Digital Pathology Research Articles

Generating and evaluating synthetic data in digital pathology through diffusion models

Generating and evaluating synthetic data in digital pathology through diffusion models

How Automation in Electronic Medical Records Can Lead to Errors in Dermatology

The integration of electronic health records has revolutionized healthcare by facilitating communication among providers, but it has also introduced significant challenges. In dermatology, a particular issue arises with the use of electronic medical record (EMR) systems that employ auto-populated fields on digital pathology requisition forms (RFs) for skin biopsies. These systems allow for the rapid selection of pre-set descriptions and differential diagnoses, which, while timesaving, frequently lead to diagnostic inaccuracies and potentially detrimental impacts on patient care. Dermatopathologists often receive biopsy specimens with RFs that list multiple provisional diagnoses based on generic, EMR-generated descriptions, which may not accurately represent the patient’s condition. This practice hampers the ability of dermatopathologists to perform effective clinicopathologic correlation, crucial for accurate diagnosis. We highlight that the convenience of EMR systems can discourage clinicians from recording detailed, accurate clinical observations. Consequently, this lack of detailed documentation prevents dermatopathologists from making informed diagnoses by correlating histologic features with clinical appearances. The current practice of using EMRs without consideration for their clinical relevance not only wastes healthcare resources but also poses a significant medico-legal risk. Therefore, refining EMR practices and integrating more comprehensive clinical data will ensure greater accuracy in dermatological diagnoses.

Using Artificial Intelligence for Personalized Cancer Treatment

The rapid evolution of artificial intelligence and the tools it has created gives rise to significant opportunities for precision oncology and promises a brighter future for the development of the sector. Precision oncology is a modern and advancing branch of cancer treatment and it is a renovated approach which specializes on specific treatment plans rather than “one-size-fits”; with advanced technology including but not limited to genomics, proteomics and bioinformatics, precision oncology focuses on optimizing therapeutic efficacy by targeting the specific molecular and cellular characteristics of a patient's cancer. This review paper explores the sophisticated and complex potential roles of artificial intelligence in cancer treatment, emphasizing on its impact in personalized treatment plans, immunotherapy and diagnostic accuracy. The immense database that AI has access to enables medical professionals to analyze imaging results and detect similarities of symptoms, genetic patterns, offering enhanced prognostic and diagnostic capabilities by eliminating human error and increasing cost effectiveness. This paper also dives into applications such as immune checkpoint inhibitors, digital pathology and demonstrations of the remarkable capability of artificial intelligence in multi-omics integration. Nonetheless, the advent of such advanced artificial intelligence tools, although promising, have not yet become the “gold standard” in oncology due to many rising agitations regarding reliability, since treatment decisions made by opaque AI algorithms are prone to error due to potential biases in AI systems and design errors. By critically examining these points, this paper evaluates both the advantages and disadvantages of the future implementation of artificial intelligence in the field of precision oncology, aiming to devise optimal integration in clinical practice.

BIOS-08. CISD: A NOVEL, USER-FRIENDLY MULTIFACETED ALGORITHM FOR DETECTING STITCHING ARTIFACTS IN LARGE-SCALE CODEX IMAGING DATASETS

Abstract BACKGROUND CODEX (CO-Detection by indEXing) imaging has emerged as a pivotal tool in cellular biology, providing unprecedented insights into cellular distributions and interactions within the tissue microenvironment, crucial for advancing computational pathology. Despite its advantages, CODEX is susceptible to stitching errors during the merging of image tiles, which can significantly distort spatial analysis by introducing artifacts that obscure true cellular architecture. METHODS Addressing this challenge, we introduce CISD (CODEX Image Stitch Detection), a novel, user-friendly algorithm optimized for the detection of stitching artifacts in large-scale CODEX datasets. Developed and validated using a comprehensive dataset from 7 matched glioblastoma patients—encompassing over 2800 images across 29 immune protein markers—CISD utilizes a combination of image preprocessing, statistical measures, and thresholding to detect stitching errors. RESULTS CISD demonstrates robust performance, achieving a median Area Under Curve (AUC) value in the gold standard dataset where 73% of markers exceed 70% AUC, and 17% surpass 90%. Further validation shows that in a secondary dataset, 45% of markers maintain AUC values over 70% with 38% exceeding 90%, validating the algorithm’s consistent efficacy across varied datasets. CONCLUSION CISD’s significance lies not only in its computational achievements but also in its accessibility, facilitating high-quality analysis of extensive datasets without complex pattern recognition requirements. This democratizes high-precision computational pathology, enabling more researchers to accurately analyze spatial data. By reducing technical barriers, CISD sets a new benchmark in CODEX imaging, enhancing data reliability and accelerating research discoveries in cellular biology. This innovation stands to substantially refine spatial analysis in digital pathology, ensuring dependable interpretation of complex biological data.

Integrating transcriptomic data and digital pathology for NRG-based prediction of prognosis and therapy response in gastric cancer

Background Cancer is characterized by its ability to resist cell death, and emerging evidence suggests a potential correlation between non-apoptotic regulated cell death (RCD), tumor progression, and therapy response. However, the prognostic significance of non-apoptotic RCD-related genes (NRGs) and their relationships with immune response in gastric cancer (GC) remain unclear. Methods In this study, RNA-seq gene expression and clinical information of GC patients were acquired from The Cancer Genome Atlas and the Gene Expression Omnibus databases. Cox and LASSO regression analyses were used to construct the NRG signature. Moreover, we developed a deep learning model based on ResNet50 to predict the NRG signature from digital pathology slides. The expression of signature hub genes was validated using real-time quantitative PCR and single-cell RNA sequencing data. Results We identified 13 NRGs as signature genes for predicting the prognosis of patients with GC. The high-risk group, characterized by higher NRG scores, demonstrated a shorter overall survival rate, increased immunosuppressive cell infiltration, and immune dysfunction. Moreover, associations were observed between the NRG signature and chemotherapeutic drug responsiveness, as well as immunotherapy effectiveness in GC patients. Furthermore, the deep learning model effectively stratified GC patients based on the NRG signature by leveraging morphological variances, showing promising results for the classification of GC patients. Validation experiments demonstrated that the expression level of SERPINE1 was significantly upregulated in GC, while the expression levels of GPX3 and APOD were significantly downregulated. Conclusion The NRG signature and its deep learning model have significant clinical implications, highlighting the importance of individualized treatment strategies based on GC subtyping. These findings provide valuable insights for guiding clinical decision-making and treatment approaches for GC.

Application of Nexus Learning framework for Personalized Medicine and Drug Development

Biological systems provide an enormous amount of knowledge on development and disease. The pharmaceutical industry is facing new possibilities and challenges with the rise of high-throughput methods to study disease and biology. The aim is to develop drugs based on acceptable therapeutic assumptions. Nexus Learning provides a variety of tools that help improve discovery and decision making using extensive, high-quality data and well-specified queries. Potential applications of NLF can be found across the whole drug development process. Clinical study target validation, prognostic biomarker discovery, and digital pathology data processing are a few examples. There has been a wide variety of applications in terms of technique and environment, with some methods producing reliable forecasts and insights. The primary problem with using NLF is that the findings provided by NL tend to be easy to understand or reproduce, which can limit their value. The collection of complete and methodical high-dimensional data remains an essential need across all domains. By addressing these challenges and increasing understanding about what is required to verify NLF methodologies, we can employ NLF to make data-driven decisions, which can speed up drug discovery and development and minimize failure rates.

Determination of Ki-67 indices in neuroendocrine tumours of the gastrointestinal tract: the past, the present, and the future.

Ki-67 index (Ki-67i) is integral to the grading of many tumours. There remains considerable variability across pathologists in methods used to determine Ki-67i and in their results. Manual counting (or "eyeballing") is widely used, but digital pathology tools such as web-based image analysis and artificial intelligence-assisted cell detection software have become available in daily pathology practice. This study aims to compare the accuracy and efficiency of manual and two digital methods of Ki-67i determination. H&E and Ki-67 immunohistochemical (IHC) slides/images of 12 gastrointestinal neuroendocrine tumours (GI-NETs) were provided to 8 pathologists to evaluate Ki-67i via manual estimation (ME; the "past"), web-based image analysis using cellular segmentation (AI4Path.ca; the "present"), and software-based image analysis with built-in AI algorithms (QuPath; the "future"). Data collected include Ki67i, time expended, total cells counted, and pathologists' confidence level in the reported result. Deviation of Ki-67i from a gold standard result (GS) was analyzed using multiple linear regression, and results were compared via paired t test. Our results found no statistically significant differences in Ki-67i deviation from GS when comparing ME and AI4P methods for all 12 cases. The QP Ki-67i detection accuracy varied significantly. ME was the method with the least time expenditure. Junior pathologists are less confident in ME. Grading consensus was comparable among all three methods. These findings suggest that while digital pathology can confer increased Ki-67i accuracy in some cases of GI-NETs, higher time expenditure and proper hotspot selection may represent barriers to the adoption of digital pathology methods in the future.

Impact of metabolism-related markers on outcomes in ovarian cancer patients: Findings of the MITO16A/MaNGO-OV2 trial.

In ovarian cancer, expression of metabolism-related markers has been investigated in several studies focusing on individual markers; however, a parallel quantitative evaluation of markers mapping to distinct metabolic processes and their prognostic value in large patient cohorts is still lacking. Here, by using immunohistochemistry followed by digital pathology, we investigated the expression of several markers related to glycolysis including monocarboxylate transporter 1 and 4 (MCT1, MCT4), glutamine metabolism (glutaminase, GLS) and hypoxia/acidosis (carbonic anhydrase 9, CA IX) in tissue microarrays of > 300 patients recruited in the MITO16A clinical trial, which involved treatment of ovarian cancer patients with carboplatin/taxol plus bevacizumab. Regarding the prognostic impact of these markers, results indicate that GLS expression correlated with progression-free survival, but this effect disappeared when data were corrected for multiple testing. All other markers showed no correlation with clinical outcome. These results indicate marked heterogeneity of expression of metabolism-associated markers in ovarian cancer; however, there was a lack of association with clinical benefit after chemotherapy/anti-vascular endothelial growth factor treatment. Notwithstanding the lack of prognostic value, knowledge of the pattern of expression of these biomarkers in tumors can be useful for patient stratification purposes when new drugs targeting these metabolic pathways will be tested.

Complex shape markers can detect alterations in the spatial distribution of cell nuclei in human lung squamous cell carcinoma: a useful tool for automatic analysis?

Lung cancer is the leading cause of cancer-related death. The use of computational methods to quantify changes that are not perceptible to the human eye is increasing in digital pathology imaging and has quickly improved detection rates at a low cost. Therefore, the present study aims to use complex computational shape markers as tools for automated analysis of the spatial distribution of cells in microscopy images of squamous cell lung carcinoma (SqCC). Photomicrographs from pathology glass slides in the LC25000 dataset were used in this study. Compared with those of the control, the fractal dimension (28%) and lacunarity (41%) of the cell nuclei changed in SqCC. The multifractal analysis revealed a significant difference in parameters Dq, α, and f(α) for all values of q (-10 to + 10), with a greater increase for more positive q values. The values at q + 10 increased by 34% for Dq, 36% for α, and 53% for f(α) in the SqCC images. The circularity, area, and perimeter also changed in the SqCC images. However, the parameters of aspect ratio, roundness, and solidity did not significantly differ between SqCC and benign tissue. The complex shape markers with the greatest changes in this study were the f(α) values for multifractality (53%) and lacunarity (41%). In conclusion, automated quantification of the spatial distribution of cell nuclei can be a fast, low-cost tool for evaluating the microscopic characteristics of SqCC; therefore, complex shape markers could be useful tools for software and artificial intelligence to detect lung carcinoma.

A digital pathology model for predicting radioiodine-avid metastases on initial post-therapeutic 131I scan in patients with papillary thyroid cancer.

Accurate postoperative assessment is critical for optimizing 131I therapy in patients with papillary thyroid cancer (PTC). This study aimed to develop a pathology model utilizing postoperative digital pathology slides to predict lymph node and/or distant metastases on post-therapeutic 131I scan after initial 131I treatment in PTC patients. A retrospective analysis was conducted on 229 PTC patients who underwent total or near-total thyroidectomy and subsequent 131I treatment after levothyroxine (LT4) withdrawal between January 2022 and August 2023. The pathology model was developed through two stages: patch-level prediction and WSI-level prediction. The clinical model was constructed using statistically significant variables identified from univariate and multivariate logistic regression analysis. Of the 229 patients, 19.6% (45/229) exhibited 131I-avid metastatic foci in post-therapeutic 131I scan. Multifactorial analysis identified stimulated thyroglobulin (sTg) as the sole independent risk factor. The AUC of the pathology model in the training and test cohorts were 0.976 (95% CI 0.948-1.000) and 0.805 (95% CI 0.660-0.951), respectively, which were significantly higher than the clinical model (AUC 0.652 and 0.548, Pall < 0.05). This model has the potential to serve as a valuable tool for clinicians in tailoring treatment strategies, thereby optimizing therapeutic outcomes for PTC patients.

Enhanced Computer-Aided Digital Imaging Technique for Predictions in Breast Cancer

Background: Breast cancer (BRCA) is the most frequently diagnosed cancer in women, with a rise in occurrences and fatalities. The field of BRCA prediction and diagnosis has witnessed significant advancements in recent years, particularly emphasizing enhanced computer-aided digital imaging techniques, and has emerged as a powerful ally in the prediction of BRCA through histopathology image analysis. A number of approaches have been suggested in recent years for the categorization of histopathology BRCA images into benign and malignant as it examines the images at cellular level. The histopathology slides must be manually analysed which is time consuming and tiresome and is prone to human error. Additionally, different laboratories occasionally have different interpretation of these images. Methods: This paper focuses on implementing a framework for Computer-Aided digital imaging technique that can serve as a decision support. With recent advancements in computing power the analysis of BRCA histopathology image samples has become easier. Stain normalization (SN), segmentation, feature extraction and classification are the steps to categorize the cancer into benign and malignant. Nuclei segmentation is a crucial step that needs to be taken into account in order to establish malignancy. These are considered essential for early diagnosis of BRCA. A unique method proposed for BRCA prediction is put forward. To maximize the prediction accuracy, the suggested method is integrated with machine learning (ML) techniques and clinical data is used to evaluate the suggested approach. Results: This strategy is adaptable to many cancer types and imaging techniques. The suggested technique is applied to clinical data and is integrated with logistic regression and K-Nearest Neighbor resulting in accuracy of 92.10% and 86.89% respectively for BRCA histopathology images. Conclusion: The objective of this work is to validate the proposed model which takes input as feature pattern for a given label. For the collected clinical samples, the model is able to classify the input as benign or malignant. The proposed model worked efficiently for different BC datasets and performed classification task successfully. Integrating mathematical model (MM) with ML model for interpreting histopathology BRCA is a potential area of research in the field of digital pathology.

Out-of-distribution detection in digital pathology: Do foundation models bring the end to reconstruction-based approaches?

Artificial intelligence (AI) has shown promising results for computational pathology tasks. However, one of the limitations in clinical practise is that these algorithms are optimised for the distribution represented by the training data. For out-of-distribution (OOD) data, they often deliver predictions with equal confidence, even though these often are incorrect. In the pursuit of OOD detection in digital pathology, this study evaluates the state-of-the-art (SOTA) in computational pathology OOD detection, based on diffusion probabilistic models, specifically by adapting the latent diffusion model (LDM) for this purpose (AnoLDM). We compare this against post-hoc methods based on the latent space of foundation models, which are SOTA in general computer vision research. The approaches are not only evaluated on data from the same medical centres as the training set, but also on several datasets with data distribution shifts. The results show that AnoLDM performs similarly well or better than diffusion model based approaches published in previous studies in computational pathology but with reduced computational costs. However, our optimal configuration of an approach based on foundation models (kang_residual) outperforms AnoLDM on OOD detection on data not experiencing any covariate shifts, with an AUROC of 96.17 versus 91.86. Interestingly, AnoLDM is more successful at handling the data distribution shifts investigated in this study. However, both AnoLDM and kang_residual suffer substantial loss in the performance under the data distribution shifts, hence future work should focus on improving the generalisation of OOD detection for computational pathology applications.

Deep learning-driven macroscopic AI segmentation model for brain tumor detection via digital pathology: Foundations for terahertz imaging-based AI diagnostics

Deep learning-driven macroscopic AI segmentation model for brain tumor detection via digital pathology: Foundations for terahertz imaging-based AI diagnostics

FASNet: Feature alignment-based method with digital pathology images in assisted diagnosis medical system

FASNet: Feature alignment-based method with digital pathology images in assisted diagnosis medical system

Construction of a combined prognostic model for pancreatic ductal adenocarcinoma based on deep learning and digital pathology images

BackgroundDeep learning has made significant advancements in the field of digital pathology, and the integration of multiple models has further improved accuracy. In this study, we aimed to construct a combined prognostic model using deep learning-extracted features from digital pathology images of pancreatic ductal adenocarcinoma (PDAC) alongside clinical predictive indicators and to explore its prognostic value.MethodsA retrospective analysis was conducted on 142 postoperative pathologically confirmed PDAC cases. These cases were divided into training (n = 114) and testing sets (n = 28) at an 8:2 ratio. Tumor whole-slide imaging features were extracted and screened to construct a pathological risk model based on a pre-trained deep learning model. Clinical and pathological data from the training set were used to select independent predictive factors for PDAC and establish a clinical risk model using LASSO, univariate, and multivariate Cox regression analyses. Based on the pathological and clinical risk models, a combined model was developed. The Harrell concordance index (C-index) was computed to assess the predictive performance of each model for PDAC survival prognosis.ResultsFor the training and testing sets, the C-index values for the clinical risk model were 0.76 and 0.75, respectively; for the pathological risk model, they were 0.82 and 0.73, respectively; and for the combined model, they were 0.86 and 0.77, respectively. The combined model exhibited appropriate calibration at 1-, 3-, and 5-year time points, as well as a superior area under the curve of the receiver operating characteristic curve and clinical net benefit compared to the single models.ConclusionsIntegrating the pathological and clinical risk models may provide a higher predictive value for survival prognosis.

Applications of artificial intelligence in digital pathology for gastric cancer.

Gastric cancer is one of the most common cancers and is one of the leading causes of cancer-related deaths in worldwide. Early diagnosis and treatment are essential for a positive outcome. The integration of artificial intelligence in the pathology field is increasingly widespread, including histopathological images analysis. In recent years, the application of digital pathology technology emerged as a potential solution to enhance the understanding and management of gastric cancer. Through sophisticated image analysis algorithms, artificial intelligence technologies facilitate the accuracy and sensitivity of gastric cancer diagnosis and treatment and personalized therapeutic strategies. This review aims to evaluate the current landscape and future potential of artificial intelligence in transforming gastric cancer pathology, so as to provide ideas for future research.

A comprehensive evaluation of an artificial intelligence based digital pathology to monitor large-scale deworming programs against soil-transmitted helminths: A study protocol.

Manual screening of a Kato-Katz (KK) thick stool smear remains the current standard to monitor the impact of large-scale deworming programs against soil-transmitted helminths (STHs). To improve this diagnostic standard, we recently designed an artificial intelligence based digital pathology system (AI-DP) for digital image capture and analysis of KK thick smears. Preliminary results of its diagnostic performance are encouraging, and a comprehensive evaluation of this technology as a cost-efficient end-to-end diagnostic to inform STH control programs against the target product profiles (TPP) of the World Health Organisation (WHO) is the next step for validation. Here, we describe the study protocol for a comprehensive evaluation of the AI-DP based on its (i) diagnostic performance, (ii) repeatability/reproducibility, (iii) time-to-result, (iv) cost-efficiency to inform large-scale deworming programs, and (v) usability in both laboratory and field settings. For each of these five attributes, we designed separate experiments with sufficient power to verify the non-inferiority of the AI-DP (KK2.0) over the manual screening of the KK stool thick smears (KK1.0). These experiments will be conducted in two STH endemic countries with national deworming programs (Ethiopia and Uganda), focussing on school-age children only. This comprehensive study will provide the necessary data to make an evidence-based decision on whether the technology is indeed performant and a cost-efficient end-to-end diagnostic to inform large-scale deworming programs against STHs. Following the protocolized collection of high-quality data we will seek approval by WHO. Through the dissemination of our methodology and statistics, we hope to support additional developments in AI-DP technologies for other neglected tropical diseases in resource-limited settings. The trial was registered on September 29, 2023 Clinicaltrials.gov (ID: NCT06055530).

Public evidence on AI products for digital pathology

Novel products applying artificial intelligence (AI)-based methods to digital pathology images are touted to have many uses and benefits. However, publicly available information for products can be variable, with few sources of independent evidence. This review aimed to identify public evidence for AI-based products for digital pathology. Key features of products on the European Economic Area/Great Britain (EEA/GB) markets were examined, including their regulatory approval, intended use, and published validation studies. There were 26 AI-based products that met the inclusion criteria and, of these, 24 had received regulatory approval via the self-certification route as General in vitro diagnostic (IVD) medical devices. Only 10 of the products (38%) had peer-reviewed internal validation studies and 11 products (42%) had peer-reviewed external validation studies. To support transparency an online register was developed using identified public evidence (https://osf.io/gb84r/), which we anticipate will provide an accessible resource on novel devices and support decision making.

Precise grading of non-muscle invasive bladder cancer with multi-scale pyramidal CNN

The grading of non-muscle invasive bladder cancer (NMIBC) continues to face challenges due to subjective interpretations, which affect the assessment of its severity. To address this challenge, we are developing an innovative artificial intelligence (AI) system aimed at objectively grading NMIBC. This system uses a novel convolutional neural network (CNN) architecture called the multi-scale pyramidal pretrained CNN to analyze both local and global pathology markers extracted from digital pathology images. The proposed CNN structure takes as input three levels of patches, ranging from small patches (e.g., 128×128\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$128 \ imes 128$$\\end{document}) to the largest size patches (512×512\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$512 \ imes 512$$\\end{document}). These levels are then fused by random forest (RF) to estimate the severity grade of NMIBC. The optimal patch sizes and other model hyperparameters are determined using a grid search algorithm. For each patch size, the proposed system has been trained on 32K patches (comprising 16K low-grade and 16K high-grade samples) and subsequently tested on 8K patches (consisting of 4K low-grade and 4K high-grade samples), all annotated by two pathologists. Incorporating light and efficient processing, defining new benchmarks in the application of AI to histopathology, the ShuffleNet-based AI system achieved notable metrics on the testing data, including 94.25% ± 0.70% accuracy, 94.47% ± 0.93% sensitivity, 94.03% ± 0.95% specificity, and a 94.29% ± 0.70% F1-score. These results highlight its superior performance over traditional models like ResNet-18. The proposed system’s robustness in accurately grading pathology demonstrates its potential as an advanced AI tool for diagnosing human diseases in the domain of digital pathology.

A deep-learning-based histopathology classifier for focal cortical dysplasia (FCD) unravels a complex scenario of comorbid FCD subtypes.

Recently, we developed a first artificial intelligence (AI)-based digital pathology classifier for focal cortical dysplasia (FCD) as defined by the ILAE classification. Herein, we tested the usefulness of the classifier in a retrospective histopathology workup scenario. Eighty-six new cases with histopathologically confirmed FCD ILAE type Ia (FCDIa), FCDIIa, FCDIIb, mild malformation of cortical development with oligodendroglial hyperplasia in epilepsy (MOGHE), or mild malformations of cortical development were selected, 20 of which had confirmed gene mosaicism. The classifier always recognized the correct histopathology diagnosis in four or more 1000 × 1000-μm digital tiles in all cases. Furthermore, the final diagnosis overlapped with the largest batch of tiles assigned by the algorithm to one diagnostic entity in 80.2% of all cases. However, 86.2% of all cases revealed more than one diagnostic category. As an example, FCDIIb was identified in all of the 23 patients with histopathologically assigned FCDIIb, whereas the classifier correctly recognized FCDIIa tiles in 19 of these cases (83%), that is, dysmorphic neurons but no balloon cells. In contrast, the classifier misdiagnosed FCDIIb tiles in seven of 23 cases histopathologically assigned to FCDIIa (33%). This mandates a second look by the signing histopathologist to either confirm balloon cells or differentiate from reactive astrocytes. The algorithm also recognized coexisting architectural dysplasia, for example, vertically oriented microcolumns as in FCDIa, in 22% of cases classified as FCDII and in 62% of cases with MOGHE. Microscopic review confirmed microcolumns in the majority of tiles, suggesting that vertically oriented architectural abnormalities are more common than previously anticipated. An AI-based diagnostic classifier will become a helpful tool in our future histopathology laboratory, in particular when large anatomical resections from epilepsy surgery require extensive resources. We also provide an open access web application allowing the histopathologist to virtually review digital tiles obtained from epilepsy surgery to corroborate their final diagnosis.