Large Datasets Research Articles

Geomechanical characterization and mineralogical correlation of compositionally diverse world-class Kazakhstani source rocks: Insights from nanoindentation testing

Extensive nanoindentation testing over a range of deflection depths of up to 4 μm yielded a large dataset, providing a viable framework for the statistical assessment of the mechanical properties, specifically elastic modulus (E) and hardness (H), of compositionally diverse organic-rich mudstone samples. The data from indentations as shallow as 300–400 nm were clustered using the k-means algorithm to identify three mechanical categories in the samples: a soft pseudophase (e.g., organic matter, gypsum, and clay minerals), a stiff pseudophase (e.g., quartz and feldspar), and a transitional composite-like pseudophase bridging the soft and hard minerals. The initially diverse values of E and H for the mechanical pseudophases were observed to converge to a constant value at indentations beyond 2–2.5 μm (varying between different samples), implying the existence of a minimal probing depth for assessing the bulk E and H of heterogeneous mudstone samples. The obtained bulk E and H values (8–21 GPa and 0.3–0.9 GPa, respectively) demonstrated a strong correlation with the mineralogical composition of the indented samples. Despite containing a notable proportion of mechanically stiff components (>45 vol%), the bulk mechanical parameters determined in this study were significantly lower than those reported for major shale formations such as the Barnett and Longmaxi Shale. This discrepancy is primarily due to the presence of organic matter with low thermal maturity (Ro < 0.6%), which constitutes <36 vol% of the samples, and a significant gypsum content, accounting for <15 vol%.The employed approach not only demonstrated the importance of choosing the proper indentation depths for investigating the mechanical properties of highly heterogeneous mudstone rocks and their constituent minerals, but it also illustrated the capability of examining various volumes of investigation using nanoindentation, approaching macroscopic values, and identifying a representative element volume (REV). The findings also provided crucial insights into the fracability and overall producibility of the investigated formations, thereby enhancing our understanding of their extraction potential.

Hypergraph-Based Multitask Feature Selection with Temporally Constrained Group Sparsity Learning on fMRI

Localizing the brain regions affected by tasks is crucial to understanding the mechanisms of brain function. However, traditional statistical analysis does not accurately identify the brain regions of interest due to factors such as sample size, task design, and statistical effects. Here, we propose a hypergraph-based multitask feature selection framework, referred to as HMTFS, which we apply to a functional magnetic resonance imaging (fMRI) dataset to extract task-related brain regions. HMTFS is characterized by its ability to construct a hypergraph through correlations between subjects, treating each subject as a node to preserve high-order information of time-varying signals. Additionally, it manages feature selection across different time windows in fMRI data as multiple tasks, facilitating time-constrained group sparse learning with a smoothness constraint. We utilize a large fMRI dataset from the Human Connectome Project (HCP) to validate the performance of HMTFS in feature selection. Experimental results demonstrate that brain regions selected by HMTFS can provide higher accuracy for downstream classification tasks compared to other competing feature selection methods and align with findings from previous neuroscience studies.

Leveraging digital footprints data for accurate estimation of the residential housing stock in the United Kingdom, 1997–2022

ABSTRACT This paper explores the use of digital footprints data to improve on the limited information available on the size and distribution of the residential housing stock in the United Kingdom over the past two decades. We use a subset of a large dataset consisting of the names and addresses of more than one billion individuals dating back to 1997 (LCRs: Linked Consumer Registers) to calibrate property lifecycle information within an authoritative geolocated address and property dataset. We validate the results of our novel calibration method against official estimates by the UK government covering the same period. We further show that our calibration method, based on individual level data, captures a much larger share of the residential housing stock since 1997 than would have been possible with other available data sources. We argue that these housing stock estimates are fundamental to the development of a comprehensive digital footprints data infrastructure that has the potential to significantly enhance population, migration, and other social statistics in the UK. More generally, this study highlights the benefits of using digital footprints data in refining, calibrating, and extending existing datasets.

Deep Learning Algorithms for Skin Disease Classification

Skin diseases are a serious concern of public health worldwide, and successful treatment needs a correct and timely diagnosis. Traditional diagnostic methods mostly depend on dermatologist’s visual observation and this leads to subjective interpretations coupled with time-consuming processes. Deep learning algorithms have lately been known as powerful means for automated medical image analysis that present more accurate and quicker results at the same time. This study analyses the usage of state-of-the-art deep learning algorithms like YOLOv8, Deep CNN, and ResNet50 used for classification of skin diseases using dermatological images. Classifying the skin conditions relies heavily on the ability to identify and extract essential features. Different skin conditions were covered under large dataset thus providing a comprehensive foundation for training and validation aimed at ensuring that the models could generalize well across different diseases. Each algorithm also employs transfer learning techniques by utilizing pre-trained models based on large image datasets in order to improve adaptability and generalization over new data types. The use of deep learning algorithms in classifying skin diseases represents a significant method to achieve efficient and accurate diagnosis with benefits to both patients and healthcare professionals as is the trend in medical image analysis. The advanced deep learning models introduced in this paper excel at classifying complex skin diseases, outperforming the machine learning approaches in performance.

Risk Factors for Perinatal Arterial Ischemic Stroke: A Machine Learning Approach.

Perinatal arterial ischemic stroke (PAIS) is a focal vascular brain injury presumed to occur between the fetal period and the first 28 days of life. It is the leading cause of hemiparetic cerebral palsy. Multiple maternal, intrapartum, delivery, and fetal factors have been associated with PAIS, but studies are limited by modest sample sizes and complex interactions between factors. Machine learning approaches use large and complex data sets to enable unbiased identification of clinical predictors but have not yet been applied to PAIS. We combined large PAIS data sets and used machine learning methods to identify clinical PAIS factors and compare this data-driven approach with previously described literature-driven clinical prediction models. Common data elements from 3 registries with patients with PAIS, the Alberta Perinatal Stroke Project, Canadian Cerebral Palsy Registry, International Pediatric Stroke Study, and a longitudinal cohort of healthy controls (Alberta Pregnancy Outcomes and Nutrition Study), were used to identify potential predictors of PAIS. Inclusion criteria were term birth and idiopathic PAIS (absence of primary causative medical condition). Data including maternal/pregnancy, intrapartum, and neonatal factors were collected between January 2003 and March 2020. Common data elements were entered into a validated random forest machine learning pipeline to identify the highest predictive features and develop a predictive model. Univariable analyses were completed post hoc to assess the relationship between each predictor and outcome. A machine learning model was developed using data from 2,571 neonates, including 527 cases (20%) and 2,044 controls (80%). With a mean of 21 features selected, the random forest machine learning approach predicted the outcome with approximately 86.5% balanced accuracy. Factors that were selected a priori through literature-driven variable selection that were also identified as most important by the machine learning model were maternal age, recreational substance exposure, tobacco exposure, intrapartum maternal fever, and low Apgar score at 5 minutes. Additional variables identified through machine learning included in utero alcohol exposure, infertility, miscarriage, primigravida, meconium, spontaneous vaginal delivery, neonatal head circumference, and 1-minute Apgar score. Overall, the machine learning model performed better (area under the curve [AUC] 0.93) than the literature-driven model (AUC 0.73). Machine learning may be an alternative, unbiased method to identify clinical predictors associated with PAIS. Identification of previously suggested and novel clinical factors requires cautious interpretation but supports the multifactorial nature of PAIS pathophysiology. Our results suggest that identification of neonates at risk of PAIS is possible.

The distribution of velocity and sediment concentration under currents in coastal seas

A power law velocity profile for engineering application was obtained based on the field data set of flow velocity and sediment concentration in coastal areas of Jinzhou, Nantong, and Zhoushan. Using theoretical analyses, van Rijn's formula for sediment concentration profile was modified. Tests of the new power-law velocity profile and the modified formula of suspended sediment concentration against a large data set were performed. The results show a good fit to the data and that these two formulae could demonstrate the basic principles of vertical distribution of flow velocity and suspended sediment concentration, respectively.

Phenotyping Hepatic Immune-Related Adverse Events in the Setting of Immune Checkpoint Inhibitor Therapy.

We present and validate a rule-based algorithm for the detection of moderate to severe liver-related immune-related adverse events (irAEs) in a real-world patient cohort. The algorithm can be applied to studies of irAEs in large data sets. We developed a set of criteria to define hepatic irAEs. The criteria include: the temporality of elevated laboratory measurements in the first 2-14 weeks of immune checkpoint inhibitor (ICI) treatment, steroid intervention within 2 weeks of the onset of elevated laboratory measurements, and intervention with a duration of at least 2 weeks. These criteria are based on the kinetics of patients who experienced moderate to severe hepatotoxicity (Common Terminology Criteria for Adverse Events grades 2-4). We applied these criteria to a retrospective cohort of 682 patients diagnosed with hepatocellular carcinoma and treated with ICI. All patients were required to have baseline laboratory measurements before and after the initiation of ICI. A set of 63 equally sampled patients were reviewed by two blinded, clinical adjudicators. Disagreements were reviewed and consensus was taken to be the ground truth. Of these, 25 patients with irAEs were identified, 16 were determined to be hepatic irAEs, 36 patients were nonadverse events, and two patients were of indeterminant status. Reviewers agreed in 44 of 63 patients, including 19 patients with irAEs (0.70 concordance, Fleiss' kappa: 0.43). By comparison, the algorithm achieved a sensitivity and specificity of identifying hepatic irAEs of 0.63 and 0.81, respectively, with a test efficiency (percent correctly classified) of 0.78 and outcome-weighted F1 score of 0.74. The algorithm achieves greater concordance with the ground truth than either individual clinical adjudicator for the detection of irAEs.

Projected fuzzy [formula omitted]-means clustering algorithm with instance penalty

At present, high-dimensional data clustering has become a vital research field in machine learning. Traditional clustering algorithms cannot perform well on high-dimensional data, where the clustering task is usually divided into two stages: dimensionality reduction first and clustering later. In general, the existing high-dimensional clustering methods usually have the following shortcomings: (1) the two-stage strategy splits the connection between clustering and dimensionality reduction; (2) these algorithms do not consider the impact of anomalous instances in high-dimensional data on clustering performance. Therefore, to address these problems, a projected fuzzy c-means clustering algorithm with instance penalty (PCIP) is proposed. Firstly, we construct an instance penalty matrix and assign an instance penalty coefficient to each sample. Secondly, a model for clustering high-dimensional data is constructed by integrating fuzzy c-means clustering (FCM) and principal component analysis (PCA). The proposed model can perform dimensionality reduction and clustering simultaneously. In addition, the time complexity of the proposed algorithm is linearly related to the number of samples n, which can efficiently deal with large data sets. The proposed PCIP algorithm is verified by experiments using clustering accuracy and normalized mutual information (NMI) as evaluation metrics. The experimental results on 10 image datasets show that the average accuracy and average NMI of the PCIP algorithm are improved by 0.0375 and 0.0275, respectively, compared to the second-ranked algorithm.

Adjoint Synthesis for Trans‐Oceanic Tsunami Waveforms and Simultaneous Inversion of Fault Geometry and Slip Distribution

AbstractTsunamis propagate over long distances and can cause widespread damage even after crossing ocean basins. Prediction of tsunamis in distant areas based on observations near their sources is critical to mitigating damage. In recent years, the accuracy of numerical models of trans‐oceanic tsunami propagation has improved significantly due to the incorporation of effects such as the solid earth response to tsunami loading and wave dispersion. However, these models are computationally expensive and have not been fully utilized for real‐time prediction. Here, we derive the adjoint operator for the linear set of equations describing deep‐ocean tsunami propagation and show how a pre‐computed database of adjoint states can achieve rapid synthesis of tsunami waveforms at target sites from nonpoint arbitrary tsunami sources. The adjoint synthesis method allows for an exhaustive parameter search for tsunami source estimation. A method for simultaneous inversion of fault geometry and slip distribution using adjoint synthesis with Sequential Monte Carlo method was proposed and applied to the 2012 Haida Gwaii earthquake tsunami. The influence of model accuracy and the amount of observed data on the estimation of tsunami sources and waveforms was examined. It was found that with a highly accurate propagation model, using only a limited amount of observed data produced source and waveform estimates very similar to the final models obtained with much larger data sets. The final inferred fault model involved megathrust slip distributed between the Haida Gwaii trench and the Queen Charlotte fault. The proposed method can also quantify the uncertainty of the waveform forecasts.

Mutational landscape of HER-2 low breast cancer.

e13040 Background: HER2-low breast cancer is defined as HER2 IHC score of 1+ or 2+/ISH- tumors. Anti-Her-2 therapies such as trastuzumab deruxtecan improve progression-free survival (PFS) and overall survival (OS) in patients with HER-2 low metastatic breast cancer. However, the mutational landscape of HER2-low BC is still not well defined. The aim of this study is to characterize the biological and clinicopathological landscape of HER2-low BC by next generation sequencing. Methods: We conducted a single institution, retrospective chart review of patients with metastatic breast cancer (MBC) from 1/2020 to 4/2023. We obtained Next generation sequencing data and correlated it with clinical outcomes. Patients with HER2 IHC score of 1+ or 2+/ISH- tumors referred as HER2-low, while with no HER2 IHC staining at all were referred as HER2-zero, and pts with HER2 IHC 3+ or HER2 IHC 2+/ISH+ referred as HER2+. Results: A total of 98 pts were enrolled:45 pts with HER2-low tumors, 29 pts with HER2+ and 24 pts with HER2-zero tumors. In the HER2-low cohort, 58% (26 pts) were initially diagnosed with loco regional disease while 42% presented with stage IV at diagnosis. On average patients were on 3 lines of therapy for metastatic disease (ranging between 1 to 12). Of those 35% received T-DXd. High Ki-67 proliferation (≥20), reported in 58% (26 pts). OS was 54.4 months, 95% CI [ 42.5,100.2]. In the HER2+ cohort, 31% of pts presented with stage IV at diagnosis. On average pts received 3 lines of therapy (ranging from 1 to 14). 69% of pts received T-DXd. High Ki-67 reported in 48% of pts. OS was 59.2 months, 95% CI [39.2,96,1]. In the HER2-zero, 62.5% of pts presented with stage IV. On average pts received 1 line of therapy (ranging between 1 to 5), 62% of pts had high Ki-67. OS of 41.6 months 95% CI [18.4, 77.9]. Conclusions: Although limited by the sample size, our data suggests that HER-2 low breast cancer tends to be more locally advanced or metastatic at the time of presentation. HER-2 low BC is associated with a higher Ki-67 index and tends to harbor more mutations. HER-2 low may not just be a target, but a unique biological entity and can be distinguished from HER2-Zero BC with higher prevalence of HR+ and other biomarkers and better OS. NGS data from a larger data set is required to better characterize HER-2 low breast cancer. We suggest that clinicians obtain NGS panel in all patients with HER-2 low MBC. [Table: see text]

Identification of dental implant systems from low-quality and distorted dental radiographs using AI trained on a large multi-center dataset

Most artificial intelligence (AI) studies have attempted to identify dental implant systems (DISs) while excluding low-quality and distorted dental radiographs, limiting their actual clinical use. This study aimed to evaluate the effectiveness of an AI model, trained on a large and multi-center dataset, in identifying different types of DIS in low-quality and distorted dental radiographs. Based on the fine-tuned pre-trained ResNet-50 algorithm, 156,965 panoramic and periapical radiological images were used as training and validation datasets, and 530 low-quality and distorted images of four types (including those not perpendicular to the axis of the fixture, radiation overexposure, cut off the apex of the fixture, and containing foreign bodies) were used as test datasets. Moreover, the accuracy performance of low-quality and distorted DIS classification was compared using AI and five periodontists. Based on a test dataset, the performance evaluation of the AI model achieved accuracy, precision, recall, and F1 score metrics of 95.05%, 95.91%, 92.49%, and 94.17%, respectively. However, five periodontists performed the classification of nine types of DISs based on four different types of low-quality and distorted radiographs, achieving a mean overall accuracy of 37.2 ± 29.0%. Within the limitations of this study, AI demonstrated superior accuracy in identifying DIS from low-quality or distorted radiographs, outperforming dental professionals in classification tasks. However, for actual clinical application of AI, extensive standardization research on low-quality and distorted radiographic images is essential.

Deep analysis of total serum N-glycome suggests glyco-signatures for phospholipase A2 receptor 1-related idiopathic membranous nephropathy diagnosis

Idiopathic membranous nephropathy (IMN) is an antibody-mediated and kidney-specific autoimmune disease, with the antigen phospholipase A2 receptor 1 (PLA2R1) accounting for approximately 70% of IMN cases. Although a variety of new podocyte target antigens and their autoantibodies have been identified, they are still of limited diagnostic and therapeutic value due to lack of high specificity and sensitivity. N-glycans play vital roles in renal system and their pathobiological relevance has become increasingly recognized in many kidney diseases, but not fully explored in IMN. To find possible glyco-signatures for PLA2R1-related IMN diagnosis, we herein established a comprehensive workflow for total serum N-glycome analysis based on our recently developed mass spectrometry (MS)-based N-glycan purification method, named Ultrafast Glycoprotein Immobilization for Glycan extraction (UltraGIG). A total of 191 N-glycans were identified from IMN patients, representing the largest N-glycome dataset in IMN. Compared to healthy controls, up-regulation of sialylation and core-fucosylation as well as down-regulation of galactosylation were observed in PLA2R1-positive IMN patients, and up-regulation of hyper-galactosylation was specific for PLA2R1-negative IMN patients. A six-glycan marker panel consisting of H4N3S1, H4N3F1, H6N4S2, H6H5F1S2, H6N5 and H6N6F1S1, was proposed to aid in the accurate diagnosis of PLA2R1-related IMN, which provided new insights into IMN biomarker study. SignificancePLA2R1-related IMN is a kidney-specific autoimmune disease with a high risk of developing end-stage renal disease (ESRD) and even kidney failure. Current biomarkers are still of limited diagnostic and therapeutic value due to lack of high specificity and sensitivity. An in-depth MS analysis of total serum N-glycome of PLA2R1-related IMN patients was conducted for the first time. We generated the largest dataset of serum N-glycome for IMN to date, and proposed a novel six-glycan marker panel that may help the accurate diagnosis of PLA2R1-related IMN.

A comparative study of YOLOv5 and YOLOv8 for corrosion segmentation tasks in metal surfaces

This study delves into the comparative efficacy of YOLOv5 and YOLOv8 in corrosion segmentation tasks. We employed three unique datasets, comprising 4942, 5501, and 6136 images, aiming to thoroughly evaluate the models’ adaptability and robustness in diverse scenarios. The assessment metrics included precision, recall, F1-score, and mean average precision. Furthermore, graphical tests offered a visual perspective on the segmentation capabilities of each architecture. Our results highlight YOLOv8’s superior speed and segmentation accuracy across datasets, further corroborated by graphical evaluations. These visual assessments were instrumental in emphasizing YOLOv8’s proficiency in handling complex corroded surfaces. However, in the largest dataset, both models encountered challenges, particularly with overlapping bounding boxes. YOLOv5 notably lagged, struggling to achieve the performance standards set by YOLOv8, especially with irregular corroded surfaces. In conclusion, our findings underscore YOLOv8’s enhanced capabilities, establishing it as a preferable choice for real-world corrosion detection tasks. This research thus offers invaluable insights, poised to redefine corrosion management strategies and guide future explorations in corrosion identification.

Identifying mining-induced chromium contamination in soil through visible-near infrared spectroscopy and machine learning

The extraction, purification, and utilization of mineral resources have been among the largest anthropogenic sources of chromium (Cr) in soil. Determining Cr contamination in soil is a key issue prior to its appropriate remediation. Nevertheless, the efficient identification of large-scale soil Cr contamination requires continuous research. The present study proposes a continental-scale method to rapidly identify soil Cr contamination using visible-near infrared spectroscopy (vis-NIR) and machine learning (ML). A large dataset containing 18,675 topsoil samples from the Land Use/Land Cover Area Frame Survey 2009 projects across Europe was compiled. Five advanced ML algorithms were compared, and hyperparameter optimization was conducted using the grid search method. Permutation importance was employed to calculate the rank of each spectral wavelength, shedding light on the most sensitive spectral wavelength for Cr contamination. Results indicate that hyperparameter optimization had the most significant performance improvement on support vector machine (SVM), exhibiting an increase in training performance from 0.795 to 0.868. The achieved optimal SVM accuracy, area under the receiver operating feature curve, sensitivity, and specificity of 0.78, 0.85, 0.85, and 0.66, respectively, indicating excellent predictive performance on the Cr contamination classification. The optimal SVM model revealed that the most important spectral band for classifying Cr contamination was 1430–1433 nm. This finding implies that the adsorption of molecular water was closely related to the classification of Cr contamination. The current study introduces the first continental-scale identification of Cr contamination using vis-NIR, which has excellent guiding significance for Cr remediation and the identification of other heavy metals using vis-NIR.

Kernel rootkit detection multi class on deep learning techniques

The harmful code application known as a rootkit is designed to be loaded and run directly from the operating system's (OSs') Kernel. Rootkits deployed in the Kernel, called Kernel-mode rootkits, can alter the OS. The intention behind these Kernel changes is to conceal the hack. Detecting a Kernel rootkit in a target machine is found to be quite challenging. Numerous techniques can be employed to modify the Kernel of a system. Kernel rootkits also create hidden access for attacks, enabling unauthorized entry to be gained by attackers on the machine. The ultimate consequence is that essential computer data can be modified, personal information can be gathered, and hackers can observe behavior. Synthetic neural networks support artificial intelligence, a branch of deep learning that models the human brain and operates on large datasets. This study proposed the Kernel rootkit detection multi-class deep learning techniques (KRDMCDLT). Deep learning algorithms are utilized to recognize the Kernel rootkit from a batch of data by selecting essential properties for learning tracking models. Thus, by identifying the OS malware, trojan assaults can be stopped before they can access infected data. This Kernel rootkit detection was tested in a Google Cloud Platform (GCP) computing system.

Tensor Golub–Kahan method based on Einstein product

The Singular Value Decomposition (SVD) of matrices is a widely used tool in scientific computing. In many applications of machine learning, data analysis, signal and image processing, the large datasets are structured into tensors, for which generalizations of SVD have already been introduced, for various types of tensor–tensor products. In this paper, we present innovative methods for approximating this generalization of SVD to tensors in the framework of the Einstein tensor product. These singular elements are called singular values and singular tensors respectively. The proposed method uses the tensor Lanczos bidiagonalization applied to the Einstein product. In most applications, as in the matrix case, the extremal singular values are of special interest. To enhance the approximation of the largest or the smallest singular triplets (singular values and left and right singular tensors), a restarted method based on Ritz augmentation is proposed. Numerical results are proposed to illustrate the effectiveness of the presented method. In addition, applications to video compression and facial recognition are presented.

Unveiling compositional images of specific proteins in individual cells by LA-ICP-MS: labelling with ruthenium red and metal nanoclusters

Unveiling compositional images of specific proteins in individual cells by LA-ICP-MS: labelling with ruthenium red and metal nanoclusters

Synthetic Optical Coherence Tomography Angiographs for Detailed Retinal Vessel Segmentation Without Human Annotations.

Optical coherence tomography angiography (OCTA) is a non-invasive imaging modality that can acquire high-resolution volumes of the retinal vasculature and aid the diagnosis of ocular, neurological and cardiac diseases. Segmenting the visible blood vessels is a common first step when extracting quantitative biomarkers from these images. Classical segmentation algorithms based on thresholding are strongly affected by image artifacts and limited signal-to-noise ratio. The use of modern, deep learning-based segmentation methods has been inhibited by a lack of large datasets with detailed annotations of the blood vessels. To address this issue, recent work has employed transfer learning, where a segmentation network is trained on synthetic OCTA images and is then applied to real data. However, the previously proposed simulations fail to faithfully model the retinal vasculature and do not provide effective domain adaptation. Because of this, current methods are unable to fully segment the retinal vasculature, in particular the smallest capillaries. In this work, we present a lightweight simulation of the retinal vascular network based on space colonization for faster and more realistic OCTA synthesis. We then introduce three contrast adaptation pipelines to decrease the domain gap between real and artificial images. We demonstrate the superior segmentation performance of our approach in extensive quantitative and qualitative experiments on three public datasets that compare our method to traditional computer vision algorithms and supervised training using human annotations. Finally, we make our entire pipeline publicly available, including the source code, pretrained models, and a large dataset of synthetic OCTA images.

A digital pathology AI model to predict immune-oncology biomarker status and pembrolizumab response in a real-world cohort of patients with colorectal and breast cancer.

e13617 Background: Deep learning (DL) algorithms have garnered significant attention for prediction of biomarker status in digitized tissue slides. Here, we trained and validated a digital pathology algorithm to predict microsatellite instability (MSI)/mismatch repair deficiency (MMRd) in colorectal cancer (CRC) and PD-L1 status in breast cancer (BC) using a large real-world patient cohort. Methods: H&amp;E and immunohistochemistry (IHC) slides of CRC and BC from Caris Life Sciences (Phoenix, AZ) with known biomarker status were digitally scanned (CRC: N=27,494, 6.6% MSI-high by next-generation sequencing [NGS]; N=25,788, 6.4% MMRd by IHC; BC: N=16,069, 15.8% PD-L1+ by IHC). A transformer-based DL model (Wagner, Sophia J., et al., 2023) was trained to predict biomarker status, with 5-fold cross-validation. Accuracy, F1 score, area under the curve (AUC), and other metrics were calculated comparing the model to NGS/IHC “truth”. Risk stratification was performed on holdout datasets of pembrolizumab-treated CRC (N=422-454) and triple negative breast cancer (TNBC; N=513) pts. Post-treatment overall survival (OS) was calculated from start of pembrolizumab to last contact, per insurance claims data. Hazard ratios (HR) and 95% confidence intervals (CI) were calculated by Cox-Proportional hazards and compared between model and “truth”. Results: When applied to holdout datasets for biomarker prediction, the validated model achieved AUC of 0.78-0.95. In CRC and BC, performance was better among primary specimens than metastatic. For MSI-H/MMRd risk stratification in pembrolizumab-treated CRC, HR for OS were comparable between model and “truth” for primary and metastatic cohorts. When applied to the TNBC holdout dataset, the model performed better than “truth” for PD-L1 risk stratification in pembrolizumab-treated pts. Conclusions: A digital pathology DL model predicts biomarker status with high accuracy and is non-inferior to NGS/IHC for risk stratification of pembrolizumab-treated CRC and TNBC pts. These results support further investigation of AI algorithms developed from large datasets as possible rapid screening tools for biomarker detection in molecular testing programs. [Table: see text]

Multitask Learning for Joint Diagnosis of Multiple Mental Disorders in Resting-State fMRI.

Facing the increasing worldwide prevalence of mental disorders, the symptom-based diagnostic criteria struggle to address the urgent public health concern due to the global shortfall in well-qualified professionals. Thanks to the recent advances in neuroimaging techniques, functional magnetic resonance imaging (fMRI) has surfaced as a new solution to characterize neuropathological biomarkers for detecting functional connectivity (FC) anomalies in mental disorders. However, the existing computer-aided diagnosis models for fMRI analysis suffer from unstable performance on large datasets. To address this issue, we propose an efficient multitask learning (MTL) framework for joint diagnosis of multiple mental disorders using resting-state fMRI data. A novel multiobjective evolutionary clustering algorithm is presented to group regions of interests (ROIs) into different clusters for FC pattern analysis. On the optimal clustering solution, the multicluster multigate mixture-of-expert model is used for the final classification by capturing the highly consistent feature patterns among related diagnostic tasks. Extensive simulation experiments demonstrate that the performance of the proposed framework is superior to that of the other state-of-the-art methods. Moreover, the potential for practical application of the framework is also validated in terms of limited computational resources, real-time analysis, and insufficient training data. The proposed model can identify the remarkable interpretative biomarkers associated with specific mental disorders for clinical interpretation analysis.