Single Classifier Research Articles

Abstract 5059: Robust single sample consensus molecular subtype classification for primary and metastatic colorectal cancer

Abstract Introduction Consensus Molecular Subtypes (CMS) represent a well-established molecular stratification framework for colorectal cancer (CRC). Existing methods for CMS classification rely on a representative input cohort as a preprocessing step (CMScaller) or have difficulty generalizing to metastatic samples (CMSclassifier). We developed Tempus CMS to overcome these limitations. Our method normalizes gene expression data from input samples to a static reference cohort to enable single sample classification of both primary and metastatic tumors. We evaluated the performance of this classifier on a large, de-identified CRC cohort comprising 8,489 samples from primary and metastatic sites. Methods To normalize input data, our method shifts and scales each gene expression value based on the mean and standard deviation of the expression of that gene in the reference cohort (n=2,787 primary and metastatic CRC). CMS calls are then generated via nearest template prediction similar to CMScaller (Eide et al., 2017). The analysis cohort was selected from clinical biopsies within 30 days of primary or metastatic diagnosis excluding reference cohort samples. CMS calls were assessed by comparing subtype prevalence to reported rates and by testing for known enrichments of pathways and genomic markers. Results Subtype prevalences for lower gastrointestinal (GI) biopsies and resections (n=5,090) were comparable to reported rates for primary CRC tumors: CMS1 - 12%, CMS2 - 28%, CMS3 - 19%, and CMS4 - 33% with 7.6% no calls. Biopsies from metastatic tissue showed site-specific changes in prevalence: CMS2 increased to 43% in liver samples (n=1,715) and CMS4 was identified in 49% of samples from metastatic sites excluding the liver (n=1,684). In addition, we found that purity had an effect on prevalence: in lower GI tumors with ≥ 50% tumor purity rates of CMS2 and CMS4 were 34% and 26%, respectively, while rates in tumors with purity between 20% and 50% were 20% CMS2 and 41% CMS4. Across all tissue groups, pathway enrichment showed significant associations, including TGF-ꞵ signaling in CMS4 (adj P ≤ 4.4e-15) and cell differentiation in CMS3 (adj P ≤ 3.0e-27). Expected enrichments of CMS markers were also observed: BRAF mutations and microsatellite instability in CMS1 and KRAS mutations in CMS3. In liver tissue, higher KRAS mutation rates were noted in CMS1 (57%) and CMS3 (83%) compared to GI tissue (35% and 67%, respectively). Conclusions In this study, we developed an enhanced, single sample CMS-calling algorithm optimized for primary and metastatic sites. Application of this algorithm in a large cohort recapitulated known biology, which suggests that the tool can be used to support clinical studies requiring robust molecular stratification. Citation Format: Andrew J. Sedgewick, Tushar Chandra, Timothy Taxter, Justin Guinney. Robust single sample consensus molecular subtype classification for primary and metastatic colorectal cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 5059.

Abstract 3637: Ultrasensitive assays for combined detection of CEA+ extracellular vesicles and soluble CEA

Abstract Colorectal cancer (CRC) is the third most common cancer worldwide and is the second leading cause of cancer death in the United States. Carcinoembryonic antigen (CEA) is a biomarker commonly used in the management of CRC. It is typically present at low levels in the blood of healthy adults but is elevated in CRC samples and some other cancers. CEA is most commonly used to monitor treatment efficacy and as a critical part of post-treatment surveillance for residual disease, recurrence, and/or progression/metastasis. The CRC recurrence rate averages >25%, and liver metastasis occurs in 25-30% of cases. Improved CEA assays could enable earlier determination of treatment failure, detection of lower levels of minimal residual disease (MRD) after treatment, or earlier detection of recurrence. CEA is a GPI-linked membrane protein secreted from CRC cells mainly by enzymatic cleavage of its GPI anchor. However, we have identified alternative secretion of CEA in association with extracellular vesicles (EVs). We hypothesized that the soluble CEA produced by canonical shedding and the alternatively secreted CEA may provide distinct information about the tumor or its microenvironment and serve as independent but complementary biomarkers of CRC. We developed ultrasensitive, multi-marker electrochemiluminescence (ECL) immunoassays for measuring intact EVs presenting CEA, CD73, or both in human plasma. We observed that plasma levels of CEA+EVs and CD73+EVs were elevated in some late-stage CRC samples compared to controls. CEA and CD73 double-positive EV (CEA+CD73+EVs) levels were also elevated in the same samples, exhibiting a 3-fold improved separation of these samples with CRC from controls, compared to detection of CEA+EVs. Additionally, approximately 25% of stage IV CRC samples had elevated CEA+EV and CEA+CD73+EV levels while exhibiting low soluble CEA. We confirmed this result in an independent sample set that included samples across all CRC stages, and also identified a few earlier-stage samples that exhibited high CEA+EVs and CEA+CD73+EVs without elevated soluble CEA. Since CEA+CD73+EVs and soluble CEA assays identified different subsets of samples, we combined the two assays into a single classifier for late stage CRC and controls, yielding improved receiver operator characteristic (ROC) with area under the curve (AUC) of 0.975 versus 0.927 for the standard soluble CEA assay or 0.788 for CEA+CD73+EVs alone. These results suggest that assaying EVs presenting multiple surface markers can increase specificity for the detection of tumor-derived EVs relative to a single marker and that assaying both soluble and EV-associated forms of CEA in plasma may improve the value of CEA measurement in CRC. Citation Format: Lucie Hebert, Misk Al-Ameen, Talis Spriggs, Collin Nelson, Evan A. Gizzie, Jeffrey L. Franklin, Robert J. Coffey, David A. Routenberg, Jacob N. Wohlstadter. Ultrasensitive assays for combined detection of CEA+ extracellular vesicles and soluble CEA [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 3637.

Improving Adversarial Robustness of Ensemble Classifiers by Diversified Feature Selection and Stochastic Aggregation

Learning-based classifiers are found to be vulnerable to attacks by adversarial samples. Some works suggested that ensemble classifiers tend to be more robust than single classifiers against evasion attacks. However, recent studies have shown that this is not necessarily the case under more realistic settings of black-box attacks. In this paper, we propose a novel ensemble approach to improve the robustness of classifiers against evasion attacks by using diversified feature selection and a stochastic aggregation strategy. Our proposed scheme includes three stages. Firstly, the adversarial feature selection algorithm is used to select a feature each time that can trade-offbetween classification accuracy and robustness, and add it to the feature vector bank. Secondly, each feature vector in the bank is used to train a base classifier and is added to the base classifier bank. Finally, m classifiers from the classifier bank are randomly selected for decision-making. In this way, it can cause each classifier in the base classifier bank to have good performance in terms of classification accuracy and robustness, and it also makes it difficult to estimate the gradients of the ensemble accurately. Thus, the robustness of classifiers can be improved without reducing the classification accuracy. Experiments performed using both Linear and Kernel SVMs on genuine datasets for spam filtering, malware detection, and handwritten digit recognition demonstrate that our proposed approach significantly improves the classifiers’ robustness against evasion attacks.

An efficient fake news classification model based on ensemble deep learning techniques

The availability and expansion of social media has made it difficult to distinguish between fake and real news. Information falsification has exponentially increased as a result of how simple it is to spread information through sharing. Social media dependability is also under jeopardy due to the extensive dissemination of false information. Therefore, it has become a research problem to automatically validate information, specifically source, content, and publisher, to identify it as true or false. Despite its limitations, machine learning (ML) has been crucial in the categorization of information. Previous studies suggested three-step methods for categorising false information on social media. In the first step of the process, the data set is subjected to a number of pre-processing processes in order to transform unstructured data sets into structured data sets. The unknowable properties of fake news and the features are extracted by the Lexicon Model in the second stage. In the third stage of this research project, a feature selection method by WOA (Whale Optimization Algorithm) for weight value to tune the classification part. Finally, a Hybrid Classification model that is hybrid with a fuzzy based Convolutional Neural Network and kernel based support vector machine is constructed in order to identify the data pertaining to bogus news. However using single classifier for fake news detection produces the insufficient accuracy. To overcome this issue in this work introduced an improved model for fake news classification. To turn unstructured data sets into structured data sets, a variety of pre-processing operations are used on the data set in the initial phase of the procedure. The unknowable properties of fake news and the features are extracted by the Lexicon Model in the second stage. In the third stage of this research project, a feature selection method by COA (Coati Optimization Algorithm) for weight value to tune the classification part. Finally, an ensemble of RNN (Recurrent Neural Networks), VGG-16 and ResNet50.A classification model was developed to recognise bogus news information. Evaluate each fake news analysis' performance in terms of accuracy, precision, recall, and F1 score. The suggested model, out of all the methodologies taken into consideration in this study, provides the highest outcomes, according to experimental findings

Improving Continuous Sign Language Recognition with Consistency Constraints and Signer Removal

Deep-learning-based continuous sign language recognition (CSLR) models typically consist of a visual module, a sequential module, and an alignment module. However, the effectiveness of training such CSLR backbones is hindered by limited training samples, rendering the use of a single connectionist temporal classification loss insufficient. To address this limitation, we propose three auxiliary tasks to enhance CSLR backbones. First, we enhance the visual module, which is particularly sensitive to the challenges posed by limited training samples, from the perspective of consistency. Specifically, since sign languages primarily rely on signers’ facial expressions and hand movements to convey information, we develop a keypoint-guided spatial attention module that directs the visual module to focus on informative regions, thereby ensuring spatial attention consistency. Furthermore, recognizing that the output features of both the visual and sequential modules represent the same sentence, we leverage this prior knowledge to better exploit the power of the backbone. We impose a sentence embedding consistency constraint between the visual and sequential modules, enhancing the representation power of both features. The resulting CSLR model, referred to as consistency-enhanced CSLR, demonstrates superior performance on signer-dependent datasets, where all signers appear during both training and testing. To enhance its robustness for the signer-independent setting, we propose a signer removal module based on feature disentanglement, effectively eliminating signer-specific information from the backbone. To validate the effectiveness of the proposed auxiliary tasks, we conduct extensive ablation studies. Notably, utilizing a transformer-based backbone, our model achieves state-of-the-art or competitive performance on five benchmarks, including PHOENIX-2014, PHOENIX-2014-T, PHOENIX-2014-SI, CSL, and CSL-Daily. Code and models are available at https://github.com/2000ZRL/LCSA_C2SLR_SRM.

An Empirical Evaluation of a Novel Ensemble Deep Neural Network Model and Explainable AI for Accurate Segmentation and Classification of Ovarian Tumors Using CT Images.

Ovarian cancer is one of the leading causes of death worldwide among the female population. Early diagnosis is crucial for patient treatment. In this work, our main objective is to accurately detect and classify ovarian cancer. To achieve this, two datasets are considered: CT scan images of patients with cancer and those without, and biomarker (clinical parameters) data from all patients. We propose an ensemble deep neural network model and an ensemble machine learning model for the automatic binary classification of ovarian CT scan images and biomarker data. The proposed model incorporates four convolutional neural network models: VGG16, ResNet 152, Inception V3, and DenseNet 101, with transformers applied for feature extraction. These extracted features are fed into our proposed ensemble multi-layer perceptron model for classification. Preprocessing and CNN tuning techniques such as hyperparameter optimization, data augmentation, and fine-tuning are utilized during model training. Our ensemble model outperforms single classifiers and machine learning algorithms, achieving a mean accuracy of 98.96%, a precision of 97.44%, and an F1-score of 98.7%. We compared these results with those obtained using features extracted by the UNet model, followed by classification with our ensemble model. The transformer demonstrated superior performance in feature extraction over the UNet, with a mean Dice score and mean Jaccard score of 0.98 and 0.97, respectively, and standard deviations of 0.04 and 0.06 for benign tumors and 0.99 and 0.98 with standard deviations of 0.01 for malignant tumors. For the biomarker data, the combination of five machine learning models-KNN, logistic regression, SVM, decision tree, and random forest-resulted in an improved accuracy of 92.8% compared to single classifiers.

Ensembles of quantum classifiers

In the current era, known as Noisy Intermediate-Scale Quantum (NISQ), encoding large amounts of data in the quantum devices is challenging and the impact of noise significantly affects the quality of the obtained results. A viable approach for the execution of quantum classification algorithms is the introduction of a well-known machine learning paradigm, namely, the ensemble methods. Indeed, the ensembles combine multiple internal classifiers, which are characterized by compact sizes due to the smaller data subsets used for training, to achieve more accurate and robust prediction performance. In this way, it is possible to reduce the qubit requirements with respect to a single larger classifier while achieving comparable or improved performance. In this work, we present an implementation and an extensive empirical evaluation of ensembles of quantum instance-based classifiers for binary classification, with the purpose of providing insights into their effectiveness, limitations, and potential for enhancing the performance of basic quantum models. In particular, three classical ensemble methods and three quantum instance-based classifiers have been taken into account here. Hence, the scheme that has been implemented (in Python) has a hybrid nature. The results (obtained on real-world datasets) have shown an accuracy advantage for the ensemble techniques with respect to the single quantum classifiers, and also an improvement in robustness. In fact, ensembles have proven effective not only in mitigating unsuitable data normalizations but also in reducing the impact of noise on quantum classifiers, enhancing their stability.

Neuromorphic auditory classification based on a single dynamical electrochemical memristor

Designing compact computing hardware and systems is highly desired for resource-restricted edge computing applications. Utilizing the rich dynamics in a physical device for computing is a unique approach in creating complex functionalities with miniaturized footprint. In this work, we developed a dynamical electrochemical memristor from a static memristor by replacing the gate material. The dynamical device possessed short-term fading dynamics and exhibited distinct frequency-dependent responses to varying input signals, enabling its use as a single device-based frequency classifier. Simulation showed that the device responses to different frequency components in a mixed-frequency signal were additive with nonlinear attenuation at higher frequency, providing a guideline in designing the system to process complex signals. We used a rate-coding scheme to convert real world auditory recordings into fixed amplitude spike trains to decouple amplitude-based information and frequency-based information and was able to demonstrate auditory classification of different animals. The work provides a new building block for temporal information processing.

Surface and Underwater Acoustic Source Recognition Using Array Feature Extraction Based on Machine Learning

Sound source recognition is an important application of passive sonar. How to distinguish between surface and underwater sound sources is always a difficult problem. In order to solve the problem of S/U sound source identification, this paper proposes a machine learning method based on array feature extraction, which has certain innovative value. Firstly, according to the experimental environment of SACLANT 1993, simulation data is generated based on KRAKEN. Secondly, the simulation data and experimental data are used to extract the array features. Thirdly, the accuracy, recall rate, F1 and accuracy of GBDT classifier in three different frequency bands are evaluated. The results show that the training model established by using the array feature extraction method can effectively solve the problem of poor accuracy of some channels in the single channel classification, and obtain good experimental accuracy. Finally, the experimental accuracy of the three bandwidths is 0.9920, 0.9857 and 0.9713, respectively.

A joint autoencoder and classifier deep neural network for AD and MCI classification

AbstractIn this article, we present a new approach to distinguish progressive mild cognitively impaired (pMCI) subjects, who eventually develop Alzheimer's disease (AD) from stable MCI (sMCI) subjects whose situation does not deteriorate into AD. The proposed approach combines the discriminating capabilities of classifiers and representation learning capacities of autoencoders into a unified architecture, and is hence termed as joint autoencoder and classifier deep neural network (JACDNN). JACDNN employs a single classifier and multiple autoencoders that are trained together to perform pattern classification. The classifier in JACDNN is trained using standard approaches to distinguish between subject from different classes using the binary cross entropy loss. The autoencoders in JACDNN, regularizes individual layers in the network used for classification to learn representations useful for reconstructing a given input. The performance of JACDNN has been evaluated on several machine learning problems pertaining to dementia, namely AD versus cognitively normal (CN) subjects, AD versus sMCI, CN versus pMCI, and pMCI versus sMCI. These problems are targeted using two datasets. The first dataset consist of gray matter (GM) features of subjects and the second dataset consist of combination of GM and white matter (WM) features. It is observed that better classification results are obtained when the classifier is built on GM and WM as compared with GM features alone. Performance comparison of JACDNN with other existing approaches has been conducted for these problems. The results clearly indicate that JACDNN performs better than other existing approaches for these problems.

Machine Learning of Plasma Proteomics Classifies Diagnosis of Interstitial Lung Disease.

Rationale: Distinguishing connective tissue disease-associated interstitial lung disease (CTD-ILD) from idiopathic pulmonary fibrosis (IPF) can be clinically challenging. Objectives: To identify proteins that separate and classify patients with CTD-ILD and those with IPF. Methods: Four registries with 1,247 patients with IPF and 352 patients with CTD-ILD were included in analyses. Plasma samples were subjected to high-throughput proteomics assays. Protein features were prioritized using recursive feature elimination to construct a proteomic classifier. Multiple machine learning models, including support vector machine, LASSO (least absolute shrinkage and selection operator) regression, random forest, and imbalanced Random Forest, were trained and tested in independent cohorts. The validated models were used to classify each case iteratively in external datasets. Measurements and Main Results: A classifier with 37 proteins (proteomic classifier 37 [PC37]) was enriched in the biological process of bronchiole development and smooth muscle proliferation and immune responses. Four machine learning models used PC37 with sex and age score to generate continuous classification values. Receiver operating characteristic curve analyses of these scores demonstrated consistent areas under the curve of 0.85-0.90 in the test cohort and 0.94-0.96 in the single-sample dataset. Binary classification demonstrated 78.6-80.4% sensitivity and 76-84.4% specificity in the test cohort and 93.5-96.1% sensitivity and 69.5-77.6% specificity in the single-sample classification dataset. Composite analysis of all machine learning models confirmed 78.2% (194 of 248) accuracy in the test cohort and 82.9% (208 of 251) in the single-sample classification dataset. Conclusions: Multiple machine learning models trained with large cohort proteomic datasets consistently distinguished CTD-ILD from IPF. Many of the identified proteins are involved in immune pathways. We further developed a novel approach for single-sample classification, which could facilitate honing the differential diagnosis of ILD in challenging cases and improve clinical decision making.

Unmasking Deception: A Proactive Approach to Detecting and Preventing Fake Job Offers

The research presents an innovative approach to combating fraudulent job postings on the internet through the utilization of machine learning-based classification techniques. By leveraging different classifiers, including single classifiers and ensemble classifiers, the study aims to discern fake job postings amidst a vast array of online listings. Through an extensive analysis of experimental results, the research identifies ensemble classifiers as the optimal choice for detecting employment scams, surpassing the efficacy of single classifiers. Employing a dataset sourced from Kaggle, the study focuses on distinguishing between real and fake job postings, with the latter comprising a minority of the dataset, as anticipated. By adhering to these structured stages, the research aims to contribute to the advancement of methods for identifying and mitigating fraudulent activities in online job postings, thereby enhancing the integrity of online recruitment processes.

MAGIC-DR: An interpretable machine-learning guided approach for acute myeloid leukemia measurable residual disease analysis.

Multiparameter flow cytometry is widely used for acute myeloid leukemia minimal residual disease testing (AML MRD) but is time consuming and demands substantial expertise. Machine learning offers potential advancements in accuracy and efficiency, but has yet to be widely adopted for this application. To explore this, we trained single cell XGBoost classifiers from 98 diagnostic AML cell populations and 30 MRD negative samples. Performance was assessed by cross-validation. Predictions were integrated with UMAP as a heatmap parameter for an augmented/interactive AML MRD analysis framework, which was benchmarked against traditional MRD analysis for 25 test cases. The results showed that XGBoost achieved a median AUC of 0.97, effectively distinguishing diverse AML cell populations from normal cells. When integrated with UMAP, the classifiers highlighted MRD populations against the background of normal events. Our pipeline, MAGIC-DR, incorporated classifier predictions and UMAP into flow cytometry standard (FCS) files. This enabled a human-in-the-loop machine learning guided MRD workflow. Validation against conventional analysis for 25 MRD samples showed 100% concordance in myeloid blast detection, with MAGIC-DR also identifying several immature monocytic populations not readily found by conventional analysis. In conclusion, Integrating a supervised classifier with unsupervised dimension reduction offers a robust method for AML MRD analysis that can be seamlessly integrated into conventional workflows. Our approach can support and augment human analysis by highlighting abnormal populations that can be gated on for quantification and further assessment. This has the potential to speed up MRD analysis, and potentially improve detection sensitivity for certain AML immunophenotypes.

Reconstructing historical urban landscapes from KH-9 HEXAGON mapping camera system imagery: an example of Hangzhou City

ABSTRACT Declassified images from the Keyhole (KH)-9 HEXAGON mapping camera system (MCS) offer fine-scale details of urban regions. However, these images have seldom been utilized in urban research due to challenges in labelling (collecting training samples), having only a single panchromatic band and classification. To tackle these limitations, this paper focuses on developing a multi-stage reconstructed historical fine-scale urban landscape (RHFUL) pipeline for KH-9 HEXAGON MCS. The proposed pipeline first integrates internalized parameters, hierarchical object-based image analysis properties and class variability to synthesize new features, abbreviated to IHC. Second, the pipeline uses a weak semi-automated supervised labelling (WSSL) approach to acquire training samples. Finally, the training samples and generated features are subjected to the SegNet deep learning architecture. The performance of each step was assessed against corresponding state-of-the-art benchmark approaches for each of synthesizing features, labelling and classification. In the proposed RHFUL pipeline, the proposed IHC provided the most salient information for urban classification, WSSL labelled urban features more accurately, and the SegNet architecture classified more accurately the urban features relative to the benchmarks. Considering the potential advantages, but also limitations of KH-9 HEXAGON MCS images, further research should be undertaken, particularly drawing on the current advances in pattern recognition techniques for contemporary digital satellite sensors.

Efficiency of Identification of Blackcurrant Powders Using Classifier Ensembles.

In the modern times of technological development, it is important to select adequate methods to support various food and industrial problems, including innovative techniques with the help of artificial intelligence (AI). Effective analysis and the speed of algorithm implementation are key points in assessing the quality of food products. Non-invasive solutions are being sought to achieve high accuracy in the classification and evaluation of various food products. This paper presents various machine learning algorithm architectures to evaluate the efficiency of identifying blackcurrant powders (i.e., blackcurrant concentrate with a density of 67 °Brix and a color coefficient of 2.352 (E520/E420) in combination with the selected carrier) based on information encoded in microscopic images acquired via scanning electron microscopy (SEM). Recognition of blackcurrant powders was performed using texture feature extraction from images aided by the gray-level co-occurrence matrix (GLCM). It was evaluated for quality using individual single classifiers and a metaclassifier based on metrics such as accuracy, precision, recall, and F1-score. The research showed that the metaclassifier, as well as a single random forest (RF) classifier most effectively identified blackcurrant powders based on image texture features. This indicates that ensembles of classifiers in machine learning is an alternative approach to demonstrate better performance than the existing traditional solutions with single neural models. In the future, such solutions could be an important tool to support the assessment of the quality of food products in real time. Moreover, ensembles of classifiers can be used for faster analysis to determine the selection of an adequate machine learning algorithm for a given problem.

Fake Job Post Detection using Machine Learning

The proliferation of online job platforms has given rise to a concerning increase in fraudulent job postings, presenting significant risks to job seekers and undermining the credibility of the job market. This research paper aims to address the pressing issue of fake job post identification by leveraging machine learning techniques. The primary objective is to develop a robust automated tool capable of accurately distinguishing between authentic and deceptive job advertisements. The proposed methodology utilizes a range of machine learning algorithms, incorporating supervised learning techniques and natural language processing methods, to analyze and classify job postings. Through the integration of both single classifiers and ensemble classifiers, the system evaluates and compares results, effectively detecting fraudulent job postings on the web. The study underscores the need for a proactive approach, acknowledging the dynamic tactics employed by scammers. Continuous refinement and adaptation of the machine learning models are emphasized to stay ahead of evolving fraudulent strategies. Ultimately, this research contributes to establishing a more secure online job market, fostering trust among job seekers and mitigating the financial and emotional risks associated with deceptive job postings.

Semi-supervised medical image classification via distance correlation minimization and graph attention regularization

We propose a novel semi-supervised learning method to leverage unlabeled data alongside minimal annotated data and improve medical imaging classification performance in realistic scenarios with limited labeling budgets to afford data annotations. Our method introduces distance correlation to minimize correlations between feature representations from different views of the same image encoded with non-coupled deep neural networks architectures. In addition, it incorporates a data-driven graph-attention based regularization strategy to model affinities among images within the unlabeled data by exploiting their inherent relational information in the feature space. We conduct extensive experiments on four medical imaging benchmark data sets involving X-ray, dermoscopic, magnetic resonance, and computer tomography imaging on single and multi-label medical imaging classification scenarios. Our experiments demonstrate the effectiveness of our method in achieving very competitive performance and outperforming several state-of-the-art semi-supervised learning methods. Furthermore, they confirm the suitability of distance correlation as a versatile dependence measure and the benefits of the proposed graph-attention based regularization for semi-supervised learning in medical imaging analysis.

Prediction of microvascular invasion in hepatocellular carcinoma based on multi-phase dynamic enhanced CT radiomics feature and multi-classifier hierarchical fusion model

To predict microvascular invasion (MVI) in hepatocellular carcinoma (HCC) using a model based on multiphase dynamic-enhanced CT (DCE-CT) radiomics feature and hierarchical fusion of multiple classifiers. We retrospectively collected preoperative DCE-CT images from 111 patients with pathologically confirmed HCC in Guangzhou First People's Hospital between January, 2016 and April, 2020. The volume of interest was outlined in the early arterial phase, late arterial phase, portal venous phase and equilibrium phase, and radiomics features of these 4 phases were extracted. Seven classifiers based on different algorithms were trained using the filtered feature subsets to obtain multiple base classifiers under each phase. According to the hierarchical fusion strategy, a multi-criteria decision-making-based weight assignment algorithm was used for fusing each base classifier under the same phase with the model after extracting the phase information to obtain the prediction model. The proposed model was evaluated using a 5-fold cross-validation and assessed for area under the ROC curve (AUC), accuracy, sensitivity, and specificity. The prediction model was also compared with the fusion models using a single phase or multiple phases, models based on a single phase with a single classifier, models with different base classifier diversities, and 8 classifier models based on other ensemble methods. The experimental results showed that the performance of the proposed model for predicting HCCMVI was optimal after incorporating the 4 phases and 7 classifiers, with AUC, accuracy, sensitivity, and specificity of 0.828, 0.766, 0.877, and 0.648, respectively. Comparative experiments showed that this prediction model outperformed the models based on a single phase with a single classifier and other ensemble models. The proposed prediction model is effective for predicting MVI in HCC with superior performance to other models.

Simcryocluster: a semantic similarity clustering method of cryo-EM images by adopting contrastive learning

BackgroundCryo-electron microscopy (Cryo-EM) plays an increasingly important role in the determination of the three-dimensional (3D) structure of macromolecules. In order to achieve 3D reconstruction results close to atomic resolution, 2D single-particle image classification is not only conducive to single-particle selection, but also a key step that affects 3D reconstruction. The main task is to cluster and align 2D single-grain images into non-heterogeneous groups to obtain sharper single-grain images by averaging calculations. The main difficulties are that the cryo-EM single-particle image has a low signal-to-noise ratio (SNR), cannot manually label the data, and the projection direction is random and the distribution is unknown. Therefore, in the low SNR scenario, how to obtain the characteristic information of the effective particles, improve the clustering accuracy, and thus improve the reconstruction accuracy, is a key problem in the 2D image analysis of single particles of cryo-EM.ResultsAiming at the above problems, we propose a learnable deep clustering method and a fast alignment weighted averaging method based on frequency domain space to effectively improve the class averaging results and improve the reconstruction accuracy. In particular, it is very prominent in the feature extraction and dimensionality reduction module. Compared with the classification method based on Bayesian and great likelihood, a large amount of single particle data is required to estimate the relative angle orientation of macromolecular single particles in the 3D structure, and we propose that the clustering method shows good results.ConclusionsSimcryoCluster can use the contrastive learning method to perform well in the unlabeled high-noise cryo-EM single particle image classification task, making it an important tool for cryo-EM protein structure determination

A basket trial design based on constrained hierarchical Bayesian model for latent subgroups

ABSTRACT It is well known a basket trial consisting of multiple cancer types has the potential of borrowing strength across the baskets defined by the cancer types, leading to an efficient design in terms of sample size and trial duration. The treatment effects in those baskets are often heterogeneous and categorized by the cancer types being sensitive or insensitive to the treatment. Hence, the assumption of exchangeability in many existing basket trials may be violated, and there is a need to design trials without this assumption. In this paper, we simplify the constrained hierarchical Bayesian model for latent subgroups (CHBM-LS) for two classifiers to deal with the potential heterogeneity of treatment effects due to the single classifier of the cancer type. Different baskets are aggregated into subgroups using a latent subgroup modeling approach. The treatment effects are similar and exchangeable to facilitate information borrowing within each latent subgroup. Applying the simplified CHBM-LS approach to the real basket trials where baskets defined by only cancer types shows better performance than other available approaches. Further simulation study also demonstrates this CHBM-LS approach outperforms other approaches with higher statistical power and better-controlled type I error rates under various scenarios.