the proposed MIVPG improves Q-former in main VL tasks.
... In such an evolving scenario, artificial intelligence (AI) technologies, including machine learning (ML), deep learning (DL), and convolutional neural networks (CNNs), emerge as transformative tools to enhance predictive algorithms precisely assessing the risk of recurrence in EBC [14][15][16]. These AI applications, particularly in digital and computational pathology, allow for the extraction of subvisual morphometric phenotypes, potentially leading to diagnostic breakthroughs [17][18][19][20][21][22][23]. However, their application both in clinical studies and real-world clinical practice faces several unresolved challenges [24]. ...
Effective risk assessment in early breast cancer is essential for informed clinical decision-making, yet consensus on defining risk categories remains challenging. This paper explores evolving approaches in risk stratification, encompassing histopathological, immunohistochemical, and molecular biomarkers alongside cutting-edge artificial intelligence (AI) techniques. Leveraging machine learning, deep learning, and convolutional neural networks, AI is reshaping predictive algorithms for recurrence risk, thereby revolutionizing diagnostic accuracy and treatment planning. Beyond detection, AI applications extend to histological subtyping, grading, lymph node assessment, and molecular feature identification, fostering personalized therapy decisions. With rising cancer rates, it is crucial to implement AI to accelerate breakthroughs in clinical practice, benefiting both patients and healthcare providers. However, it is important to recognize that while AI offers powerful automation and analysis tools, it lacks the nuanced understanding, clinical context, and ethical considerations inherent to human pathologists in patient care. Hence, the successful integration of AI into clinical practice demands collaborative efforts between medical experts and computational pathologists to optimize patient outcomes.
... WSIs present tumor morphology in very detailed, gigapixel resolution, which is extremely useful in cancer studies, enabling computational analysis and remote assessment by experts. Automatic computational analysis of WSI is a rapidly expanding field in medical image analysis, which can alleviate pathologists' workloads and help reduce the chance of diagnostic errors (Janowczyk and Madabhushi, 2016;Aeffner et al., 2019;Niazi et al., 2019;Baheti et al., 2023c). The broader use of WSI has, in turn, resulted in substantial developments in computational analysis of histopathology imaging, particularly for gaining novel insights from population-based studies. ...
... Prognostic stratification through WSIs represents a burgeoning field at the crossroads of medical imaging and machine learning. WSIs offer valuable insights into a patient's tissue sample, encompassing its histological features, morphology, physiology, and biology (Aeffner et al., 2019). A crucial step in our study was the reclassification of TCGA-GBM (Scarpace et al., 2016) and TCGA-LGG (Pedano et al., 2016) datasets, aligning them with the 2021 WHO CNS classification criteria (Louis et al., 2021). ...
Introduction
Glioblastoma (GBM) is a highly aggressive malignant tumor of the central nervous system that displays varying molecular and morphological profiles, leading to challenging prognostic assessments. Stratifying GBM patients according to overall survival (OS) from H&E-stained whole slide images (WSI) using advanced computational methods is challenging, but with direct clinical implications.
Methods
This work is focusing on GBM (IDH-wildtype, CNS WHO Gr.4) cases, identified from the TCGA-GBM and TCGA-LGG collections after considering the 2021 WHO classification criteria. The proposed approach starts with patch extraction in each WSI, followed by comprehensive patch-level curation to discard artifactual content, i.e., glass reflections, pen markings, dust on the slide, and tissue tearing. Each patch is then computationally described as a feature vector defined by a pre-trained VGG16 convolutional neural network. Principal component analysis provides a feature representation of reduced dimensionality, further facilitating identification of distinct groups of morphology patterns, via unsupervised k-means clustering.
Results
The optimal number of clusters, according to cluster reproducibility and separability, is automatically determined based on the rand index and silhouette coefficient, respectively. Our proposed approach achieved prognostic stratification accuracy of 83.33% on a multi-institutional independent unseen hold-out test set with sensitivity and specificity of 83.33%.
Discussion
We hypothesize that the quantification of these clusters of morphology patterns, reflect the tumor's spatial heterogeneity and yield prognostic relevant information to distinguish between short and long survivors using a decision tree classifier. The interpretability analysis of the obtained results can contribute to furthering and quantifying our understanding of GBM and potentially improving our diagnostic and prognostic predictions.
... Deep learning (DL), is a subset of ML emerged in the 1980s, and utilizes multi-layer neural networks to process data in a way that mimics human neural connections [17]. Essentially, these AI approaches are designed to extract meaningful image representations, which are subsequently processed by specific machine classifiers, based on specific criteria (segmentation, diagnostics, or prognostics) using supervised or unsupervised methods [18,19]. ...
... These insights bridge the gap to the emerging field of tumor profiling and precision oncology, emphasizing the need for predictive assays to stratify patients for personalized treatments. The analytical prowess of AI holds the potential to accelerate the discovery of novel histopathological biomarkers, predicting disease progression, tumor recurrence, and treatment response [18,25]. Recent years have seen increased AI implementation in histological imaging analysis, revolutionizing personalized therapy steps, including tumor detection, staging, and molecular subtyping for enhanced patient care. ...
Simple Summary
This review highlights the profound impact of digital pathology (DP) and artificial intelligence (AI) on advancing cancer diagnosis and treatment. DP enables pathologists to access, analyze, and share high-resolution images, enhancing diagnostic accuracy and fostering remote collaboration. AI further refines cancer diagnosis by automating tasks and facilitating spatial analysis of the tumor microenvironment (TME), leading to the discovery of novel biomarkers. Immunoscore (IS), an AI-assisted immune assay, exhibits robust potential in improving cancer diagnosis, prognosis, and treatment selection, surpassing traditional staging systems. Integrating DP and AI, particularly the IS biomarker, into clinical practice promises to enhance personalized cancer therapy. The research underscores a pivotal leap forward in pathology, stressing the imperative of incorporating AI-driven technologies for improved cancer patient care and outcomes. This exploration aims to provide insights into the transformative potential of DP in cancer management, influencing the clinical community towards more effective diagnostic and therapeutic strategies.
Abstract
(1) Background: Digital pathology (DP) is transforming the landscape of clinical practice, offering a revolutionary approach to traditional pathology analysis and diagnosis. (2) Methods: This innovative technology involves the digitization of traditional glass slides which enables pathologists to access, analyze, and share high-resolution whole-slide images (WSI) of tissue specimens in a digital format. By integrating cutting-edge imaging technology with advanced software, DP promises to enhance clinical practice in numerous ways. DP not only improves quality assurance and standardization but also allows remote collaboration among experts for a more accurate diagnosis. Artificial intelligence (AI) in pathology significantly improves cancer diagnosis, classification, and prognosis by automating various tasks. It also enhances the spatial analysis of tumor microenvironment (TME) and enables the discovery of new biomarkers, advancing their translation for therapeutic applications. (3) Results: The AI-driven immune assays, Immunoscore (IS) and Immunoscore-Immune Checkpoint (IS-IC), have emerged as powerful tools for improving cancer diagnosis, prognosis, and treatment selection by assessing the tumor immune contexture in cancer patients. Digital IS quantitative assessment performed on hematoxylin–eosin (H&E) and CD3+/CD8+ stained slides from colon cancer patients has proven to be more reproducible, concordant, and reliable than expert pathologists’ evaluation of immune response. Outperforming traditional staging systems, IS demonstrated robust potential to enhance treatment efficiency in clinical practice, ultimately advancing cancer patient care. Certainly, addressing the challenges DP has encountered is essential to ensure its successful integration into clinical guidelines and its implementation into clinical use. (4) Conclusion: The ongoing progress in DP holds the potential to revolutionize pathology practices, emphasizing the need to incorporate powerful AI technologies, including IS, into clinical settings to enhance personalized cancer therapy.
... Recently, the Food and Drug Administration (FDA) has approved the use of digital pathology for primary diagnosis [30]. Nevertheless, the abundance and intricate information among different cell types, along with the spatial context provided by digital pathology, has underscored the necessity for precise analysis of large datasets [31][32][33]. Thus, implementing robust and reproducible AI-based methods might potentially resolve the challenges faced by oncologists and pathologists. ...
The diagnosis of cancer is typically based on histopathological sections or biopsies on glass slides. Artificial intelligence (AI) approaches have greatly enhanced our ability to extract quantitative information from digital histopathology images as a rapid growth in oncology data. Gynecological cancers are major diseases affecting women's health worldwide. They are characterized by high mortality and poor prognosis, underscoring the critical importance of early detection, treatment, and identification of prognostic factors. This review highlights the various clinical applications of AI in gynecological cancers using digitized histopathology slides. Particularly, deep learning models have shown promise in accurately diagnosing, classifying histopathological subtypes, and predicting treatment response and prognosis. Furthermore, the integration with transcriptomics, proteomics, and other multi-omics techniques can provide valuable insights into the molecular features of diseases. Despite the considerable potential of AI, substantial challenges remain. Further improvements in data acquisition and model optimization are required, and the exploration of broader clinical applications, such as the biomarker discovery, need to be explored.
... Upon reviewing the document on the application of digital pathology in the analysis of p16/Ki-67 dual staining, a deeper exploration reveals that this innovative approach significantly enhances the detection and interpretation of precancerous and cancerous lesions in cervical cytology [98,99]. By leveraging sophisticated image analysis algorithms, digital pathology enables a standardized and automated assessment of stained slides, reducing subjective interpretation variability [100]. This method improves diagnostic accuracy, particularly in identifying high-grade cervical intraepithelial neoplasia. ...
Cervical cancer, primarily caused by high-risk human papillomavirus (HR-HPV) types 16 and 18, is a major global health concern. Persistent HR-HPV infection can progress from reversible precancerous lesions to invasive cervical cancer, which is driven by the oncogenic activity of human papillomavirus (HPV) genes, particularly E6 and E7. Traditional screening methods, including cytology and HPV testing, have limited sensitivity and specificity. This review explores the application of p16/Ki-67 dual-staining cytology for cervical cancer screening. This advanced immunocytochemical method allows for simultaneously detecting p16 and Ki-67 proteins within cervical epithelial cells, offering a more specific approach for triaging HPV-positive women. Dual staining and traditional methods are compared, demonstrating their high sensitivity and negative predictive value but low specificity. The increased sensitivity of dual staining results in higher detection rates of CIN2+ lesions, which is crucial for preventing cervical cancer progression. However, its low specificity may lead to increased false-positive results and unnecessary biopsies. The implications of integrating dual staining into contemporary screening strategies, particularly considering the evolving landscape of HPV vaccination and changes in HPV genotype prevalence, are also discussed. New guidelines and further research are necessary to elucidate the long-term effects of integrating dual staining into screening protocols.
... Digital pathology is a general term that refers to the process of digitizing histopathology, immunohistochemistry or cytology slides using whole-slide scanners, along with the interpretation, management, and analysis of these digitized whole-slide images (WSIs) using computational approaches (computational pathology) [95,96]. The digital pathology slides can be stored in a centralized repository, enabling remote access for manual review by a pathologist or automated evaluation by a data algorithm. ...
Advancements in data acquisition and computational methods are generating a large amount of heterogeneous biomedical data from diagnostic domains such as clinical imaging, pathology, and next-generation sequencing (NGS), which help characterize individual differences in patients. However, this information needs to be available and suitable to promote and support scientific research and technological development, supporting the effective adoption of the precision medicine approach in clinical practice. Digital biobanks can catalyze this process, facilitating the sharing of curated and standardized imaging data, clinical, pathological and molecular data, crucial to enable the development of a comprehensive and personalized data-driven diagnostic approach in disease management and fostering the development of computational predictive models. This work aims to frame this perspective, first by evaluating the state of standardization of individual diagnostic domains and then by identifying challenges and proposing a possible solution towards an integrative approach that can guarantee the suitability of information that can be shared through a digital biobank. Our analysis of the state of the art shows the presence and use of reference standards in biobanks and, generally, digital repositories for each specific domain. Despite this, standardization to guarantee the integration and reproducibility of the numerical descriptors generated by each domain, e.g. radiomic, pathomic and -omic features, is still an open challenge. Based on specific use cases and scenarios, an integration model, based on the JSON format, is proposed that can help address this problem. Ultimately, this work shows how, with specific standardization and promotion efforts, the digital biobank model can become an enabling technology for the comprehensive study of diseases and the effective development of data-driven technologies at the service of precision medicine.
