The development of artificial intelligence in the histological diagnosis of Inflammatory Bowel Disease (IBD-AI).

BACKGROUND: Inflammatory bowel disease (IBD) includes Crohn's Disease (CD) and Ulcerative Colitis (UC). Correct diagnosis requires the identification of precise morphological features such basal plasmacytosis. However, histopathological interpretation can be challenging, and it is subject to high variability. AIM: The IBD-Artificial Intelligence (AI) project aims at the development of an AI-based evaluation system to support the diagnosis of IBD, semi-automatically quantifying basal plasmacytosis. METHODS: A deep learning model was trained to detect and quantify plasma cells on a public dataset of 4981 annotated images. The model was then tested on an external validation cohort of 356 intestinal biopsies of CD, UC and healthy controls. AI diagnostic performance was calculated compared to human gold standard. RESULTS: The system correctly found that CD and UC samples had a greater prevalence of basal plasma cells with mean number of PCs within ROIs of 38.22 (95 % CI: 31.73, 49.04) for CD, 55.16 (46.57, 65.93) for UC, and 17.25 (CI: 12.17, 27.05) for controls. Overall, OR=4.968 (CI: 1.835, 14.638) was found for IBD compared to normal mucosa (CD: +59 %; UC: +129 %). Additionally, as expected, UC samples were found to have more plasma cells in colon than CD cases. CONCLUSION: Our model accurately replicated human assessment of basal plasmacytosis, underscoring the value of AI models as a potential aid IBD diagnosis.

Multiomic foundations of human prefrontal cortex tissue function.

The prefrontal cortex (PFC) is a region of the brain that in humans is involved in the production of higher-order functions such as cognition, emotion, perception, and behavior. Neurotransmission in the PFC produces higher-order functions by integrating information from other areas of the brain. At the foundation of neurotransmission, and by extension at the foundation of higher-order brain functions, are an untold number of coordinated molecular processes involving the DNA sequence variants in the genome, RNA transcripts in the transcriptome, and proteins in the proteome. These "multiomic" foundations are poorly understood in humans, perhaps in part because most modern studies that characterize the molecular state of the human PFC use tissue obtained when neurotransmission and higher-order brain functions have ceased (i.e., the postmortem state). Here, analyses are presented on data generated for the Living Brain Project (LBP) to investigate whether PFC tissue from individuals with intact higher-order brain function has characteristic multiomic foundations. Two complementary strategies were employed towards this end. The first strategy was to identify in PFC samples obtained from living study participants a signature of RNA transcript expression associated with neurotransmission measured intracranially at the time of PFC sampling, in some cases while participants performed a task engaging higher-order brain functions. The second strategy was to perform multiomic comparisons between PFC samples obtained from individuals with intact higher-order brain function at the time of sampling (i.e., living study participants) and PFC samples obtained in the postmortem state. RNA transcript expression within multiple PFC cell types was associated with fluctuations of dopaminergic, serotonergic, and/or noradrenergic neurotransmission in the substantia nigra measured while participants played a computer game that engaged higher-order brain functions. A subset of these associations - termed the "transcriptional program associated with neurotransmission" (TPAWN) - were reproduced in analyses of brain RNA transcript expression and intracranial neurotransmission data obtained from a second LBP cohort and from a cohort in an independent study. RNA transcripts involved in TPAWN were found to be (1) enriched for RNA transcripts associated with measures of neurotransmission in rodent and cell models, (2) enriched for RNA transcripts encoded by evolutionarily constrained genes, (3) depleted of RNA transcripts regulated by common DNA sequence variants, and (4) enriched for RNA transcripts implicated in higher-order brain functions by human population genetic studies. In PFC excitatory neurons of living study participants, higher expression of the genes in TPAWN tracked with higher expression of RNA transcripts that in rodent PFC samples are markers of a class of excitatory neurons that connect the PFC to deep brain structures. TPAWN was further reproduced by RNA transcript expression patterns differentiating living PFC samples from postmortem PFC samples, and significant differences between living and postmortem PFC samples were additionally observed with respect to (1) the expression of most primary RNA transcripts, mature RNA transcripts, and proteins, (2) the splicing of most primary RNA transcripts into mature RNA transcripts, (3) the patterns of co-expression between RNA transcripts and proteins, and (4) the effects of some DNA sequence variants on RNA transcript and protein expression. Taken together, this report highlights that studies of brain tissue obtained in a safe and ethical manner from large cohorts of living individuals can help advance understanding of the multiomic foundations of brain function.

Visceral Adiposity and Progression of ADPKD: A Cohort Study of Patients From the TEMPO 3:4 Trial.

RATIONALE & OBJECTIVE: Body mass index (BMI) is an independent predictor of kidney disease progression in individuals with autosomal dominant polycystic kidney disease (ADPKD). Adipocytes do not simply act as a fat reservoir but are active endocrine organs. We hypothesized that greater visceral abdominal adiposity would associate with more rapid kidney growth in ADPKD and influence the efficacy of tolvaptan. STUDY DESIGN: A retrospective cohort study. SETTING & PARTICIPANTS: 1,053 patients enrolled in the TEMPO 3:4 tolvaptan trial with ADPKD and at high risk of rapid disease progression. PREDICTOR: Estimates of visceral adiposity extracted from coronal plane magnetic resonance imaging (MRI) scans using deep learning. OUTCOME: Annual change in total kidney volume (TKV) and effect of tolvaptan on kidney growth. ANALYTICAL APPROACH: Multinomial logistic regression and linear mixed models. RESULTS: In fully adjusted models, the highest tertile of visceral adiposity was associated with greater odds of annual change in TKV of≥7% versus<5% (odds ratio [OR], 4.78 [95% CI, 3.03-7.47]). The association was stronger in women than men (interaction P<0.01). In linear mixed models with an outcome of percent change in TKV per year, tolvaptan efficacy (% change in TKV) was reduced with higher visceral adiposity (3-way interaction of treatment ∗ time ∗ visceral adiposity, P=0.002). Visceral adiposity significantly improved classification performance of predicting rapid annual percent change in TKV for individuals with a normal BMI (DeLong's test z score: -2.03; P=0.04). Greater visceral adiposity was not associated with estimated glomerular filtration rate (eGFR) slope in the overall cohort; however, visceral adiposity was associated with more rapid decline in eGFR slope (below the median) in women (fully adjusted OR, 1.06 [95% CI, 1.01-1.11] per 10 unit increase in visceral adiposity) but not men (OR, 0.98 [95% CI, 0.95-1.02]). LIMITATIONS: Retrospective; rapid progressors; computational demand of deep learning. CONCLUSIONS: Visceral adiposity that can be quantified by MRI in the coronal plane using a deep learning segmentation model independently associates with more rapid kidney growth and improves classification of rapid progression in individuals with a normal BMI. Tolvaptan efficacy decreases with increasing visceral adiposity. PLAIN-LANGUAGE SUMMARY: We analyzed images from a previous study with the drug tolvaptan conducted in patients with autosomal dominant polycystic kidney disease (ADPKD) to measure the amount of fat tissue surrounding the kidneys (visceral fat). We had previously shown body mass index can predict kidney growth in this population; now we determined whether visceral fat was an important factor associated with kidney growth. Using a machine learning tool to automate measurement of fat in images, we observed that visceral fat was independently associated with kidney growth, that it was a better predictor of faster kidney growth in lean patients than body mass index, and that having more visceral fat made treatment of ADPKD with tolvaptan less effective.

AI models for automated segmentation of engineered polycystic kidney tubules.

Autosomal dominant polycystic kidney disease (ADPKD) is a monogenic, rare disease, characterized by the formation of multiple cysts that grow out of the renal tubules. Despite intensive attempts to develop new drugs or repurpose existing ones, there is currently no definitive cure for ADPKD. This is primarily due to the complex and variable pathogenesis of the disease and the lack of models that can faithfully reproduce the human phenotype. Therefore, the development of models that allow automated detection of cysts' growth directly on human kidney tissue is a crucial step in the search for efficient therapeutic solutions. Artificial Intelligence methods, and deep learning algorithms in particular, can provide powerful and effective solutions to such tasks, and indeed various architectures have been proposed in the literature in recent years. Here, we comparatively review state-of-the-art deep learning segmentation models, using as a testbed a set of sequential RGB immunofluorescence images from 4 in vitro experiments with 32 engineered polycystic kidney tubules. To gain a deeper understanding of the detection process, we implemented both pixel-wise and cyst-wise performance metrics to evaluate the algorithms. Overall, two models stand out as the best performing, namely UNet++ and UACANet: the latter uses a self-attention mechanism introducing some explainability aspects that can be further exploited in future developments, thus making it the most promising algorithm to build upon towards a more refined cyst-detection platform. UACANet model achieves a cyst-wise Intersection over Union of 0.83, 0.91 for Recall, and 0.92 for Precision when applied to detect large-size cysts. On all-size cysts, UACANet averages at 0.624 pixel-wise Intersection over Union. The code to reproduce all results is freely available in a public GitHub repository.

Evolving educational landscape in pathology: a comprehensive bibliometric and visual analysis including digital teaching and learning resources.

AIMS: Pathology education is a core component of medical training, and its literature is critical for refining educational modalities. We performed a cross-sectional bibliometric analysis to explore publications on pathology education, focusing on new medical education technologies. METHODS: The analysis identified 64 pathology journals and 53 keywords. Relevant articles were collected using a web application, PaperScraper, developed to accelerate literature search. Citation data were collected from multiple sources. Descriptive statistics, with time period analysis, were performed using Microsoft Excel and visualised with Flourish Studio. Two article groups were further investigated with a bibliometric software, VOSViewer, to establish co-authorship and keyword relationships. RESULTS: 8946 citations were retrieved from 905 selected articles. Most articles were published in the last decade (447, 49.4%). The top journals were Archives of Pathology & Laboratory Medicine (184), Human Pathology (122) and the American Journal of Clinical Pathology (117). The highest number of citations was found for Human Pathology (2120), followed by Archives of Pathology & Laboratory Medicine (2098) and American Journal of Clinical Pathology (1142). Authors with different backgrounds had the greatest number of articles and citations. 12 co-authorship, 3 keyword and 8 co-citation clusters were found for the social media/online resources group, 8 co-authorship, 4 keyword and 7 co-citation clusters for the digital pathology/virtual microscopy/mobile technologies group. CONCLUSIONS: The analysis revealed a significant increase in publications over time. The emergence of digital teaching and learning resources played a major role in this growth. Overall, these findings underscore the transformative potential of technology in pathology education.

A Weakly Supervised Deep Learning Framework for Whole Slide Classification to Facilitate Digital Pathology in Animal Study.

The pathology of animal studies is crucial for toxicity evaluations and regulatory assessments, but the manual examination of slides by pathologists remains time-consuming and requires extensive training. One inherent challenge in this process is the interobserver variability, which can compromise the consistency and accuracy of a study. Artificial intelligence (AI) has demonstrated its ability to automate similar examinations in clinical applications with enhanced efficiency, consistency, and accuracy. However, training AI models typically relies on costly pixel-level annotation of injured regions and is often not available for animal pathology. To address this, we developed the PathologAI system, a "weakly" supervised approach for WSI classification in rat images without explicit lesion annotation at the pixel level. Using rat liver imaging data from the Open TG-GATEs system, PathologAI was applied to predict necrosis of n = 816 WSIs (377 controls). TG-GATEs studied 170 compounds at three dose levels (low, middle, and high) for four time points (3, 7, 14, and 28 days). PathologAI first preprocessed WSIs at the tile level to generate a high-level representation with a Generative Adversarial Network architecture. The prediction of liver necrosis relied on an ensemble model of 5 CNN classifiers trained on 335 WSIs. The ensemble model achieved notable classification accuracy on the holdout test set: 87% among 87 control slides free of findings, 83% among 120 controls with spontaneous necrosis, 67% among 147 treated animals with spontaneous minimal or slight necrosis, and 59% among 127 treated animals with minimal or slight necrosis caused by the treatment. Importantly, PathologAI was able to discriminate WSIs with spontaneous necrosis from those with treatment related necrosis and discriminated mild lesion level findings (slight vs minimal) and between treatment dose levels. PathologAI could provide an inexpensive and rapid screening tool to assist the digital pathology analysis in preclinical applications and general toxicological studies.

Quantification of the Immune Content in Neuroblastoma: Deep Learning and Topological Data Analysis in Digital Pathology.

We introduce here a novel machine learning (ML) framework to address the issue of the quantitative assessment of the immune content in neuroblastoma (NB) specimens. First, the EUNet, a U-Net with an EfficientNet encoder, is trained to detect lymphocytes on tissue digital slides stained with the CD3 T-cell marker. The training set consists of 3782 images extracted from an original collection of 54 whole slide images (WSIs), manually annotated for a total of 73,751 lymphocytes. Resampling strategies, data augmentation, and transfer learning approaches are adopted to warrant reproducibility and to reduce the risk of overfitting and selection bias. Topological data analysis (TDA) is then used to define activation maps from different layers of the neural network at different stages of the training process, described by persistence diagrams (PD) and Betti curves. TDA is further integrated with the uniform manifold approximation and projection (UMAP) dimensionality reduction and the hierarchical density-based spatial clustering of applications with noise (HDBSCAN) algorithm for clustering, by the deep features, the relevant subgroups and structures, across different levels of the neural network. Finally, the recent TwoNN approach is leveraged to study the variation of the intrinsic dimensionality of the U-Net model. As the main task, the proposed pipeline is employed to evaluate the density of lymphocytes over the whole tissue area of the WSIs. The model achieves good results with mean absolute error 3.1 on test set, showing significant agreement between densities estimated by our EUNet model and by trained pathologists, thus indicating the potentialities of a promising new strategy in the quantification of the immune content in NB specimens. Moreover, the UMAP algorithm unveiled interesting patterns compatible with pathological characteristics, also highlighting novel insights into the dynamics of the intrinsic dataset dimensionality at different stages of the training process. All the experiments were run on the Microsoft Azure cloud platform.

Integrative Network Fusion: A Multi-Omics Approach in Molecular Profiling.

Recent technological advances and international efforts, such as The Cancer Genome Atlas (TCGA), have made available several pan-cancer datasets encompassing multiple omics layers with detailed clinical information in large collection of samples. The need has thus arisen for the development of computational methods aimed at improving cancer subtyping and biomarker identification from multi-modal data. Here we apply the Integrative Network Fusion (INF) pipeline, which combines multiple omics layers exploiting Similarity Network Fusion (SNF) within a machine learning predictive framework. INF includes a feature ranking scheme (rSNF) on SNF-integrated features, used by a classifier over juxtaposed multi-omics features (juXT). In particular, we show instances of INF implementing Random Forest (RF) and linear Support Vector Machine (LSVM) as the classifier, and two baseline RF and LSVM models are also trained on juXT. A compact RF model, called rSNFi, trained on the intersection of top-ranked biomarkers from the two approaches juXT and rSNF is finally derived. All the classifiers are run in a 10x5-fold cross-validation schema to warrant reproducibility, following the guidelines for an unbiased Data Analysis Plan by the US FDA-led initiatives MAQC/SEQC. INF is demonstrated on four classification tasks on three multi-modal TCGA oncogenomics datasets. Gene expression, protein expression and copy number variants are used to predict estrogen receptor status (BRCA-ER, N = 381) and breast invasive carcinoma subtypes (BRCA-subtypes, N = 305), while gene expression, miRNA expression and methylation data is used as predictor layers for acute myeloid leukemia and renal clear cell carcinoma survival (AML-OS, N = 157; KIRC-OS, N = 181). In test, INF achieved similar Matthews Correlation Coefficient (MCC) values and 97% to 83% smaller feature sizes (FS), compared with juXT for BRCA-ER (MCC: 0.83 vs. 0.80; FS: 56 vs. 1801) and BRCA-subtypes (0.84 vs. 0.80; 302 vs. 1801), improving KIRC-OS performance (0.38 vs. 0.31; 111 vs. 2319). INF predictions are generally more accurate in test than one-dimensional omics models, with smaller signatures too, where transcriptomics consistently play the leading role. Overall, the INF framework effectively integrates multiple data levels in oncogenomics classification tasks, improving over the performance of single layers alone and naive juxtaposition, and provides compact signature sizes.

Predictability of drug-induced liver injury by machine learning.

BACKGROUND: Drug-induced liver injury (DILI) is a major concern in drug development, as hepatotoxicity may not be apparent at early stages but can lead to life threatening consequences. The ability to predict DILI from in vitro data would be a crucial advantage. In 2018, the Critical Assessment Massive Data Analysis group proposed the CMap Drug Safety challenge focusing on DILI prediction. METHODS AND RESULTS: The challenge data included Affymetrix GeneChip expression profiles for the two cancer cell lines MCF7 and PC3 treated with 276 drug compounds and empty vehicles. Binary DILI labeling and a recommended train/test split for the development of predictive classification approaches were also provided. We devised three deep learning architectures for DILI prediction on the challenge data and compared them to random forest and multi-layer perceptron classifiers. On a subset of the data and for some of the models we additionally tested several strategies for balancing the two DILI classes and to identify alternative informative train/test splits. All the models were trained with the MAQC data analysis protocol (DAP), i.e., 10x5 cross-validation over the training set. In all the experiments, the classification performance in both cross-validation and external validation gave Matthews correlation coefficient (MCC) values below 0.2. We observed minimal differences between the two cell lines. Notably, deep learning approaches did not give an advantage on the classification performance. DISCUSSION: We extensively tested multiple machine learning approaches for the DILI classification task obtaining poor to mediocre performance. The results suggest that the CMap expression data on the two cell lines MCF7 and PC3 are not sufficient for accurate DILI label prediction. REVIEWERS: This article was reviewed by Maciej Kandula and Pawel P. Labaj.

Evaluating reproducibility of AI algorithms in digital pathology with DAPPER.

Artificial Intelligence is exponentially increasing its impact on healthcare. As deep learning is mastering computer vision tasks, its application to digital pathology is natural, with the promise of aiding in routine reporting and standardizing results across trials. Deep learning features inferred from digital pathology scans can improve validity and robustness of current clinico-pathological features, up to identifying novel histological patterns, e.g., from tumor infiltrating lymphocytes. In this study, we examine the issue of evaluating accuracy of predictive models from deep learning features in digital pathology, as an hallmark of reproducibility. We introduce the DAPPER framework for validation based on a rigorous Data Analysis Plan derived from the FDA's MAQC project, designed to analyze causes of variability in predictive biomarkers. We apply the framework on models that identify tissue of origin on 787 Whole Slide Images from the Genotype-Tissue Expression (GTEx) project. We test three different deep learning architectures (VGG, ResNet, Inception) as feature extractors and three classifiers (a fully connected multilayer, Support Vector Machine and Random Forests) and work with four datasets (5, 10, 20 or 30 classes), for a total of 53, 000 tiles at 512 × 512 resolution. We analyze accuracy and feature stability of the machine learning classifiers, also demonstrating the need for diagnostic tests (e.g., random labels) to identify selection bias and risks for reproducibility. Further, we use the deep features from the VGG model from GTEx on the KIMIA24 dataset for identification of slide of origin (24 classes) to train a classifier on 1, 060 annotated tiles and validated on 265 unseen ones. The DAPPER software, including its deep learning pipeline and the Histological Imaging-Newsy Tiles (HINT) benchmark dataset derived from GTEx, is released as a basis for standardization and validation initiatives in AI for digital pathology.

Machine learning models for predicting endocrine disruption potential of environmental chemicals.

We introduce here ML4Tox, a framework offering Deep Learning and Support Vector Machine models to predict agonist, antagonist, and binding activities of chemical compounds, in this case for the estrogen receptor ligand-binding domain. The ML4Tox models have been developed with a 10 × 5-fold cross-validation schema on the training portion of the CERAPP ToxCast dataset, formed by 1677 chemicals, each described by 777 molecular features. On the CERAPP "All Literature" evaluation set (agonist: 6319 compounds; antagonist 6539; binding 7283), ML4Tox significantly improved sensitivity over published results on all three tasks, with agonist: 0.78 vs 0.56; antagonist: 0.69 vs 0.11; binding: 0.66 vs 0.26.