[Artificial intelligence for pathology-how, where, and why?] / Künstliche Intelligenz in der Pathologie ­ wie, wo und warum?

Artificial intelligence promises many innovations and simplifications in pathology, but also raises just as many questions and uncertainties. In this article, we provide a brief overview of the current status, the goals already achieved by existing algorithms, and the remaining challenges.

Complete digital pathology transition: A large multi-center experience.

INTRODUCTION: Transitioning from glass slide pathology to digital pathology for primary diagnostics requires an appropriate laboratory information system, an image management system, and slide scanners; it also reinforces the need for sophisticated pathology informatics including synoptic reporting. Previous reports have discussed the transition itself and relevant considerations for it, but not the selection criteria and considerations for the infrastructure. OBJECTIVE: To describe the process used to evaluate slide scanners, image management systems, and synoptic reporting systems for a large multisite institution. METHODS: Six network hospitals evaluated six slide scanners, three image management systems, and three synoptic reporting systems. Scanners were evaluated based on the quality of image, speed, ease of operation, and special capabilities (including z-stacking, fluorescence and others). Image management and synoptic reporting systems were evaluated for their ease of use and capacity. RESULTS: Among the scanners evaluated, the Leica GT450 produced the highest quality images, while the 3DHistech Pannoramic provided fluorescence and superior z-stacking. The newest generation of scanners, released relatively recently, performed better than slightly older scanners from major manufacturers Although the Olympus VS200 was not fully vetted due to not meeting all inclusion criteria, it is discussed herein due to its exceptional versatility. For Image Management Software, the authors believe that Sectra is, at the time of writing the best developed option, but this could change in the very near future as other systems improve their capabilities. All synoptic reporting systems performed impressively. CONCLUSIONS: Specifics regarding quality and abilities of different components will change rapidly with time, but large pathology practices considering such a transition should be aware of the issues discussed and evaluate the most current generation to arrive at appropriate conclusions.

Authentic Pathology Specimen Reception: A Valuable Resource for Developing Biomedical Science Student Competencies and Employability.

Background/Introduction: The pathology specimen reception is fundamental to the services provided by Biomedical Science laboratories worldwide. To ensure patient safety and that samples are of adequate quality to send for analysis, prospective Biomedical Scientists should have a robust knowledge of the processes involved and the acceptance criteria of the pathology specimen reception. This knowledge has been highlighted by employers as a current gap in Biomedical Science graduates and therefore needs to be addressed within higher education settings. To do this, this study aimed to 1) design a practical session to simulate the key processes of the pathology specimen reception and 2) to understand Biomedical Science students' opinions on these activities and the development of transferable skills required for post-graduate employment. Methods: The practical session was designed based on industrial requirements and academic knowledge of student skill sets to ensure suitability. Qualitative information regarding participant demographics and career interests was acquired through open-answer or multiple-choice questions. Quantitative student feedback was acquired via questionnaires utilising a 5-point Likert scale (n = 77). Results: The scenario-based practical session provided students with a positive learning experience with 98.7% of participants enjoying the session, with 87.0% stating they learned a lot by completing the session. It was also identified that participants preferred this style of learning to that of conventional higher education teaching modalities with 97.4% stating they would prefer simulated employment focussed scenarios embedded into the curriculum more often. The majority of participants also thought this session was helpful for the development of their key transferrable skills including teamworking, communication, and confidence. When stratified based on demographic data, there was minimal difference between cohorts and in the majority of cases, those participants from non-traditional university entry backgrounds had a more positive experience and better transferable skill development following the completion of this style of learning experience. Conclusion: This study highlights simulation-based learning as a tool to develop core Biomedical Science knowledge, build student graduate capital, and ensure the preparedness of students for post-graduation employment.

How to use AI in pathology.

AI plays an important role in pathology, both in clinical practice supporting pathologists in their daily work, and in research discovering novel biomarkers for improved patient care. Still, AI is in its starting phase, and many pathology labs still need to transition to a digital workflow to be able to enjoy the benefits of AI. In this perspective, we explain the major benefits of AI in pathology, highlight key requirements that need to be met and example how to use it in a typical workflow.

High-Order Correlation-Guided Slide-Level Histology Retrieval With Self-Supervised Hashing.

Histopathological Whole Slide Images (WSIs) play a crucial role in cancer diagnosis. It is of significant importance for pathologists to search for images sharing similar content with the query WSI, especially in the case-based diagnosis. While slide-level retrieval could be more intuitive and practical in clinical applications, most methods are designed for patch-level retrieval. A few recently unsupervised slide-level methods only focus on integrating patch features directly, without perceiving slide-level information, and thus severely limits the performance of WSI retrieval. To tackle the issue, we propose a High-Order Correlation-Guided Self-Supervised Hashing-Encoding Retrieval (HSHR) method. Specifically, we train an attention-based hash encoder with slide-level representation in a self-supervised manner, enabling it to generate more representative slide-level hash codes of cluster centers and assign weights for each. These optimized and weighted codes are leveraged to establish a similarity-based hypergraph, in which a hypergraph-guided retrieval module is adopted to explore high-order correlations in the multi-pairwise manifold to conduct WSI retrieval. Extensive experiments on multiple TCGA datasets with over 24,000 WSIs spanning 30 cancer subtypes demonstrate that HSHR achieves state-of-the-art performance compared with other unsupervised histology WSI retrieval methods.

Toxicologic Pathology Forum: Opinion on Performing Good Laboratory Practice Histopathology Evaluation for Nonclinical Toxicity Studies in a Remote Location.

Toxicologic/veterinary pathologists are working remotely from Good Laboratory Practice (GLP) test facilities (TFs) in increasing numbers, most commonly in home-office settings. A study pathologist (SP) generating data on GLP-compliant nonclinical studies must be keenly aware of applicable national GLP regulations and comply with TF and protocol requirements. This Toxicological Pathology Forum Opinion Piece will summarize primary areas of emphasis for the SP generating GLP data using glass slides. Peer review and digital review of whole slide images are out of scope for this opinion piece. Key GLP considerations for primary pathology on glass slides are discussed with respect to SP location and employment status, including pathologist qualifications, specimen management, facilities, equipment, archive, and quality assurance. Notable differences between national GLP regulations of the United States, the United Kingdom, Germany, the Netherlands, France, Ireland, Switzerland, Italy, and Israel are presented. With the understanding that each combination of location and employment is unique, the authors provide a general overview of considerations for successful remote GLP work.

Expansion microscopy: A revolution in diagnostic pathology.

Expansion microscopy (ExM) is a recent discovery in diagnostic pathology and microbiology that promotes a physical magnification of the tissue specimen instead of optical magnification. It not only improves the resolution of the specimen but also enhances the diagnostic precision, and permits nanoscale imaging of the specimen. Optical microscopy is routinely used in histopathology and microbiology for ages. Due to its simplicity, compatibility with different types of specimens, and ease of operation it is accepted by pathologists. ExM and its variants have been widely tested in different types of tissue specimens and microbiological specimens and yielded brilliant results.

Nationwide implementation of a multifaceted tailored strategy to improve uptake of standardized structured reporting in pathology: an effect and process evaluation.

BACKGROUND: Implementation strategies are aimed at improving guideline adherence. Both effect and process evaluations are conducted to provide insights into the success or failure of these strategies. In our study, we evaluate the nationwide implementation of standardized structured reporting (SSR) in pathology. METHODS: An interrupted time series analysis was conducted to evaluate the effect of a previously developed implementation strategy, which consisted of various digitally available elements, on SSR in pathology laboratories. A segmented regression analysis was performed to analyze the change in mean SSR percentages directly after the strategy introduction for pathology reporting and specific subcategories. In addition, we analyzed the change in trend in the weekly percentages after strategy introduction, also for subgroups of tumor groups, retrieval methods, and type of laboratory. The change in SSR use after the strategy introduction was determined for all pathology laboratories. We further conducted a process evaluation in which the exposure to the strategy elements was determined. Experiences of the users with all strategy elements and the remaining barriers and potential strategy elements were evaluated through an eSurvey. We also tested whether exposure to a specific element and a combination of elements resulted in a higher uptake of SSR after strategy introduction. RESULTS: There was a significant increase in an average use of SSR after the strategy introduction for reporting of gastrointestinal (p=.018) and urological (p=.003) oncological diagnoses. A significant increase was present for all oncological resections as a group (p=.007). Thirty-three out of 42 pathology laboratories increased SSR use after the strategy introduction. The "Feedback button", an option within the templates for SSR to provide feedback to the provider and one of the elements of the implementation strategy, was most frequently used by the SSR users, and effectiveness results showed that it increased average SSR use after the strategy introduction. Barriers were still present for SSR implementation. CONCLUSIONS: Nationwide SSR implementation improved for specific tumor groups and retrieval methods. The next step will be to further improve the use of SSR and, simultaneously, to further develop potential benefits of high SSR use, focusing on re-using discrete pathology data. In this way, we can facilitate proper treatment decisions in oncology.

An autofocus algorithm considering wavelength changes for large scale microscopic hyperspectral pathological imaging system.

Microscopic hyperspectral imaging technology has been widely used to acquire pathological information of tissue sections. Autofocus is one of the most important steps in microscopic hyperspectral imaging systems to capture large scale or even whole slide images of pathological slides with high quality and high speed. However, there are quite few autofocus algorithm put forward for the microscopic hyperspectral imaging system. Therefore, this article proposes a Laplace operator based autofocus algorithm for microscopic hyperspectral imaging system which takes the influence of wavelength changes into consideration. Through the proposed algorithm, the focal length for each wavelength can be adjusted automatically to ensure that each single band image can be autofocused precisely with adaptive image sharpness evaluation method. In addition, to increase the capture speed, the relationship of wavelength and focal length is derived and the focal offsets among different single band images are calculated for pre-focusing. We have employed the proposed method on our own datasets and the experimental results show that it can capture large-scale microscopic hyperspectral pathology images with high precise.

Non-invasive scoring of cellular atypia in keratinocyte cancers in 3D LC-OCT images using Deep Learning.

Diagnosis based on histopathology for skin cancer detection is today's gold standard and relies on the presence or absence of biomarkers and cellular atypia. However it suffers drawbacks: it requires a strong expertise and is time-consuming. Moreover the notion of atypia or dysplasia of the visible cells used for diagnosis is very subjective, with poor inter-rater agreement reported in the literature. Lastly, histology requires a biopsy which is an invasive procedure and only captures a small sample of the lesion, which is insufficient in the context of large fields of cancerization. Here we demonstrate that the notion of cellular atypia can be objectively defined and quantified with a non-invasive in-vivo approach in three dimensions (3D). A Deep Learning (DL) algorithm is trained to segment keratinocyte (KC) nuclei from Line-field Confocal Optical Coherence Tomography (LC-OCT) 3D images. Based on these segmentations, a series of quantitative, reproducible and biologically relevant metrics is derived to describe KC nuclei individually. We show that, using those metrics, simple and more complex definitions of atypia can be derived to discriminate between healthy and pathological skins, achieving Area Under the ROC Curve (AUC) scores superior than 0.965, largely outperforming medical experts on the same task with an AUC of 0.766. All together, our approach and findings open the door to a precise quantitative monitoring of skin lesions and treatments, offering a promising non-invasive tool for clinical studies to demonstrate the effects of a treatment and for clinicians to assess the severity of a lesion and follow the evolution of pre-cancerous lesions over time.

Deep-ReAP: Deep Representations And Partial label learning for Multi-pathology Classification.

Automated detection of pathology in images with multiple pathologies is one of the most challenging problems in medical diagnostics. The primary hurdles for automated systems include data imbalance across pathology categories and structural variations in pathological manifestations across patients. In this work, we present a novel method to detect a minimal dataset to train deep learning models that classify and explain multiple pathologies through the deep representations. We implement partial label learning with 1% false labels to identify the under-fit pathological categories that need further training followed by fine-tuning the deep representations. The proposed method identifies 54% of available training images as optimal for explainable classification of upto 7 pathological categories that can co-exist in 36 various combinations in retinal images, with overall precision/recall/Fß scores of 57%/87%/80%. Thus, the proposed method can lead to explainable inferencing for multi-label medical image data sets.

Multi-pathology detection and lesion localization in WCE videos by using the instance segmentation approach.

The majority of current systems for automatic diagnosis considers the detection of a unique and previously known pathology. Considering specifically the diagnosis of lesions in the small bowel using endoscopic capsule images, very few consider the possible existence of more than one pathology and when they do, they are mainly detection based systems therefore unable to localize the suspected lesions. Such systems do not fully satisfy the medical community, that in fact needs a system that detects any pathology and eventually more than one, when they coexist. In addition, besides the diagnostic capability of these systems, localizing the lesions in the image has been of great interest to the medical community, mainly for training medical personnel purposes. So, nowadays, the inclusion of the lesion location in automatic diagnostic systems is practically mandatory. Multi-pathology detection can be seen as a multi-object detection task and as each frame can contain different instances of the same lesion, instance segmentation seems to be appropriate for the purpose. Consequently, we argue that a multi-pathology system benefits from using the instance segmentation approach, since classification and segmentation modules are both required complementing each other in lesion detection and localization. According to our best knowledge such a system does not yet exist for the detection of WCE pathologies. This paper proposes a multi-pathology system that can be applied to WCE images, which uses the Mask Improved RCNN (MI-RCNN), a new mask subnet scheme which has shown to significantly improve mask predictions of the high performing state-of-the-art Mask-RCNN and PANet systems. A novel training strategy based on the second momentum is also proposed for the first time for training Mask-RCNN and PANet based systems. These approaches were tested using the public database KID, and the included pathologies were bleeding, angioectasias, polyps and inflammatory lesions. Experimental results show significant improvements for the proposed versions, reaching increases of almost 7% over the PANet model when the new proposed training approach was employed.

Inter-laboratory performance of ICE histopathology scoring to identify UN GHS Category 1 surfactants and non-extreme pH detergents.

The inter-laboratory performance of Isolated Chicken Eye (ICE) histopathology scoring was assessed for predicting EU CLP/UN GHS Cat. 1 surfactants. Furthermore, the predictive capacity of ICE histopathology was evaluated for the combined dataset of surfactants and existing data for non-extreme pH (2 < pH < 11.5) detergents. Use of ICE histopathology led to increased sensitivity compared to the ICE test method alone for surfactants. When combined with the existing dataset of detergents, use of histopathology in addition to the standard ICE test method decreased the false negative rates from 64% (14/22) to 27% (6/22); increased accuracy from 53% (16/30) to 77% (23/30); and led to acceptable level of false positives (from 0/8 to 1/8 (12.5%). Moreover, good reproducibility of ICE histopathology predictions conducted on the same slides was found between pathologists and peer-reviewers from three independent laboratories (10/12 or 83%) and over time. Use of ICE histopathology was therefore found suitable to predict EU CLP/UN GHS Cat. 1 surfactants and non-extreme pH detergents. In addition, appropriate reproducibility of ICE histopathology was found, provided that i) an internal peer-review system was in place; ii) original slides were assessed to enable evaluation of three dimensional effects; and iii) appropriate training and proficiency appraisal were conducted.

The medical student's guide to pathology residency, fellowships, and careers.

Recent studies have shown that relatively few MD, DO, and underrepresented in medicine (URM) students and physicians are matching into pathology residency in the United States (US). In the 2021 Main Residency Match, just 33.6% of filled pathology residency positions were taken by senior year students at US allopathic medical schools. This has been attributed to the fact that pathology is not a required rotation in most US medical schools, pathology is often taught in an integrated curriculum in the US where is does not stand out as a distinct field, and because the COVID-19 pandemic led to a suspension of in-person pathology rotations and electives. Ultimately, many US medical students fail to consider pathology as a career pathway. The objective of this article is to provide medical students with basic information, in the form of frequently asked questions (FAQs), about pathology training and career opportunities. This was accomplished by forming a team of MD and DO pathology attendings, pathology trainees, and a medical student from multiple institutions to create a pathology guide for medical students. This guide includes information about post-sophomore fellowships, 5 major pathology residency tracks, more than 20 fellowship pathways, and allopathic and osteopathic board examinations. This guide also contains photographs and descriptions of major pathology sub-specialties, including the daily and on-call duties and responsibilities of pathology residents. The exciting future of pathology is also discussed. This guide supports the agenda of the College of American Pathologists' (CAP) Pathologist Pipeline Initiative to improve student recruitment into pathology.

Digital pathology as a solution for working from home.

No abstract available.