Explainable AI (xAI) for Anatomic Pathology

Explainable AI for Medical Image Analysis in Medical Cyber-Physical Systems: Enhancing Transparency and Trustworthiness of IoMT

This study explores the application of explainable artificial intelligence (XAI) in the context of medical image analysis within medical cyber-physical systems (MCPS) to enhance transparency and trustworthiness. Meanwhile, this study proposes an explainable framework that integrates machine learning and knowledge reasoning. The explainability of the model is realized when the framework evolution target feature results and reasoning results are the same and are relatively reliable. However, using these technologies also presents new challenges, including the need to ensure the security and privacy of patient data from Internet of Medical Things (IoMT). Therefore, attack detection is an essential aspect of MCPS security. For the MCPS model with only sensor attacks, the necessary and sufficient conditions for detecting attacks are given based on the definition of sparse observability. The corresponding attack detector and state estimator are designed by assuming that some IoMT sensors are under protection. It is expounded that the IoMT sensors under protection play an important role in improving the efficiency of attack detection and state estimation. The experimental results show that the XAI in the context of medical image analysis within MCPS improves the accuracy of lesion classification, effectively removes low-quality medical images, and realizes the explainability of recognition results. This helps doctors understand the logic of the system's decision-making and can choose whether to trust the results based on the explanation given by the framework.

Benchmarking pathology foundation models for non-neoplastic pathology in the placenta

Machine learning (ML) applications within diagnostic histopathology have been extremely successful. While many successful models have been built using general-purpose models trained largely on everyday objects, there is a recent trend toward pathology-specific foundation models, trained using histopathology images. Pathology foundation models show strong performance on cancer detection and subtyping, grading, and predicting molecular diagnoses. However, we have noticed lacunae in the testing of foundation models. Nearly all the benchmarks used to test them are focused on cancer. Neoplasia is an important pathologic mechanism and key concern in much of clinical pathology, but it represents one of many pathologic bases of disease. Non-neoplastic pathology dominates findings in the placenta, a critical organ in human development, as well as a specimen commonly encountered in clinical practice. Very little to none of the data used in training pathology foundation models is placenta. Thus, placental pathology is doubly out of distribution, representing a useful challenge for foundation models. We developed benchmarks for estimation of gestational age, classifying normal tissue, identifying inflammation in the umbilical cord and membranes, and in classification of macroscopic lesions including villous infarction, intervillous thrombus, and perivillous fibrin deposition. We tested 5 pathology foundation models and 4 non-pathology models for each benchmark in tasks including zero-shot K-nearest neighbor classification and regression, content-based image retrieval, supervised regression, and whole-slide attention-based multiple instance learning. In each task, the best performing model was a pathology foundation model. However, the gap between pathology and non-pathology models was diminished in tasks related to inflammation or those in which a supervised task was performed using model embeddings. Performance was comparable among pathology foundation models. Among non-pathology models, ResNet consistently performed worse, while models from the present decade showed better performance. Future work could examine the impact of incorporating placental data into foundation model training.

Toward efficient slide-level grading of liver biopsy via explainable deep learning framework

In the context of chronic liver diseases, where variability in progression necessitates early and precise diagnosis, this study addresses the limitations of traditional histological analysis and the shortcomings of existing deep learning approaches. A novel patch-level classification model employing multi-scale feature extraction and fusion was developed to enhance the grading accuracy and interpretability of liver biopsies, analyzing 1322 cases across various staining methods. The study also introduces a slide-level aggregation framework, comparing different diagnostic models, to efficiently integrate local histological information. Results from extensive validation show that the slide-level model consistently achieved high F1 scores, notably 0.9 for inflammatory activity and steatosis, and demonstrated rapid diagnostic capabilities with less than one minute per slide on average. The patch-level model also performed well, with an F1 score of 0.64 for ballooning and 0.99 for other indicators, and proved transferable to public datasets. The conclusion drawn is that the proposed analytical framework offers a reliable basis for the diagnosis and treatment of chronic liver diseases, with the added benefit of robust interpretability, suggesting its practical utility in clinical settings.

Ulcerative Colitis, LAIR1 and TOX2 Expression, and Colorectal Cancer Deep Learning Image Classification Using Convolutional Neural Networks

BACKGROUND

Ulcerative colitis is a chronic inflammatory bowel disease of the colon mucosa associated with a higher risk of colorectal cancer.

OBJECTIVE

This study classified hematoxylin and eosin (H&E) histological images of ulcerative colitis, normal colon, and colorectal cancer using artificial intelligence (deep learning).

METHODS

A convolutional neural network (CNN) was designed and trained to classify the three types of diagnosis, including 35 cases of ulcerative colitis (n = 9281 patches), 21 colon control (n = 12,246), and 18 colorectal cancer (n = 63,725). The data were partitioned into training (70%) and validation sets (10%) for training the network, and a test set (20%) to test the performance on the new data. The CNNs included transfer learning from ResNet-18, and a comparison with other CNN models was performed. Explainable artificial intelligence for computer vision was used with the Grad-CAM technique, and additional LAIR1 and TOX2 immunohistochemistry was performed in ulcerative colitis to analyze the immune microenvironment.

RESULTS

Conventional clinicopathological analysis showed that steroid-requiring ulcerative colitis was characterized by higher endoscopic Baron and histologic Geboes scores and LAIR1 expression in the lamina propria, but lower TOX2 expression in isolated lymphoid follicles (all p values < 0.05) compared to mesalazine-responsive ulcerative colitis. The CNN classification accuracy was 99.1% for ulcerative colitis, 99.8% for colorectal cancer, and 99.1% for colon control. The Grad-CAM heatmap confirmed which regions of the images were the most important. The CNNs also differentiated between steroid-requiring and mesalazine-responsive ulcerative colitis based on H&E, LAIR1, and TOX2 staining. Additional classification of 10 new cases of colorectal cancer (adenocarcinoma) were correctly classified.

CONCLUSIONS

CNNs are especially suited for image classification in conditions such as ulcerative colitis and colorectal cancer; LAIR1 and TOX2 are relevant immuno-oncology markers in ulcerative colitis.

Seeing the random forest through the decision trees. Supporting learning health systems from histopathology with machine learning models: Challenges and opportunities

This paper discusses some overlooked challenges faced when working with machine learning models for histopathology and presents a novel opportunity to support "Learning Health Systems" with them. Initially, the authors elaborate on these challenges after separating them according to their mitigation strategies: those that need innovative approaches, time, or future technological capabilities and those that require a conceptual reappraisal from a critical perspective. Then, a novel opportunity to support "Learning Health Systems" by integrating hidden information extracted by ML models from digitalized histopathology slides with other healthcare big data is presented.

Artificial Intelligence Applications in Lymphoma Diagnosis and Management: Opportunities, Challenges, and Future Directions

Lymphoma, a heterogeneous group of blood cancers, presents significant diagnostic and therapeutic challenges due to its complex subtypes and variable clinical outcomes. Artificial intelligence (AI) has emerged as a promising tool to enhance the accuracy and efficiency of lymphoma pathology. This review explores the potential of AI in lymphoma diagnosis, classification, prognosis prediction, and treatment planning, as well as addressing the challenges and future directions in this rapidly evolving field.

Current role of cytopathology in the molecular and computational era: The perspective of young pathologists

Cytopathology represents a well established diagnostic approach because of its limited cost, reliability, and minimal invasiveness with respect to other methodologies. The evolving complexity of the different classifications systems and the implementation of ancillary techniques to refine the diagnosis is progressively helping in the risk of malignancy stratification, and the adoption of next-generation sequencing techniques contributes to enrich this valuable tool with predictive information, which is always more essential in the tailored medicine era. The recent introduction of digital and computational pathology is further boosting the potentialities of cytopathology, aiding in the interpretation of samples to improve the cost effectiveness of large screening programs and the diagnostic efficiency within intermediate/atypical categories. Moreover, the adoption of artificial intelligence tools is promising to complement molecular investigations, representing a stimulating perspective in the cytopathology field. In this work, the authors tried to summarize the multifaceted nature of this complex and evolving field of pathology, synthesizing the most recent advances and providing the young pathologists' perspective on this fascinating world.

Recommendations for Performance Evaluation of Machine Learning in Pathology: A Concept Paper From the College of American Pathologists

CONTEXT.—

Machine learning applications in the pathology clinical domain are emerging rapidly. As decision support systems continue to mature, laboratories will increasingly need guidance to evaluate their performance in clinical practice. Currently there are no formal guidelines to assist pathology laboratories in verification and/or validation of such systems. These recommendations are being proposed for the evaluation of machine learning systems in the clinical practice of pathology.

OBJECTIVE.—

To propose recommendations for performance evaluation of in vitro diagnostic tests on patient samples that incorporate machine learning as part of the preanalytical, analytical, or postanalytical phases of the laboratory workflow. Topics described include considerations for machine learning model evaluation including risk assessment, predeployment requirements, data sourcing and curation, verification and validation, change control management, human-computer interaction, practitioner training, and competency evaluation.

DATA SOURCES.—

An expert panel performed a review of the literature, Clinical and Laboratory Standards Institute guidance, and laboratory and government regulatory frameworks.

CONCLUSIONS.—

Review of the literature and existing documents enabled the development of proposed recommendations. This white paper pertains to performance evaluation of machine learning systems intended to be implemented for clinical patient testing. Further studies with real-world clinical data are encouraged to support these proposed recommendations. Performance evaluation of machine learning models is critical to verification and/or validation of in vitro diagnostic tests using machine learning intended for clinical practice.

Utilizing radiomics for differential diagnosis of inverted papilloma and chronic rhinosinusitis with polyps based on unenhanced CT scans

To evaluate whether radiomics models based on unenhanced paranasal sinuses CT images could be a useful tool for differentiating inverted papilloma (IP) from chronic rhinosinusitis with polyps (CRSwNP). This retrospective study recruited 240 patients with CRSwNP and 106 patients with IP from three centers. 253 patients from Qilu Hospital were randomly divided into the training set (n = 151) and the internal validation set (n = 102) with a ratio of 6:4. 93 patients from the other two centers were used as the external validation set. The patients with the unilateral disease (n = 115) from Qilu Hospital were selected to further develop a subgroup analysis. Lesion segmentation was manually delineated in CT images. Least absolute shrinkage and selection operator algorithm was performed for feature reduction and selection. Decision tree, support vector machine, random forest, and adaptive boosting regressor were employed to establish the differential diagnosis models. 43 radiomic features were selected for modeling. Among the models, RF achieved the best results, with an AUC of 0.998, 0.943, and 0.934 in the training set, the internal validation set, and the external validation set, respectively. In the subgroup analysis, RF achieved an AUC of 0.999 in the training set and 0.963 in the internal validation set. The proposed radiomics models offered a non-invasion and accurate differential approach between IP and CRSwNP and has some significance in guiding clinicians determining the best treatment plans, as well as predicting the prognosis.

The AI ethics of digital COVID-19 diagnosis and their legal, medical, technological, and operational managerial implications

The COVID-19 pandemic has given rise to a broad range of research from fields alongside and beyond the core concerns of infectiology, epidemiology, and immunology. One significant subset of this work centers on machine learning-based approaches to supporting medical decision-making around COVID-19 diagnosis. To date, various challenges, including IT issues, have meant that, notwithstanding this strand of research on digital diagnosis of COVID-19, the actual use of these methods in medical facilities remains incipient at best, despite their potential to relieve pressure on scarce medical resources, prevent instances of infection, and help manage the difficulties and unpredictabilities surrounding the emergence of new mutations. The reasons behind this research-application gap are manifold and may imply an interdisciplinary dimension. We argue that the discipline of AI ethics can provide a framework for interdisciplinary discussion and create a roadmap for the application of digital COVID-19 diagnosis, taking into account all disciplinary stakeholders involved. This article proposes such an ethical framework for the practical use of digital COVID-19 diagnosis, considering legal, medical, operational managerial, and technological aspects of the issue in accordance with our diverse research backgrounds and noting the potential of the approach we set out here to guide future research.

HistoEM: A Pathologist-Guided and Explainable Workflow Using Histogram Embedding for Gland Classification

Pathologists have, over several decades, developed criteria for diagnosing and grading prostate cancer. However, this knowledge has not, so far, been included in the design of convolutional neural networks (CNN) for prostate cancer detection and grading. Further, it is not known whether the features learned by machine-learning algorithms coincide with diagnostic features used by pathologists. We propose a framework that enforces algorithms to learn the cellular and subcellular differences between benign and cancerous prostate glands in digital slides from hematoxylin and eosin-stained tissue sections. After accurate gland segmentation and exclusion of the stroma, the central component of the pipeline, named HistoEM, utilizes a histogram embedding of features from the latent space of the CNN encoder. Each gland is represented by 128 feature-wise histograms that provide the input into a second network for benign vs cancer classification of the whole gland. Cancer glands are further processed by a U-Net structured network to separate low-grade from high-grade cancer. Our model demonstrates similar performance compared with other state-of-the-art prostate cancer grading models with gland-level resolution. To understand the features learned by HistoEM, we first rank features based on the distance between benign and cancer histograms and visualize the tissue origins of the 2 most important features. A heatmap of pixel activation by each feature is generated using Grad-CAM and overlaid on nuclear segmentation outlines. We conclude that HistoEM, similar to pathologists, uses nuclear features for the detection of prostate cancer. Altogether, this novel approach can be broadly deployed to visualize computer-learned features in histopathology images.

Why do errors arise in artificial intelligence diagnostic tools in histopathology and how can we minimize them?

Artificial intelligence (AI)-based diagnostic tools can offer numerous benefits to the field of histopathology, including improved diagnostic accuracy, efficiency and productivity. As a result, such tools are likely to have an increasing role in routine practice. However, all AI tools are prone to errors, and these AI-associated errors have been identified as a major risk in the introduction of AI into healthcare. The errors made by AI tools are different, in terms of both cause and nature, to the errors made by human pathologists. As highlighted by the National Institute for Health and Care Excellence, it is imperative that practising pathologists understand the potential limitations of AI tools, including the errors made. Pathologists are in a unique position to be gatekeepers of AI tool use, maximizing patient benefit while minimizing harm. Furthermore, their pathological knowledge is essential to understanding when, and why, errors have occurred and so to developing safer future algorithms. This paper summarises the literature on errors made by AI diagnostic tools in histopathology. These include erroneous errors, data concerns (data bias, hidden stratification, data imbalances, distributional shift, and lack of generalisability), reinforcement of outdated practices, unsafe failure mode, automation bias, and insensitivity to impact. Methods to reduce errors in both tool design and clinical use are discussed, and the practical roles for pathologists in error minimisation are highlighted. This aims to inform and empower pathologists to move safely through this seismic change in practice and help ensure that novel AI tools are adopted safely.

Built to Last? Reproducibility and Reusability of Deep Learning Algorithms in Computational Pathology

Recent progress in computational pathology has been driven by deep learning. While code and data availability are essential to reproduce findings from preceding publications, ensuring a deep learning model's reusability is more challenging. For that, the codebase should be well-documented and easy to integrate into existing workflows and models should be robust toward noise and generalizable toward data from different sources. Strikingly, only a few computational pathology algorithms have been reused by other researchers so far, let alone employed in a clinical setting. To assess the current state of reproducibility and reusability of computational pathology algorithms, we evaluated peer-reviewed articles available in PubMed, published between January 2019 and March 2021, in 5 use cases: stain normalization; tissue type segmentation; evaluation of cell-level features; genetic alteration prediction; and inference of grading, staging, and prognostic information. We compiled criteria for data and code availability and statistical result analysis and assessed them in 160 publications. We found that only one-quarter (41 of 160 publications) made code publicly available. Among these 41 studies, three-quarters (30 of 41) analyzed their results statistically, half of them (20 of 41) released their trained model weights, and approximately a third (16 of 41) used an independent cohort for evaluation. Our review is intended for both pathologists interested in deep learning and researchers applying algorithms to computational pathology challenges. We provide a detailed overview of publications with published code in the field, list reusable data handling tools, and provide criteria for reproducibility and reusability.

Deep Learning-Based Computational Cytopathologic Diagnosis of Metastatic Breast Carcinoma in Pleural Fluid

A Pleural effusion cytology is vital for treating metastatic breast cancer; however, concerns have arisen regarding the low accuracy and inter-observer variability in cytologic diagnosis. Although artificial intelligence-based image analysis has shown promise in cytopathology research, its application in diagnosing breast cancer in pleural fluid remains unexplored. To overcome these limitations, we evaluate the diagnostic accuracy of an artificial intelligence-based model using a large collection of cytopathological slides, to detect the malignant pleural effusion cytology associated with breast cancer. This study includes a total of 569 cytological slides of malignant pleural effusion of metastatic breast cancer from various institutions. We extracted 34,221 augmented image patches from whole-slide images and trained and validated a deep convolutional neural network model (DCNN) (Inception-ResNet-V2) with the images. Using this model, we classified 845 randomly selected patches, which were reviewed by three pathologists to compare their accuracy. The DCNN model outperforms the pathologists by demonstrating higher accuracy, sensitivity, and specificity compared to the pathologists (81.1% vs. 68.7%, 95.0% vs. 72.5%, and 98.6% vs. 88.9%, respectively). The pathologists reviewed the discordant cases of DCNN. After re-examination, the average accuracy, sensitivity, and specificity of the pathologists improved to 87.9, 80.2, and 95.7%, respectively. This study shows that DCNN can accurately diagnose malignant pleural effusion cytology in breast cancer and has the potential to support pathologists.

Artificial intelligence innovation in healthcare: Relevance of reporting guidelines for clinical translation from bench to bedside

Artificial intelligence (AI) and digital innovation are transforming healthcare. Technologies such as machine learning in image analysis, natural language processing in medical chatbots and electronic medical record extraction have the potential to improve screening, diagnostics and prognostication, leading to precision medicine and preventive health. However, it is crucial to ensure that AI research is conducted with scientific rigour to facilitate clinical implementation. Therefore, reporting guidelines have been developed to standardise and streamline the development and validation of AI technologies in health. This commentary proposes a structured approach to utilise these reporting guidelines for the translation of promising AI techniques from research and development into clinical translation, and eventual widespread implementation from bench to bedside.

Opportunities and Challenges with Artificial Intelligence in Genomics

The development of artificial intelligence and machine learning algorithms may allow for advances in patient care. There are existing and potential applications in cancer diagnosis and monitoring, identification of at-risk groups of individuals, classification of genetic variants, and even prediction of patient ancestry. This article provides an overview of some current and future applications of artificial intelligence in genomic medicine, in addition to discussing challenges and considerations when bringing these tools into clinical practice.

Spatial immunophenotyping of the tumour microenvironment in non-small cell lung cancer

INTRODUCTION

Immune cells in the tumour microenvironment are associated with prognosis and response to therapy. We aimed to comprehensively characterise the spatial immune phenotypes in the mutational and clinicopathological background of non-small cell lung cancer (NSCLC).

METHODS

We established a multiplexed fluorescence imaging pipeline to spatially quantify 13 immune cell subsets in 359 NSCLC cases: CD4 effector cells (CD4-Eff), CD4 regulatory cells (CD4-Treg), CD8 effector cells (CD8-Eff), CD8 regulatory cells (CD8-Treg), B-cells, natural killer cells, natural killer T-cells, M1 macrophages (M1), CD163+ myeloid cells (CD163), M2 macrophages (M2), immature dendritic cells (iDCs), mature dendritic cells (mDCs) and plasmacytoid dendritic cells (pDCs).

RESULTS

CD4-Eff cells, CD8-Eff cells and M1 macrophages were the most abundant immune cells invading the tumour cell compartment and indicated a patient group with a favourable prognosis in the cluster analysis. Likewise, single densities of lymphocytic subsets (CD4-Eff, CD4-Treg, CD8-Treg, B-cells and pDCs) were independently associated with longer survival. However, when these immune cells were located close to CD8-Treg cells, the favourable impact was attenuated. In the multivariable Cox regression model, including cell densities and distances, the densities of M1 and CD163 cells and distances between cells (CD8-Treg-B-cells, CD8-Eff-cancer cells and B-cells-CD4-Treg) demonstrated positive prognostic impact, whereas short M2-M1 distances were prognostically unfavourable.

CONCLUSION

We present a unique spatial profile of the in situ immune cell landscape in NSCLC as a publicly available data set. Cell densities and cell distances contribute independently to prognostic information on clinical outcomes, suggesting that spatial information is crucial for diagnostic use.

How, for whom, and in what contexts will artificial intelligence be adopted in pathology? A realist interview study

OBJECTIVE

There is increasing interest in using artificial intelligence (AI) in pathology to improve accuracy and efficiency. Studies of clinicians' perceptions of AI have found only moderate acceptability, suggesting further research is needed regarding integration into clinical practice. This study aimed to explore stakeholders' theories concerning how and in what contexts AI is likely to become integrated into pathology.

MATERIALS AND METHODS

A literature review provided tentative theories that were revised through a realist interview study with 20 pathologists and 5 pathology trainees. Questions sought to elicit whether, and in what ways, the tentative theories fitted with interviewees' perceptions and experiences. Analysis focused on identifying the contextual factors that may support or constrain uptake of AI in pathology.

RESULTS

Interviews highlighted the importance of trust in AI, with interviewees emphasizing evaluation and the opportunity for pathologists to become familiar with AI as means for establishing trust. Interviewees expressed a desire to be involved in design and implementation of AI tools, to ensure such tools address pressing needs, but needs vary by subspecialty. Workflow integration is desired but whether AI tools should work automatically will vary according to the task and the context.

CONCLUSIONS

It must not be assumed that AI tools that provide benefit in one subspecialty will provide benefit in others. Pathologists should be involved in the decision to introduce AI, with opportunity to assess strengths and weaknesses. Further research is needed concerning the evidence required to satisfy pathologists regarding the benefits of AI.

An update on computational pathology tools for genitourinary pathology practice: A review paper from the Genitourinary Pathology Society (GUPS)

Machine learning has been leveraged for image analysis applications throughout a multitude of subspecialties. This position paper provides a perspective on the evolutionary trajectory of practical deep learning tools for genitourinary pathology through evaluating the most recent iterations of such algorithmic devices. Deep learning tools for genitourinary pathology demonstrate potential to enhance prognostic and predictive capacity for tumor assessment including grading, staging, and subtype identification, yet limitations in data availability, regulation, and standardization have stymied their implementation.

Artificial intelligence applications in predicting the behavior of gastrointestinal cancers in pathology

Recent research has provided a wealth of data supporting the application of artificial intelligence (AI)-based applications in routine pathology practice. Indeed, it is clear that these methods can significantly support an accurate and rapid diagnosis by eliminating errors, increasing reliability, and improving workflow. In addition, the effectiveness of AI in the pathological evaluation of prognostic parameters associated with behavior, course, and treatment in many types of tumors has also been noted. Regarding gastrointestinal system (GIS) cancers, the contribution of AI methods to pathological diagnosis has been investigated in many studies. On the other hand, studies focusing on AI applications in evaluating parameters to determine tumor behavior are relatively few. For this purpose, the potential of AI models has been studied over a broad spectrum, from tumor subtyping to the identification of new digital biomarkers. The capacity of AI to infer genetic alterations of cancer tissues from digital slides has been demonstrated. Although current data suggest the merit of AI-based approaches in assessing tumor behavior in GIS cancers, a wide range of challenges still need to be solved, from laboratory infrastructure to improving the robustness of algorithms, before incorporating AI applications into real-life GIS pathology practice. This review aims to present data from AI applications in evaluating pathological parameters related to the behavior of GIS cancer with an overview of the opportunities and challenges encountered in implementing AI in pathology.

Deep Learning for Skin Melanocytic Tumors in Whole-Slide Images: A Systematic Review

The rise of Artificial Intelligence (AI) has shown promising performance as a support tool in clinical pathology workflows. In addition to the well-known interobserver variability between dermatopathologists, melanomas present a significant challenge in their histological interpretation. This study aims to analyze all previously published studies on whole-slide images of melanocytic tumors that rely on deep learning techniques for automatic image analysis. Embase, Pubmed, Web of Science, and Virtual Health Library were used to search for relevant studies for the systematic review, in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) checklist. Articles from 2015 to July 2022 were included, with an emphasis placed on the used artificial intelligence methods. Twenty-eight studies that fulfilled the inclusion criteria were grouped into four groups based on their clinical objectives, including pathologists versus deep learning models (n = 10), diagnostic prediction (n = 7); prognosis (n = 5), and histological features (n = 6). These were then analyzed to draw conclusions on the general parameters and conditions of AI in pathology, as well as the necessary factors for better performance in real scenarios.

HER2 classification in breast cancer cells: A new explainable machine learning application for immunohistochemistry

The immunohistochemical (IHC) evaluation of epidermal growth factor 2 (HER2) for the diagnosis of breast cancer is still qualitative with a high degree of inter-observer variability, and thus requires the incorporation of complementary techniques such as fluorescent in situ hybridization (FISH) to resolve the diagnosis. Implementing automatic algorithms to classify IHC biomarkers is crucial for typifying the tumor and deciding on therapy for each patient with better performance. The present study aims to demonstrate that, using an explainable Machine Learning (ML) model for the classification of HER2 photomicrographs, it is possible to determine criteria to improve the value of IHC analysis. We trained a logistic regression-based supervised ML model with 393 IHC microscopy images from 131 patients, to discriminate between upregulated and normal expression of the HER2 protein. Pathologists' diagnoses (IHC only) vs. the final diagnosis complemented with FISH (IHC + FISH) were used as training outputs. Basic performance metrics and receiver operating characteristic curve analysis were used together with an explainability algorithm based on Shapley Additive exPlanations (SHAP) values to understand training differences. The model could discriminate amplified IHC from normal expression with better performance when the training output was the IHC + FISH final diagnosis (IHC vs. IHC + FISH: area under the curve, 0.94 vs. 0.81). This may be explained by the increased analytical impact of the membrane distribution criteria over the global intensity of the signal, according to SHAP value interpretation. The classification model improved its performance when the training input was the final diagnosis, downplaying the weighting of the intensity of the IHC signal, suggesting that to improve pathological diagnosis before FISH consultation, it is necessary to emphasize subcellular patterns of staining.

Application of Artificial Intelligence in Pathology: Trends and Challenges

Given the recent success of artificial intelligence (AI) in computer vision applications, many pathologists anticipate that AI will be able to assist them in a variety of digital pathology tasks. Simultaneously, tremendous advancements in deep learning have enabled a synergy with artificial intelligence (AI), allowing for image-based diagnosis on the background of digital pathology. There are efforts for developing AI-based tools to save pathologists time and eliminate errors. Here, we describe the elements in the development of computational pathology (CPATH), its applicability to AI development, and the challenges it faces, such as algorithm validation and interpretability, computing systems, reimbursement, ethics, and regulations. Furthermore, we present an overview of novel AI-based approaches that could be integrated into pathology laboratory workflows.

Human and machine: Better at pathology together?

Meaningful integration of artificial intelligence (AI) will transform the application of "big data" for patient care, diagnosis, and research. In this issue of Cancer Cell, Chen et al. describe a transparent system to integrate histopathology and molecular data to predict outcomes and identify novel biomarkers in cancer.

Deep-learning prediction of amyloid deposition from early-phase amyloid positron emission tomography imaging

OBJECTIVE

While the use of biomarkers for the detection of early and preclinical Alzheimer's Disease has become essential, the need to wait for over an hour after injection to obtain sufficient image quality can be challenging for patients with suspected dementia and their caregivers. This study aimed to develop an image-based deep-learning technique to generate delayed uptake patterns of amyloid positron emission tomography (PET) images using only early-phase images obtained from 0-20 min after radiotracer injection.

METHODS

We prepared pairs of early and delayed [¹¹C]PiB dynamic images from 253 patients (cognitively normal n = 32, fronto-temporal dementia n = 39, mild cognitive impairment n = 19, Alzheimer's disease n = 163) as a training dataset. The neural network was trained with the early images as the input, and the output was the corresponding delayed image. A U-net convolutional neural network (CNN) and a conditional generative adversarial network (C-GAN) were used for the deep-learning architecture and the data augmentation methods, respectively. Then, an independent test data set consisting of early-phase amyloid PET images (n = 19) was used to generate corresponding delayed images using the trained network. Two nuclear medicine physicians interpreted the actual delayed images and predicted delayed images for amyloid positivity. In addition, the concordance of the actual delayed and predicted delayed images was assessed statistically.

RESULTS

The concordance of amyloid positivity between the actual versus AI-predicted delayed images was 79%(κ = 0.60) and 79% (κ = 0.59) for each physician, respectively. In addition, the physicians' agreement rate was at 89% (κ = 0.79) when the same image was interpreted. And, the actual versus AI-predicted delayed images were not readily distinguishable (correct answer rate, 55% and 47% for each physician, respectively). The statistical comparison of the actual versus the predicted delated images indicated that the peak signal-to-noise ratio (PSNR) was 21.8 dB ± 2.2 dB, and the structural similarity index (SSIM) was 0.45 ± 0.04.

CONCLUSION

This study demonstrates the feasibility of an image-based deep-learning framework to predict delayed patterns of Amyloid PET uptake using only the early phase images. This AI-based image generation method has the potential to reduce scan time for amyloid PET and increase the patient throughput, without sacrificing diagnostic accuracy for amyloid positivity.

What works where and how for uptake and impact of artificial intelligence in pathology: A review of theories for a realist evaluation (Preprint)

BACKGROUND

There is increasing interest in the use of artificial intelligence (AI) in pathology to increase accuracy and efficiency. To date, studies of clinicians' perceptions of AI have found only moderate acceptability, suggesting the need for further research regarding how to integrate it into clinical practice.

OBJECTIVE

The aim of the study was to determine contextual factors that may support or constrain the uptake of AI in pathology.

METHODS

To go beyond a simple listing of barriers and facilitators, we drew on the approach of realist evaluation and undertook a review of the literature to elicit stakeholders' theories of how, for whom, and in what circumstances AI can provide benefit in pathology. Searches were designed by an information specialist and peer-reviewed by a second information specialist. Searches were run on the arXiv.org repository, MEDLINE, and the Health Management Information Consortium, with additional searches undertaken on a range of websites to identify gray literature. In line with a realist approach, we also made use of relevant theory. Included documents were indexed in NVivo 12, using codes to capture different contexts, mechanisms, and outcomes that could affect the introduction of AI in pathology. Coded data were used to produce narrative summaries of each of the identified contexts, mechanisms, and outcomes, which were then translated into theories in the form of context-mechanism-outcome configurations.

RESULTS

A total of 101 relevant documents were identified. Our analysis indicates that the benefits that can be achieved will vary according to the size and nature of the pathology department's workload and the extent to which pathologists work collaboratively; the major perceived benefit for specialist centers is in reducing workload. For uptake of AI, pathologists' trust is essential. Existing theories suggest that if pathologists are able to "make sense" of AI, engage in the adoption process, receive support in adapting their work processes, and can identify potential benefits to its introduction, it is more likely to be accepted.

CONCLUSIONS

For uptake of AI in pathology, for all but the most simple quantitative tasks, measures will be required that either increase confidence in the system or provide users with an understanding of the performance of the system. For specialist centers, efforts should focus on reducing workload rather than increasing accuracy. Designers also need to give careful thought to usability and how AI is integrated into pathologists' workflow.

From What to Why, the Growing Need for a Focus Shift Toward Explainability of AI in Digital Pathology

While it is impossible to deny the performance gains achieved through the incorporation of deep learning (DL) and other artificial intelligence (AI)-based techniques in pathology, minimal work has been done to answer the crucial question of why these algorithms predict what they predict. Tracing back classification decisions to specific input features allows for the quick identification of model bias as well as providing additional information toward understanding underlying biological mechanisms. In digital pathology, increasing the explainability of AI models would have the largest and most immediate impact for the image classification task. In this review, we detail some considerations that should be made in order to develop models with a focus on explainability.

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.

Integrating artificial intelligence in pathology: a qualitative interview study of users' experiences and expectations

Recent progress in the development of artificial intelligence (AI) has sparked enthusiasm for its potential use in pathology. As pathology labs are currently starting to shift their focus towards AI implementation, a better understanding how AI tools can be optimally aligned with the medical and social context of pathology daily practice is urgently needed. Strikingly, studies often fail to mention the ways in which AI tools should be integrated in the decision-making processes of pathologists, nor do they address how this can be achieved in an ethically sound way. Moreover, the perspectives of pathologists and other professionals within pathology concerning the integration of AI within pathology remains an underreported topic. This article aims to fill this gap in the literature and presents the first in-depth interview study in which professionals' perspectives on the possibilities, conditions and prerequisites of AI integration in pathology are explicated. The results of this study have led to the formulation of three concrete recommendations to support AI integration, namely: (1) foster a pragmatic attitude toward AI development, (2) provide task-sensitive information and training to health care professionals working in pathology departments and (3) take time to reflect upon users' changing roles and responsibilities.

Artificial intelligence in pathological evaluation of gastrointestinal cancers

The integration of artificial intelligence (AI) has shown promising benefits in many fields of diagnostic histopathology, including for gastrointestinal cancers (GCs), such as tumor identification, classification, and prognosis prediction. In parallel, recent evidence suggests that AI may help reduce the workload in gastrointestinal pathology by automatically detecting tumor tissues and evaluating prognostic parameters. In addition, AI seems to be an attractive tool for biomarker/genetic alteration prediction in GC, as it can contain a massive amount of information from visual data that is complex and partially understandable by pathologists. From this point of view, it is suggested that advances in AI could lead to revolutionary changes in many fields of pathology. Unfortunately, these findings do not exclude the possibility that there are still many hurdles to overcome before AI applications can be safely and effectively applied in actual pathology practice. These include a broad spectrum of challenges from needs identification to cost-effectiveness. Therefore, unlike other disciplines of medicine, no histopathology-based AI application, including in GC, has ever been approved either by a regulatory authority or approved for public reimbursement. The purpose of this review is to present data related to the applications of AI in pathology practice in GC and present the challenges that need to be overcome for their implementation.

Overview of Explainable Artificial Intelligence for Prognostic and Health Management of Industrial Assets Based on Preferred Reporting Items for Systematic Reviews and Meta-Analyses

Surveys on explainable artificial intelligence (XAI) are related to biology, clinical trials, fintech management, medicine, neurorobotics, and psychology, among others. Prognostics and health management (PHM) is the discipline that links the studies of failure mechanisms to system lifecycle management. There is a need, which is still absent, to produce an analytical compilation of PHM-XAI works. In this paper, we use preferred reporting items for systematic reviews and meta-analyses (PRISMA) to present a state of the art on XAI applied to PHM of industrial assets. This work provides an overview of the trend of XAI in PHM and answers the question of accuracy versus explainability, considering the extent of human involvement, explanation assessment, and uncertainty quantification in this topic. Research articles associated with the subject, since 2015 to 2021, were selected from five databases following the PRISMA methodology, several of them related to sensors. The data were extracted from selected articles and examined obtaining diverse findings that were synthesized as follows. First, while the discipline is still young, the analysis indicates a growing acceptance of XAI in PHM. Second, XAI offers dual advantages, where it is assimilated as a tool to execute PHM tasks and explain diagnostic and anomaly detection activities, implying a real need for XAI in PHM. Third, the review shows that PHM-XAI papers provide interesting results, suggesting that the PHM performance is unaffected by the XAI. Fourth, human role, evaluation metrics, and uncertainty management are areas requiring further attention by the PHM community. Adequate assessment metrics to cater to PHM needs are requested. Finally, most case studies featured in the considered articles are based on real industrial data, and some of them are related to sensors, showing that the available PHM-XAI blends solve real-world challenges, increasing the confidence in the artificial intelligence models' adoption in the industry.

A Computer-Aided Diagnosis System and Thyroid Imaging Reporting and Data System for Dual Validation of Ultrasound-Guided Fine-Needle Aspiration of Indeterminate Thyroid Nodules

Purpose

The fully automatic AI-Sonic computer-aided design (CAD) system was employed for the detection and diagnosis of benign and malignant thyroid nodules. The aim of this study was to investigate the efficiency of the AI-Sonic CAD system with the use of a deep learning algorithm to improve the diagnostic accuracy of ultrasound-guided fine-needle aspiration (FNA).

Methods

A total of 138 thyroid nodules were collected from 124 patients and diagnosed by an expert, a novice, and the Thyroid Imaging Reporting and Data System (TI-RADS). Diagnostic efficiency and feasibility were compared among the expert, novice, and CAD system. The application of the CAD system to enhance the diagnostic efficiency of novices was assessed. Moreover, with the experience of the expert as the gold standard, the values of features detected by the CAD system were also analyzed. The efficiency of FNA was compared among the expert, novice, and CAD system to determine whether the CAD system is helpful for the management of FNA.

Result

In total, 56 malignant and 82 benign thyroid nodules were collected from the 124 patients (mean age, 46.4 ± 12.1 years; range, 12–70 years). The diagnostic area under the curve of the CAD system, expert, and novice were 0.919, 0.891, and 0.877, respectively (p < 0.05). In regard to feature detection, there was no significant differences in the margin and composition between the benign and malignant nodules (p > 0.05), while echogenicity and the existence of echogenic foci were of great significance (p < 0.05). For the recommendation of FNA, the results showed that the CAD system had better performance than the expert and novice (p < 0.05).

Conclusions

Precise diagnosis and recommendation of FNA are continuing hot topics for thyroid nodules. The CAD system based on deep learning had better accuracy and feasibility for the diagnosis of thyroid nodules, and was useful to avoid unnecessary FNA. The CAD system is potentially an effective auxiliary approach for diagnosis and asymptomatic screening, especially in developing areas.

In situ functional cell phenotyping reveals microdomain networks in colorectal cancer recurrence

Tumors are dynamic ecosystems comprising localized niches (microdomains), possessing distinct compositions and spatial configurations of cancer and non-cancer cell populations. Microdomain-specific network signaling coevolves with a continuum of cell states and functional plasticity associated with disease progression and therapeutic responses. We present LEAPH, an unsupervised machine learning algorithm for identifying cell phenotypes, which applies recursive steps of probabilistic clustering and spatial regularization to derive functional phenotypes (FPs) along a continuum. Combining LEAPH with pointwise mutual information and network biology analyses enables the discovery of outcome-associated microdomains visualized as distinct spatial configurations of heterogeneous FPs. Utilization of an immunofluorescence-based (51 biomarkers) image dataset of colorectal carcinoma primary tumors (n = 213) revealed microdomain-specific network dysregulation supporting cancer stem cell maintenance and immunosuppression that associated selectively with the recurrence phenotype. LEAPH enables an explainable artificial intelligence platform providing insights into pathophysiological mechanisms and novel drug targets to inform personalized therapeutic strategies.

Whole-slide imaging, tissue image analysis, and artificial intelligence in veterinary pathology: An updated introduction and review

Since whole-slide imaging has been commercially available for over 2 decades, digital pathology has become a constantly expanding aspect of the pathology profession that will continue to significantly impact how pathologists conduct their craft. While some aspects, such as whole-slide imaging for archiving, consulting, and teaching, have gained broader acceptance, other facets such as quantitative tissue image analysis and artificial intelligence-based assessments are still met with some reservations. While most vendors in this space have focused on diagnostic applications, that is, viewing one or few slides at a time, some are developing solutions tailored more specifically to the various aspects of veterinary pathology including updated diagnostic, discovery, and research applications. This has especially advanced the use of digital pathology in toxicologic pathology and drug development, for primary reads as well as peer reviews. It is crucial that pathologists gain a deeper understanding of digital pathology and tissue image analysis technology and their applications in order to fully use these tools in a way that enhances and improves the pathologist's assessment as well as work environment. This review focuses on an updated introduction to the basics of digital pathology and image analysis and introduces emerging topics around artificial intelligence and machine learning.

Comparative analysis of machine learning approaches to classify tumor mutation burden in lung adenocarcinoma using histopathology images

Both histologic subtypes and tumor mutation burden (TMB) represent important biomarkers in lung cancer, with implications for patient prognosis and treatment decisions. Typically, TMB is evaluated by comprehensive genomic profiling but this requires use of finite tissue specimens and costly, time-consuming laboratory processes. Histologic subtype classification represents an established component of lung adenocarcinoma histopathology, but can be challenging and is associated with substantial inter-pathologist variability. Here we developed a deep learning system to both classify histologic patterns in lung adenocarcinoma and predict TMB status using de-identified Hematoxylin and Eosin (H&E) stained whole slide images. We first trained a convolutional neural network to map histologic features across whole slide images of lung cancer resection specimens. On evaluation using an external data source, this model achieved patch-level area under the receiver operating characteristic curve (AUC) of 0.78–0.98 across nine histologic features. We then integrated the output of this model with clinico-demographic data to develop an interpretable model for TMB classification. The resulting end-to-end system was evaluated on 172 held out cases from TCGA, achieving an AUC of 0.71 (95% CI 0.63–0.80). The benefit of using histologic features in predicting TMB is highlighted by the significant improvement this approach offers over using the clinical features alone (AUC of 0.63 [95% CI 0.53–0.72], p = 0.002). Furthermore, we found that our histologic subtype-based approach achieved performance similar to that of a weakly supervised approach (AUC of 0.72 [95% CI 0.64–0.80]). Together these results underscore that incorporating histologic patterns in biomarker prediction for lung cancer provides informative signals, and that interpretable approaches utilizing these patterns perform comparably with less interpretable, weakly supervised approaches.

Requirements for implementation of artificial intelligence in the practice of gastrointestinal pathology

Tremendous advances in artificial intelligence (AI) in medical image analysis have been achieved in recent years. The integration of AI is expected to cause a revolution in various areas of medicine, including gastrointestinal (GI) pathology. Currently, deep learning algorithms have shown promising benefits in areas of diagnostic histopathology, such as tumor identification, classification, prognosis prediction, and biomarker/genetic alteration prediction. While AI cannot substitute pathologists, carefully constructed AI applications may increase workforce productivity and diagnostic accuracy in pathology practice. Regardless of these promising advances, unlike the areas of radiology or cardiology imaging, no histopathology-based AI application has been approved by a regulatory authority or for public reimbursement. Thus, implying that there are still some obstacles to be overcome before AI applications can be safely and effectively implemented in real-life pathology practice. The challenges have been identified at different stages of the development process, such as needs identification, data curation, model development, validation, regulation, modification of daily workflow, and cost-effectiveness balance. The aim of this review is to present challenges in the process of AI development, validation, and regulation that should be overcome for its implementation in real-life GI pathology practice.

Current Challenges and Future Opportunities for XAI in Machine Learning-Based Clinical Decision Support Systems: A Systematic Review

Machine Learning and Artificial Intelligence (AI) more broadly have great immediate and future potential for transforming almost all aspects of medicine. However, in many applications, even outside medicine, a lack of transparency in AI applications has become increasingly problematic. This is particularly pronounced where users need to interpret the output of AI systems. Explainable AI (XAI) provides a rationale that allows users to understand why a system has produced a given output. The output can then be interpreted within a given context. One area that is in great need of XAI is that of Clinical Decision Support Systems (CDSSs). These systems support medical practitioners in their clinic decision-making and in the absence of explainability may lead to issues of under or over-reliance. Providing explanations for how recommendations are arrived at will allow practitioners to make more nuanced, and in some cases, life-saving decisions. The need for XAI in CDSS, and the medical field in general, is amplified by the need for ethical and fair decision-making and the fact that AI trained with historical data can be a reinforcement agent of historical actions and biases that should be uncovered. We performed a systematic literature review of work to-date in the application of XAI in CDSS. Tabular data processing XAI-enabled systems are the most common, while XAI-enabled CDSS for text analysis are the least common in literature. There is more interest in developers for the provision of local explanations, while there was almost a balance between post-hoc and ante-hoc explanations, as well as between model-specific and model-agnostic techniques. Studies reported benefits of the use of XAI such as the fact that it could enhance decision confidence for clinicians, or generate the hypothesis about causality, which ultimately leads to increased trustworthiness and acceptability of the system and potential for its incorporation in the clinical workflow. However, we found an overall distinct lack of application of XAI in the context of CDSS and, in particular, a lack of user studies exploring the needs of clinicians. We propose some guidelines for the implementation of XAI in CDSS and explore some opportunities, challenges, and future research needs.

A Real-Time Prescriptive Solution for Explainable Cyber-Fraud Detection Within the iGaming Industry

This paper presents a real-time fully autonomous prescriptive solution for explainable cyber-fraud detection within the iGaming industry. We demonstrate how our solution facilitates the time-consuming task of player risk and fraud assessment through prescriptive analytics. Our tool leverages machine learning algorithms and advancements in the field of eXplainable AI to derive smarter predictions empowered by local interpretable explanations in real-time. Our best-performing pipeline was able to predict fraudulent behaviour with an average precision of 84.2% and an area under the receiver operating characteristics of 0.82 on our dataset. We also addressed the phenomenon of concept-drift and discussed our empirical and data-driven strategy for detecting and dealing with this problem. Finally, we cover how local interpretable explanations can help adopt a pro-active stance in fighting fraud.

PSSPNN: PatchShuffle Stochastic Pooling Neural Network for an Explainable Diagnosis of COVID-19 with Multiple-Way Data Augmentation

AIM

COVID-19 has caused large death tolls all over the world. Accurate diagnosis is of significant importance for early treatment.

METHODS

In this study, we proposed a novel PSSPNN model for classification between COVID-19, secondary pulmonary tuberculosis, community-captured pneumonia, and healthy subjects. PSSPNN entails five improvements: we first proposed the n-conv stochastic pooling module. Second, a novel stochastic pooling neural network was proposed. Third, PatchShuffle was introduced as a regularization term. Fourth, an improved multiple-way data augmentation was used. Fifth, Grad-CAM was utilized to interpret our AI model.

RESULTS

The 10 runs with random seed on the test set showed our algorithm achieved a microaveraged F1 score of 95.79%. Moreover, our method is better than nine state-of-the-art approaches.

CONCLUSION

This proposed PSSPNN will help assist radiologists to make diagnosis more quickly and accurately on COVID-19 cases.

Mini Review: The Last Mile-Opportunities and Challenges for Machine Learning in Digital Toxicologic Pathology

The 2019 manuscript by the Special Interest Group on Digital Pathology and Image Analysis of the Society of Toxicologic pathology suggested that a synergism between artificial intelligence (AI) and machine learning (ML) technologies and digital toxicologic pathology would improve the daily workflow and future impact of toxicologic pathologists globally. Now 2 years later, the authors of this review consider whether, in their opinion, there is any evidence that supports that thesis. Specifically, we consider the opportunities and challenges for applying ML (the study of computer algorithms that are able to learn from example data and extrapolate the learned information to unseen data) algorithms in toxicologic pathology and how regulatory bodies are navigating this rapidly evolving field. Although we see similarities with the "Last Mile" metaphor, the weight of evidence suggests that toxicologic pathologists should approach ML with an equal dose of skepticism and enthusiasm. There are increasing opportunities for impact in our field that leave the authors cautiously excited and optimistic. Toxicologic pathologists have the opportunity to critically evaluate ML applications with a "call-to-arms" mentality. Why should we be late adopters? There is ample evidence to encourage engagement, growth, and leadership in this field.

Challenges in the Development, Deployment, and Regulation of Artificial Intelligence in Anatomic Pathology

Significant advances in artificial intelligence (AI), deep learning, and other machine-learning approaches have been made in recent years, with applications found in almost every industry, including health care. AI has proved to be capable of completing a spectrum of mundane to complex medically oriented tasks previously performed only by boarded physicians, most recently assisting with the detection of cancers difficult to find on histopathology slides. Although computers will not replace pathologists any time soon, properly designed AI-based tools hold great potential for increasing workflow efficiency and diagnostic accuracy in the practice of pathology. Recent trends, such as data augmentation, crowdsourcing for generating annotated data sets, and unsupervised learning with molecular and/or clinical outcomes versus human diagnoses as a source of ground truth, are eliminating the direct role of pathologists in algorithm development. Proper integration of AI-based systems into anatomic-pathology practice will necessarily require fully digital imaging platforms, an overhaul of legacy information-technology infrastructures, modification of laboratory/pathologist workflows, appropriate reimbursement/cost-offsetting models, and ultimately, the active participation of pathologists to encourage buy-in and oversight. Regulations tailored to the nature and limitations of AI are currently in development and, when instituted, are expected to promote safe and effective use. This review addresses the challenges in AI development, deployment, and regulation to be overcome prior to its widespread adoption in anatomic pathology.