ROC operating point selection for classification of imbalanced data with application to computer-aided polyp detection in CT colonography

Clinical target volume delineation quality assurance for MRI-guided prostate radiotherapy using deep learning with uncertainty estimation

BACKGROUND AND PURPOSE

Previous studies on automatic delineation quality assurance (QA) have mostly focused on CT-based planning. As MRI-guided radiotherapy is increasingly utilized in prostate cancer treatment, there is a need for more research on MRI-specific automatic QA. This work proposes a clinical target volume (CTV) delineation QA framework based on deep learning (DL) for MRI-guided prostate radiotherapy.

MATERIALS AND METHODS

The proposed workflow utilized a 3D dropblock ResUnet++ (DB-ResUnet++) to generate multiple segmentation predictions via Monte Carlo dropout which were used to compute an average delineation and area of uncertainty. A logistic regression (LR) classifier was employed to classify the manual delineation as pass or discrepancy based on the spatial association between the manual delineation and the network's outputs. This approach was evaluated on a multicentre MRI-only prostate radiotherapy dataset and compared with our previously published QA framework based on AN-AG Unet.

RESULTS

The proposed framework achieved an area under the receiver operating curve (AUROC) of 0.92, a true positive rate (TPR) of 0.92 and a false positive rate of 0.09 with an average processing time per delineation of 1.3 min. Compared with our previous work using AN-AG Unet, this method generated fewer false positive detections at the same TPR with a much faster processing speed.

CONCLUSION

To the best of our knowledge, this is the first study to propose an automatic delineation QA tool using DL with uncertainty estimation for MRI-guided prostate radiotherapy, which can potentially be used for reviewing prostate CTV delineation in multicentre clinical trials.

Generalizability of 3D CNN models for age estimation in diverse youth populations using structural MRI

Recently, several studies have investigated the neurodevelopment of psychiatric disorders using brain data acquired via structural magnetic resonance imaging (sMRI). These analyses have shown the potential of sMRI data to provide a relatively precise characterization of brain structural biomarkers. Despite these advances, a relatively unexplored question is how reliable and consistent a model is when assessing subjects from other independent datasets. In this study, we investigate the performance and generalizability of the same model architecture trained from distinct datasets comprising youths in diverse stages of neurodevelopment and with different mental health conditions. We employed models with the same 3D convolutional neural network (CNN) architecture to assess autism spectrum disorder (ASD), attention deficit hyperactivity disorder (ADHD), brain age, and a measure of dimensional psychopathology, the Child Behavior Checklist (CBCL) total score. The investigated datasets include the Autism Brain Imaging Data Exchange II (ABIDE-II, N = 580), Attention Deficit Hyperactivity Disorder (ADHD-200, N = 922), Brazilian High-Risk Cohort Study (BHRCS, N = 737), and Adolescent Brain Cognitive Development (ABCD, N = 11,031). Models' performance and interpretability were assessed within each dataset (for diagnosis tasks) and inter-datasets (for age estimation). Despite the demographic and phenotypic differences of the subjects, all models presented significant estimations for age (p value < 0.001) within and between datasets. In addition, most models showed a moderate to high correlation in age estimation. The results, including the models' brain regions of interest (ROI), were analyzed and discussed in light of the youth neurodevelopmental structural changes. Among other interesting discoveries, we found that less confounded training datasets produce models with higher generalization capacity.

Automatic classification of distal radius fracture using a two-stage ensemble deep learning framework

Distal radius fractures (DRFs) are one of the most common types of wrist fracture and can be subdivided into intra- and extra-articular fractures. Compared with extra-articular DRFs which spare the joint surface, intra-articular DRFs extend to the articular surface and can be more difficult to treat. Identification of articular involvement can provide valuable information about the characteristics of fracture patterns. In this study, a two-stage ensemble deep learning framework was proposed to differentiate intra- and extra-articular DRFs automatically on posteroanterior (PA) view wrist X-rays. The framework firstly detects the distal radius region of interest (ROI) using an ensemble model of YOLOv5 networks, which imitates the clinicians' search pattern of zooming in on relevant regions to assess abnormalities. Secondly, an ensemble model of EfficientNet-B3 networks classifies the fractures in the detected ROIs into intra- and extra-articular. The framework achieved an area under the receiver operating characteristic curve of 0.82, an accuracy of 0.81, a true positive rate of 0.83 and a false positive rate of 0.27 (specificity of 0.73) for differentiating intra- from extra-articular DRFs. This study has demonstrated the potential in automatic DRF characterization using deep learning on clinically acquired wrist radiographs and can serve as a baseline for further research in incorporating multi-view information for fracture classification.

Modeling risk of hypoglycemia during and following physical activity in people with type 1 diabetes using explainable mixed-effects machine learning

BACKGROUND

Physical activity (PA) can cause increased hypoglycemia (glucose <70 mg/dL) risk in people with type 1 diabetes (T1D). We modeled the probability of hypoglycemia during and up to 24 h following PA and identified key factors associated with hypoglycemia risk.

METHODS

We leveraged a free-living dataset from Tidepool comprised of glucose measurements, insulin doses, and PA data from 50 individuals with T1D (6448 sessions) for training and validating machine learning models. We also used data from the T1Dexi pilot study that contains glucose management and PA data from 20 individuals with T1D (139 session) for assessing the accuracy of the best performing model on an independent test dataset. We used mixed-effects logistic regression (MELR) and mixed-effects random forest (MERF) to model hypoglycemia risk around PA. We identified risk factors associated with hypoglycemia using odds ratio and partial dependence analysis for the MELR and MERF models, respectively. Prediction accuracy was measured using the area under the receiver operating characteristic curve (AUROC).

RESULTS

The analysis identified risk factors significantly associated with hypoglycemia during and following PA in both MELR and MERF models including glucose and body exposure to insulin at the start of PA, low blood glucose index 24 h prior to PA, and PA intensity and timing. Both models showed overall hypoglycemia risk peaking 1 h after PA and again 5-10 h after PA, which is consistent with the hypoglycemia risk pattern observed in the training dataset. Time following PA impacted hypoglycemia risk differently across different PA types. Accuracy of hypoglycemia prediction using the fixed effects of the MERF model was highest when predicting hypoglycemia during the first hour following the start of PA (AUROCVALIDATION = 0.83 and AUROCTESTING = 0.86) and decreased when predicting hypoglycemia in the 24 h after PA (AUROCVALIDATION = 0.66 and AUROCTESTING = 0.68).

CONCLUSION

Hypoglycemia risk after the start of PA can be modeled using mixed-effects machine learning to identify key risk factors that may be used within decision support and insulin delivery systems. We published the population-level MERF model online for others to use.

Slope Entropy Characterisation: The Role of the δ Parameter

Many time series entropy calculation methods have been proposed in the last few years. They are mainly used as numerical features for signal classification in any scientific field where data series are involved. We recently proposed a new method, Slope Entropy (SlpEn), based on the relative frequency of differences between consecutive samples of a time series, thresholded using two input parameters, γ and δ. In principle, δ was proposed to account for differences in the vicinity of the 0 region (namely, ties) and, therefore, was usually set at small values such as 0.001. However, there is no study that really quantifies the role of this parameter using this default or other configurations, despite the good SlpEn results so far. The present paper addresses this issue, removing δ from the SlpEn calculation to assess its real influence on classification performance, or optimising its value by means of a grid search in order to find out if other values beyond the 0.001 value provide significant time series classification accuracy gains. Although the inclusion of this parameter does improve classification accuracy according to experimental results, gains of 5% at most probably do not support the additional effort required. Therefore, SlpEn simplification could be seen as a real alternative.

Colon and Rectal Cancer Management in Low-Resource Settings

Colorectal cancer (CRC) incidence is rising in low- and middle-income countries, which also face disproportionate mortality from CRC, mainly due to diagnosis at late stages. Various challenges to CRC care exist at multiple societal levels in underserved populations. In this article, barriers to CRC care, strategies for screening, and treatment in resource-limited settings, and future directions are discussed within a global context.

Automatic Identification of Snoring and Groaning Segments in Acoustic Recordings

Sleep-related breathing disorders have severe impact on the quality of lives of those suffering from them. These disorders present with a variety of symptoms, out of which snoring and groaning are very common. This paper presents an algorithm to identify and classify segments of acoustic respiratory sound recordings that contain both groaning and snoring events. The recordings were obtained from a database containing 20 subjects from which features based on the Mel-frequency cepstral coefficients (MFCC) were extracted. In the first stage of the algorithm, segments of recordings consisting of either snoring or groaning episodes - without classifying them - were identified. In the second stage, these segments were further differentiated into individual groaning or snoring events. The algorithm in the first stage achieved a sensitivity and specificity of 90.5% ±2.9% and 90.0% ±1.6% respectively, using a RUSBoost model. In the second stage, a random forest classifier was used, and the accuracies for groan and snore events were 78.1% ±4.7% and 78.4% ±4.7% respectively.

Analysis of the benefits of imputation models over traditional QSAR models for toxicity prediction

Recently, imputation techniques have been adapted to predict activity values among sparse bioactivity matrices, showing improvements in predictive performance over traditional QSAR models. These models are able to use experimental activity values for auxiliary assays when predicting the activity of a test compound on a specific assay. In this study, we tested three different multi-task imputation techniques on three classification-based toxicity datasets: two of small scale (12 assays each) and one large scale with 417 assays. Moreover, we analyzed in detail the improvements shown by the imputation models. We found that test compounds that were dissimilar to training compounds, as well as test compounds with a large number of experimental values for other assays, showed the largest improvements. We also investigated the impact of sparsity on the improvements seen as well as the relatedness of the assays being considered. Our results show that even a small amount of additional information can provide imputation methods with a strong boost in predictive performance over traditional single task and multi-task predictive models.

Discriminating Bacterial Infection from Other Causes of Fever Using Body Temperature Entropy Analysis

Body temperature is usually employed in clinical practice by strict binary thresholding, aiming to classify patients as having fever or not. In the last years, other approaches based on the continuous analysis of body temperature time series have emerged. These are not only based on absolute thresholds but also on patterns and temporal dynamics of these time series, thus providing promising tools for early diagnosis. The present study applies three time series entropy calculation methods (Slope Entropy, Approximate Entropy, and Sample Entropy) to body temperature records of patients with bacterial infections and other causes of fever in search of possible differences that could be exploited for automatic classification. In the comparative analysis, Slope Entropy proved to be a stable and robust method that could bring higher sensitivity to the realm of entropy tools applied in this context of clinical thermometry. This method was able to find statistically significant differences between the two classes analyzed in all experiments, with sensitivity and specificity above 70% in most cases.

A neural network for glomerulus classification based on histological images of kidney biopsy

Background

Computer-aided diagnosis (CAD) systems based on medical images could support physicians in the decision-making process. During the last decades, researchers have proposed CAD systems in several medical domains achieving promising results.

CAD systems play an important role in digital pathology supporting pathologists in analyzing biopsy slides by means of standardized and objective workflows. In the proposed work, we designed and tested a novel CAD system module based on image processing techniques and machine learning, whose objective was to classify the condition affecting renal corpuscles (glomeruli) between sclerotic and non-sclerotic. Such discrimination is useful for the biopsy slides evaluation performed by pathologists.

Results

We collected 26 digital slides taken from the kidneys of 19 donors with Periodic Acid-Schiff staining. Expert pathologists have conducted the slides preparation, digital acquisition and glomeruli annotations. Before setting the classifiers, we evaluated several feature extraction techniques from the annotated regions. Then, a feature reduction procedure followed by a shallow artificial neural network allowed discriminating between the glomeruli classes.

We evaluated the workflow considering an independent dataset (i.e., processing images not used in the training procedure). Ten independent runs of the training algorithm, and evaluation, allowed achieving MCC and Accuracy of 0.95 (± 0.01) and 0.99 (standard deviation < 0.00), respectively. We also obtained good precision (0.9844 ± 0.0111) and recall (0.9310 ± 0.0153).

Conclusions

Results on the test set confirm that the proposed workflow is consistent and reliable for the investigated domain, and it can support the clinical practice of discriminating the two classes of glomeruli. Analyses on misclassifications show that the involved images are usually affected by staining artefacts or present partial sections due to slice preparation and staining processes. In clinical practice, however, pathologists discard images showing such artefacts.

Developing Non-Laboratory Cardiovascular Risk Assessment Charts and Validating Laboratory and Non-Laboratory-Based Models

Background:

Developing simplified risk assessment model based on non-laboratory risk factors that could determine cardiovascular risk as accurately as laboratory-based one can be valuable, particularly in developing countries where there are limited resources.

Objective:

To develop a simplified non-laboratory cardiovascular disease risk assessment chart based on previously reported laboratory-based chart and evaluate internal and external validation, and recalibration of both risk models to assess the performance of risk scoring tools in other population.

Methods:

A 10-year non-laboratory-based risk prediction chart was developed for fatal and non-fatal CVD using Cox Proportional Hazard regression. Data from the Isfahan Cohort Study (ICS), a population-based study among 6504 adults aged ≥ 35 years, followed-up for at least ten years was used for the non-laboratory-based model derivation. Participants were followed up until the occurrence of CVD events. Tehran Lipid and Glucose Study (TLGS) data was used to evaluate the external validity of both non-laboratory and laboratory risk assessment models in other populations rather than one used in the model derivation.

Results:

The discrimination and calibration analysis of the non-laboratory model showed the following values of Harrell’s C: 0.73 (95% CI 0.71–0.74), and Nam-D’Agostino χ²:11.01 (p = 0.27), respectively. The non-laboratory model was in agreement and classified high risk and low risk patients as accurately as the laboratory one. Both non-laboratory and laboratory risk prediction models showed good discrimination in the external validation, with Harrell’s C of 0.77 (95% CI 0.75–0.78) and 0.78 (95% CI 0.76–0.79), respectively.

Conclusions:

Our simplified risk assessment model based on non-laboratory risk factors could determine cardiovascular risk as accurately as laboratory-based one. This approach can provide simple risk assessment tool where laboratory testing is unavailable, inconvenient, and costly.

CTA-Based Non-invasive Estimation of Pressure Gradients Across a CoA: a Validation Against Cardiac Catheterisation

Non-invasive estimation of pressure gradients across a coarctation of the aorta (CoA) can reduce the need for diagnostic cardiac catheterisation. We aimed to validate two novel computational strategies-target-value approaching (TVA) and target-value fixing (TVF)-together with unrefined Doppler estimates, and to compare their diagnostic performance in identifying critical pressure drops for 40 patients. Compared to catheterisation, no statistically significant difference was demonstrated with TVA (P = 0.086), in contrast to TVF (P = 0.005) and unrefined Doppler echocardiography (P < 0.001). TVA manifested the strongest correlation with catheterisation (r = 0.93), compared to TVF (r = 0.83) and echocardiography (r = 0.67) (all P < 0.001). In discriminating pressure gradients greater than 20 mmHg, TVA, TVF, and echocardiography had respective sensitivities of 0.92, 0.88, and 0.80; specificities of 0.93, 0.80, and 0.73; and AUCs of 0.96, 0.89, and 0.80. The TVA strategy may serve as an effective and easily implemented approach to be used in clinical management of patients with CoA. Graphical Abstract Central illustration. Pressure gradients estimated using Doppler echocardiography and two novel computational strategies (TVA and TVF) were compared with cardiac catheterisation for 40 patients. TVA and TVF utilised the CTA images to obtain the CoA anatomy and Doppler echocardiography velocimetry to obtain velocity data for the assignment of CFD boundary conditions.

Hyperspectral image processing for the identification and quantification of lentiviral particles in fluid samples

Optical spectroscopic techniques have been commonly used to detect the presence of biofilm-forming pathogens (bacteria and fungi) in the agro-food industry. Recently, near-infrared (NIR) spectroscopy revealed that it is also possible to detect the presence of viruses in animal and vegetal tissues. Here we report a platform based on visible and NIR (VNIR) hyperspectral imaging for non-contact, reagent free detection and quantification of laboratory-engineered viral particles in fluid samples (liquid droplets and dry residue) using both partial least square-discriminant analysis and artificial feed-forward neural networks. The detection was successfully achieved in preparations of phosphate buffered solution and artificial saliva, with an equivalent pixel volume of 4 nL and lowest concentration of 800 TU·\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\upmu$$\end{document}μL⁻¹. This method constitutes an innovative approach that could be potentially used at point of care for rapid mass screening of viral infectious diseases and monitoring of the SARS-CoV-2 pandemic.

Duplicate Detection of Spike Events: A Relevant Problem in Human Single-Unit Recordings

Single-unit recordings in the brain of behaving human subjects provide a unique opportunity to advance our understanding of neural mechanisms of cognition. These recordings are exclusively performed in medical centers during diagnostic or therapeutic procedures. The presence of medical instruments along with other aspects of the hospital environment limit the control of electrical noise compared to animal laboratory environments. Here, we highlight the problem of an increased occurrence of simultaneous spike events on different recording channels in human single-unit recordings. Most of these simultaneous events were detected in clusters previously labeled as artifacts and showed similar waveforms. These events may result from common external noise sources or from different micro-electrodes recording activity from the same neuron. To address the problem of duplicate recorded events, we introduce an open-source algorithm to identify these artificial spike events based on their synchronicity and waveform similarity. Applying our method to a comprehensive dataset of human single-unit recordings, we demonstrate that our algorithm can substantially increase the data quality of these recordings. Given our findings, we argue that future studies of single-unit activity recorded under noisy conditions should employ algorithms of this kind to improve data quality.

GHOST: Adjusting the Decision Threshold to Handle Imbalanced Data in Machine Learning

Machine learning classifiers trained on class imbalanced data are prone to overpredict the majority class. This leads to a larger misclassification rate for the minority class, which in many real-world applications is the class of interest. For binary data, the classification threshold is set by default to 0.5 which, however, is often not ideal for imbalanced data. Adjusting the decision threshold is a good strategy to deal with the class imbalance problem. In this work, we present two different automated procedures for the selection of the optimal decision threshold for imbalanced classification. A major advantage of our procedures is that they do not require retraining of the machine learning models or resampling of the training data. The first approach is specific for random forest (RF), while the second approach, named GHOST, can be potentially applied to any machine learning classifier. We tested these procedures on 138 public drug discovery data sets containing structure-activity data for a variety of pharmaceutical targets. We show that both thresholding methods improve significantly the performance of RF. We tested the use of GHOST with four different classifiers in combination with two molecular descriptors, and we found that most classifiers benefit from threshold optimization. GHOST also outperformed other strategies, including random undersampling and conformal prediction. Finally, we show that our thresholding procedures can be effectively applied to real-world drug discovery projects, where the imbalance and characteristics of the data vary greatly between the training and test sets.

Pregnancy complications in G20210A mutation carriers associated with high prothrombin activity

OBJECTIVE

To study the association between high activity of Factor II (prothrombin) in blood plasma with G20210A mutation and the development of great obstetrical syndromes.

MATERIAL AND METHODS

A prospective clinical cohort study was conducted on 290 pregnant women (average age 31.7 ± 4.7 years old). The main group was made up of 140 G20210A patients, while the control group comprised 150 women with the wild G20210G type. The aim was to evaluate the activity of Factor II in the venous blood plasma during the stages of pregnancy with regard to trophoblast invasion waves. As per results, association analysis of Factor II activity value and gestational complications was carried out.

RESULTS

In the control group, the median (Me) of Factor II activity ranged from 108% (preconception period) to 144% (pregnancy) [95% CI 130-150]. In patients with the GA type, the value was significantly higher in related periods, ranging from 149 to 181% [95% CI 142-195], p < 0.0001. With Factor II activity ranging from 148.5 to 180.6%, pregnancies in the main group had no complications. Higher levels of Factor II activity were associated with the development of early and/or severe preeclampsia (PE) and fetal growth retardation (FGR).

CONCLUSION

The data obtained regarding Factor II activity in blood plasma, juxtaposed with the development of great obstetrical syndromes, allow to assume that manifestation of G20210A in early and/or severe PE and FGR is associated with this coagulation factor's level of activity. Threshold value of the Factor II activity with G20210A mutation, allowing to predict the development of PE, comprised 171.0% at the preconception stage (AUC - 0.86; p < 0.0001) and within 7-8 weeks of gestation it was 181.3% (AUC - 0.84; p < 0.0001).

Estimating Gender and Age from Brain Structural MRI of Children and Adolescents: A 3D Convolutional Neural Network Multitask Learning Model

Despite recent advances, assessing biological measurements for neuropsychiatric disorders is still a challenge, where confounding variables such as gender and age (as a proxy for neurodevelopment) play an important role. This study explores brain structural magnetic resonance imaging (sMRI) from two public data sets (ABIDE-II and ADHD-200) with healthy control (HC, N = 894), autism spectrum disorder (ASD, N = 251), and attention deficit hyperactivity disorder (ADHD, N = 357) individuals. We used gray and white matter preprocessed via voxel-based morphometry (VBM) to train a 3D convolutional neural network with a multitask learning strategy to estimate gender, age, and mental health status from structural brain differences. Gradient-based methods were employed to generate attention maps, providing clinically relevant identification of most representative brain regions for models' decision-making. This approach resulted in satisfactory predictions for gender and age. ADHD-200-trained models, evaluated in 10-fold cross-validation procedures on test set, obtained a mean absolute error (MAE) of 1.43 years (±0.22 SD) for age prediction and an area under the curve (AUC) of 0.85 (±0.04 SD) for gender classification. In out-of-sample validation, the best-performing ADHD-200 models satisfactorily predicted age (MAE = 1.57 years) and gender (AUC = 0.89) in the ABIDE-II data set. The models' accuracy was in line with the current state-of-the-art machine learning applications in neuroimaging. Key regions for models' accuracy were presented as a meaningful graphical output. New implementations, such as the use of VBM along with a 3D convolutional neural network multitask learning model and a brain imaging graphical output, reinforce the relevance of the proposed workflow.

Comparative Effectiveness of Machine Learning Approaches for Predicting Gastrointestinal Bleeds in Patients Receiving Antithrombotic Treatment

IMPORTANCE

Anticipating the risk of gastrointestinal bleeding (GIB) when initiating antithrombotic treatment (oral antiplatelets or anticoagulants) is limited by existing risk prediction models. Machine learning algorithms may result in superior predictive models to aid in clinical decision-making.

OBJECTIVE

To compare the performance of 3 machine learning approaches with the commonly used HAS-BLED (hypertension, abnormal kidney and liver function, stroke, bleeding, labile international normalized ratio, older age, and drug or alcohol use) risk score in predicting antithrombotic-related GIB.

DESIGN, SETTING, AND PARTICIPANTS

This retrospective cross-sectional study used data from the OptumLabs Data Warehouse, which contains medical and pharmacy claims on privately insured patients and Medicare Advantage enrollees in the US. The study cohort included patients 18 years or older with a history of atrial fibrillation, ischemic heart disease, or venous thromboembolism who were prescribed oral anticoagulant and/or thienopyridine antiplatelet agents between January 1, 2016, and December 31, 2019.

EXPOSURES

A cohort of patients prescribed oral anticoagulant and thienopyridine antiplatelet agents was divided into development and validation cohorts based on date of index prescription. The development cohort was used to train 3 machine learning models to predict GIB at 6 and 12 months: regularized Cox proportional hazards regression (RegCox), random survival forests (RSF), and extreme gradient boosting (XGBoost).

MAIN OUTCOMES AND MEASURES

The performance of the models for predicting GIB in the validation cohort, evaluated using the area under the receiver operating characteristic curve (AUC), sensitivity, specificity, positive predictive value, and prediction density plots. Relative importance scores were used to identify the variables that were most influential in the top-performing machine learning model.

RESULTS

In the entire study cohort of 306 463 patients, 166 177 (54.2%) were male, 193 648 (63.2%) were White, the mean (SD) age was 69.0 (12.6) years, and 12 322 (4.0%) had experienced a GIB. In the validation data set, the HAS-BLED model had an AUC of 0.60 for predicting GIB at 6 months and 0.59 at 12 months. The RegCox model performed the best in the validation set, with an AUC of 0.67 at 6 months and 0.66 at 12 months. XGBoost was similar, with AUCs of 0.67 at 6 months and 0.66 at 12 months, whereas for RSF, AUCs were 0.62 at 6 months and 0.60 at 12 months. The variables with the highest importance scores in the RegCox model were prior GI bleed (importance score, 0.72); atrial fibrillation, ischemic heart disease, and venous thromboembolism combined (importance score, 0.38); and use of gastroprotective agents (importance score, 0.32).

CONCLUSIONS AND RELEVANCE

In this cross-sectional study, the machine learning models examined showed similar performance in identifying patients at high risk for GIB after being prescribed antithrombotic agents. Two models (RegCox and XGBoost) performed modestly better than the HAS-BLED score. A prospective evaluation of the RegCox model compared with HAS-BLED may provide a better understanding of the clinical impact of improved performance.

A Survey on Machine Learning and Deep Learning-based Computer-Aided Methods for Detection of Polyps in CT Colonography

BACKGROUND

Colon cancer generally begins as a neoplastic growth of tissue, called polyps, originating from the inner lining of the colon wall. Most colon polyps are considered harmless but over the time, they can evolve into colon cancer, which, when diagnosed in later stages, is often fatal. Hence, time is of the essence in the early detection of polyps and the prevention of colon cancer.

METHODS

To aid this endeavor, many computer-aided methods have been developed, which use a wide array of techniques to detect, localize and segment polyps from CT Colonography images. In this paper, a comprehensive state-of-the-art method is proposed and categorize this work broadly using the available classification techniques using Machine Learning and Deep Learning.

CONCLUSION

The performance of each of the proposed approach is analyzed with existing methods and also how they can be used to tackle the timely and accurate detection of colon polyps.

A deep learning integrated radiomics model for identification of coronavirus disease 2019 using computed tomography

Since its first outbreak, Coronavirus Disease 2019 (COVID-19) has been rapidly spreading worldwide and caused a global pandemic. Rapid and early detection is essential to contain COVID-19. Here, we first developed a deep learning (DL) integrated radiomics model for end-to-end identification of COVID-19 using CT scans and then validated its clinical feasibility. We retrospectively collected CT images of 386 patients (129 with COVID-19 and 257 with other community-acquired pneumonia) from three medical centers to train and externally validate the developed models. A pre-trained DL algorithm was utilized to automatically segment infected lesions (ROIs) on CT images which were used for feature extraction. Five feature selection methods and four machine learning algorithms were utilized to develop radiomics models. Trained with features selected by L1 regularized logistic regression, classifier multi-layer perceptron (MLP) demonstrated the optimal performance with AUC of 0.922 (95% CI 0.856-0.988) and 0.959 (95% CI 0.910-1.000), the same sensitivity of 0.879, and specificity of 0.900 and 0.887 on internal and external testing datasets, which was equivalent to the senior radiologist in a reader study. Additionally, diagnostic time of DL-MLP was more efficient than radiologists (38 s vs 5.15 min). With an adequate performance for identifying COVID-19, DL-MLP may help in screening of suspected cases.

Fever Time Series Analysis Using Slope Entropy. Application to Early Unobtrusive Differential Diagnosis

Fever is a readily measurable physiological response that has been used in medicine for centuries. However, the information provided has been greatly limited by a plain thresholding approach, overlooking the additional information provided by temporal variations and temperature values below such threshold that are also representative of the subject status. In this paper, we propose to utilize continuous body temperature time series of patients that developed a fever, in order to apply a method capable of diagnosing the specific underlying fever cause only by means of a pattern relative frequency analysis. This analysis was based on a recently proposed measure, Slope Entropy, applied to a variety of records coming from dengue and malaria patients, among other fever diseases. After an input parameter customization, a classification analysis of malaria and dengue records took place, quantified by the Matthews Correlation Coefficient. This classification yielded a high accuracy, with more than 90% of the records correctly labelled in some cases, demonstrating the feasibility of the approach proposed. This approach, after further studies, or combined with more measures such as Sample Entropy, is certainly very promising in becoming an early diagnosis tool based solely on body temperature temporal patterns, which is of great interest in the current Covid-19 pandemic scenario.

Radiomics Features Predict CIC Mutation Status in Lower Grade Glioma

MRI in combination with genomic markers are critical in the management of gliomas. Radiomics and radiogenomics analysis facilitate the quantitative assessment of tumor properties which can be used to model both molecular subtype and predict disease progression. In this work, we report on the Drosophila gene capicua (CIC) mutation biomarker effects alongside radiomics features on the predictive ability of CIC mutation status in lower-grade gliomas (LGG). Genomic data of lower grade glioma (LGG) patients from The Cancer Genome Atlas (TCGA) (n = 509) and corresponding MR images from TCIA (n = 120) were utilized. Following tumor segmentation, radiomics features were extracted from T1, T2, T2 Flair, and T1 contrast enhanced (CE) images. Lasso feature reduction was used to obtain the most important MR image features and then logistic regression used to predict CIC mutation status. In our study, CIC mutation rarely occurred in Astrocytoma but has a high probability of occurrence in Oligodendroglioma. The presence of CIC mutation was found to be associated with better survival of glioma patients (p < 1e−4, HR: 0.2445), even with co-occurrence of IDH mutation and 1p/19q co-deletion (p = 0.0362, HR: 0.3674). An eleven-feature model achieved glioma prediction accuracy of 94.2% (95% CI, 94.03–94.38%), a six-feature model achieved oligodendroglioma prediction accuracy of 92.3% (95% CI, 91.70–92.92%). MR imaging and its derived image of gliomas with CIC mutation appears more complex and non-uniform but are associated with lower malignancy. Our study identified CIC as a potential prognostic factor in glioma which has close associations with survival. MRI radiomic features could predict CIC mutation, and reflect less malignant manifestations such as milder necrosis and larger tumor volume in MRI and its derived images that could help clinical judgment.

A Machine Learning Approach for Long-Term Prognosis of Bladder Cancer based on Clinical and Molecular Features

Improving the consistency and reproducibility of bladder cancer prognoses necessitates the development of accurate, predictive prognostic models. Current methods of determining the prognosis of bladder cancer patients rely on manual decision-making, including factors with high intra- and inter-observer variability, such as tumor grade. To advance the long-term prediction of bladder cancer prognoses, we developed and tested a computational model to predict the 10-year overall survival outcome using population-based bladder cancer data, without considering tumor grade classification. The resulted predictive model demonstrated promising performance using a combination of clinical and molecular features, and was also strongly related to patient overall survival in Cox models. Our study suggests that machine learning methods can provide reliable long-term prognoses for bladder cancer patients, without relying on the less consistent tumor grade. If validated in clinical trials, this automated approach could guide and improve personalized management and treatment for bladder cancer patients.

Designing of precise vaccine construct against visceral leishmaniasis through predicted epitope ensemble: A contemporary approach

Visceral leishmaniasis (VL) caused by Leishmania donovani is a fatal parasitic disease affecting primarily the poor population in endemic countries. Increasing number of deaths as well as resistant to existing drugs necessitates the development of an effective vaccine for successful treatment of VL. The present study employed a combinatorial approach for designing monomer vaccine construct against L. donovani by applying forecasted B- and T- cell epitopes from 4 genome derived antigenic proteins having secretory signal peptides and glycophosphatidylinositol (GPI) anchors with ≤ 1 transmembrane helix. The forecasted population coverage of chosen T cell epitope ensemble (combined HLA class I and II) cover 99.14 % of world-wide human population. The predicted 3D structure of vaccine constructs (VC1/VC2) were modeled using homology modeling approach and docked to innate immune receptors TLR-2 and TLR-4 with respective docking energies -1231.4/-910.3 and -1119.4/-1476 kcal/mol. Overall, the aforementioned designed vaccine constructs were found appropriate for including in self-assembly protein nanoparticles (SAPN) for further study in developing cutting-edge precision vaccine against VL in short duration with cost-effective manner.

Optical 4D signal detection in turbid water by multi-dimensional integral imaging using spatially distributed and temporally encoded multiple light sources

We propose an underwater optical signal detection system based on multi-dimensional integral imaging with spatially distributed multiple light sources and four-dimensional (4D) spatial-temporal correlation. We demonstrate our system for the detection of optical signals in turbid water. A 4D optical signal is generated from a three-dimensional (3D) spatial distribution of underwater light sources, which are temporally encoded using spread spectrum techniques. The optical signals are captured by an array of cameras, and 3D integral imaging reconstruction is performed, followed by multi-dimensional correlation to detect the optical signal. Inclusion of multiple light sources located at different depths allows for successful signal detection at turbidity levels not feasible using only a single light source. We consider the proposed system under varied turbidity levels using both Pseudorandom and Gold Codes for temporal signal coding. We also compare the effectiveness of the proposed underwater optical signal detection system to a similar system using only a single light source and compare between conventional and integral imaging-based signal detection. The underwater signal detection capabilities are measured through performance-based metrics such as receiver operating characteristic (ROC) curves, the area under the curve (AUC), and the number of detection errors. Furthermore, statistical analysis, including Kullback-Leibler divergence and Bhattacharya distance, shows improved performance of the proposed multi-source integral imaging underwater system. The proposed integral-imaging based approach is shown to significantly outperform conventional imaging-based methods.

Multi-scale characterizations of colon polyps via computed tomographic colonography

Texture features have played an essential role in the field of medical imaging for computer-aided diagnosis. The gray-level co-occurrence matrix (GLCM)-based texture descriptor has emerged to become one of the most successful feature sets for these applications. This study aims to increase the potential of these features by introducing multi-scale analysis into the construction of GLCM texture descriptor. In this study, we first introduce a new parameter - stride, to explore the definition of GLCM. Then we propose three multi-scaling GLCM models according to its three parameters, (1) learning model by multiple displacements, (2) learning model by multiple strides (LMS), and (3) learning model by multiple angles. These models increase the texture information by introducing more texture patterns and mitigate direction sparsity and dense sampling problems presented in the traditional Haralick model. To further analyze the three parameters, we test the three models by performing classification on a dataset of 63 large polyp masses obtained from computed tomography colonoscopy consisting of 32 adenocarcinomas and 31 benign adenomas. Finally, the proposed methods are compared to several typical GLCM-texture descriptors and one deep learning model. LMS obtains the highest performance and enhances the prediction power to 0.9450 with standard deviation 0.0285 by area under the curve of receiver operating characteristics score which is a significant improvement.

Prediction of Atypical Ductal Hyperplasia Upgrades Through a Machine Learning Approach to Reduce Unnecessary Surgical Excisions

PURPOSE

Surgical excision is currently recommended for all occurrences of atypical ductal hyperplasia (ADH) found on core needle biopsies for malignancy diagnoses and treatment of lesions. The excision of all ADH lesions may lead to overtreatment, which results in invasive surgeries for benign lesions in many women. A machine learning method to predict ADH upgrade may help clinicians and patients decide whether combined active surveillance and hormonal therapy is a reasonable alternative to surgical excision.

METHODS

The following six machine learning models were developed to predict ADH upgrade from core needle biopsy: gradient-boosting trees, random forest, radial support vector machine (SVM), weighted K-nearest neighbors (KNN), logistic elastic net, and logistic regression. The study cohort consisted of 128 lesions from 124 women at a tertiary academic care center in New Hampshire who had ADH on core needle biopsy and who underwent an associated surgical excision from 2011 to 2017.

RESULTS

The best-performing models were gradient-boosting trees (area under the curve [AUC], 68%; accuracy, 78%) and random forest (AUC, 67%; accuracy, 77%). The top five most important features that determined ADH upgrade were age at biopsy, lesion size, number of biopsies, needle gauge, and personal and family history of breast cancer. Using the random forest model, 98% of all malignancies would have been diagnosed through surgical biopsies, whereas 16% of unnecessary surgeries on benign lesions could have been avoided (ie, 87% sensitivity at 45% specificity).

CONCLUSION

These results add to the growing body of support for machine learning models as useful aids for clinicians and patients in decisions about the clinical management of ADH.

NMR-Based Metabolomic Approach Tracks Potential Serum Biomarkers of Disease Progression in Patients with Type 2 Diabetes Mellitus

Type 2 diabetes mellitus (T2DM) is a metabolic disorder characterized by chronic hyperglycemia associated with alterations in carbohydrate, lipid, and protein metabolism. The prognosis of T2DM patients is highly dependent on the development of complications, and therefore the identification of biomarkers of T2DM progression, with minimally invasive techniques, is a huge need. In the present study, we applied a ¹H-Nuclear Magnetic Resonance (¹H-NMR)-based metabolomic approach coupled with multivariate data analysis to identify serum metabolite profiles associated with T2DM development and progression. To perform this, we compared the serum metabolome of non-diabetic subjects, treatment-naïve non-complicated T2DM patients, and T2DM patients with complications in insulin monotherapy. Our analysis revealed a significant reduction of alanine, glutamine, glutamate, leucine, lysine, methionine, tyrosine, and phenylalanine in T2DM patients with respect to non-diabetic subjects. Moreover, isoleucine, leucine, lysine, tyrosine, and valine levels distinguished complicated patients from patients without complications. Overall, the metabolic pathway analysis suggested that branched-chain amino acid (BCAA) metabolism is significantly compromised in T2DM patients with complications, while perturbation in the metabolism of gluconeogenic amino acids other than BCAAs characterizes both early and advanced T2DM stages. In conclusion, we identified a metabolic serum signature associated with T2DM stages. These data could be integrated with clinical characteristics to build a composite T2DM/complications risk score to be validated in a prospective cohort.

ClusterSignificance: a bioconductor package facilitating statistical analysis of class cluster separations in dimensionality reduced data

Summary

Multi-dimensional data generated via high-throughput experiments is increasingly used in conjunction with dimensionality reduction methods to ascertain if resulting separations of the data correspond with known classes. This is particularly useful to determine if a subset of the variables, e.g. genes in a specific pathway, alone can separate samples into these established classes. Despite this, the evaluation of class separations is often subjective and performed via visualization. Here we present the ClusterSignificance package; a set of tools designed to assess the statistical significance of class separations downstream of dimensionality reduction algorithms. In addition, we demonstrate the design and utility of the ClusterSignificance package and utilize it to determine the importance of long non-coding RNA expression in the identity of multiple hematological malignancies.

Availability and implementation

ClusterSignificance is an R package available via Bioconductor (https://bioconductor.org/packages/ClusterSignificance) under GPL-3.

Contact

dan.grander@ki.se.

Supplementary information

Supplementary data are available at Bioinformatics online.

A novel approach for dimension reduction of microarray

This paper proposes a new hybrid search technique for feature (gene) selection (FS) using Independent component analysis (ICA) and Artificial Bee Colony (ABC) called ICA+ABC, to select informative genes based on a Naïve Bayes (NB) algorithm. An important trait of this technique is the optimization of ICA feature vector using ABC. ICA+ABC is a hybrid search algorithm that combines the benefits of extraction approach, to reduce the size of data and wrapper approach, to optimize the reduced feature vectors. This hybrid search technique is facilitated by evaluating the performance of ICA+ABC on six standard gene expression datasets of classification. Extensive experiments were conducted to compare the performance of ICA+ABC with the results obtained from recently published Minimum Redundancy Maximum Relevance (mRMR) +ABC algorithm for NB classifier. Also to check the performance that how ICA+ABC works as feature selection with NB classifier, compared the combination of ICA with popular filter techniques and with other similar bio inspired algorithm such as Genetic Algorithm (GA) and Particle Swarm Optimization (PSO). The result shows that ICA+ABC has a significant ability to generate small subsets of genes from the ICA feature vector, that significantly improve the classification accuracy of NB classifier compared to other previously suggested methods.

Modified Mahalanobis Taguchi System for Imbalance Data Classification

The Mahalanobis Taguchi System (MTS) is considered one of the most promising binary classification algorithms to handle imbalance data. Unfortunately, MTS lacks a method for determining an efficient threshold for the binary classification. In this paper, a nonlinear optimization model is formulated based on minimizing the distance between MTS Receiver Operating Characteristics (ROC) curve and the theoretical optimal point named Modified Mahalanobis Taguchi System (MMTS). To validate the MMTS classification efficacy, it has been benchmarked with Support Vector Machines (SVMs), Naive Bayes (NB), Probabilistic Mahalanobis Taguchi Systems (PTM), Synthetic Minority Oversampling Technique (SMOTE), Adaptive Conformal Transformation (ACT), Kernel Boundary Alignment (KBA), Hidden Naive Bayes (HNB), and other improved Naive Bayes algorithms. MMTS outperforms the benchmarked algorithms especially when the imbalance ratio is greater than 400. A real life case study on manufacturing sector is used to demonstrate the applicability of the proposed model and to compare its performance with Mahalanobis Genetic Algorithm (MGA).

Texture Feature Extraction and Analysis for Polyp Differentiation via Computed Tomography Colonography

Image textures in computed tomography colonography (CTC) have great potential for differentiating non-neoplastic from neoplastic polyps and thus can advance the current CTC detection-only paradigm to a new level with diagnostic capability. However, image textures are frequently compromised, particularly in low-dose CT imaging. Furthermore, texture feature extraction may vary, depending on the polyp spatial orientation variation, resulting in variable results. To address these issues, this study proposes an adaptive approach to extract and analyze the texture features for polyp differentiation. Firstly, derivative (e.g. gradient and curvature) operations are performed on the CT intensity image to amplify the textures with adequate noise control. Then Haralick co-occurrence matrix (CM) is used to calculate texture measures along each of the 13 directions (defined by the first and second order image voxel neighbors) through the polyp volume in the intensity, gradient and curvature images. Instead of taking the mean and range of each CM measure over the 13 directions as the so-called Haralick texture features, Karhunen-Loeve transform is performed to map the 13 directions into an orthogonal coordinate system so that the resulted texture features are less dependent on the polyp orientation variation. These simple ideas for amplifying textures and stabilizing spatial variation demonstrated a significant impact for the differentiating task by experiments using 384 polyp datasets, of which 52 are non-neoplastic polyps and the rest are neoplastic polyps. By the merit of area under the curve of receiver operating characteristic, the innovative ideas achieved differentiation capability of 0.8016, indicating the CTC diagnostic feasibility.

Volumetric texture features from higher-order images for diagnosis of colon lesions via CT colonography

PURPOSE

Differentiation of colon lesions according to underlying pathology, e.g., neoplastic and non-neoplastic lesions, is of fundamental importance for patient management. Image intensity-based textural features have been recognized as useful biomarker for the differentiation task. In this paper, we introduce texture features from higher-order images, i.e., gradient and curvature images, beyond the intensity image, for that task.

METHODS

Based on the Haralick texture analysis method, we introduce a virtual pathological model to explore the utility of texture features from high-order differentiations, i.e., gradient and curvature, of the image intensity distribution. The texture features were validated on a database consisting of 148 colon lesions, of which 35 are non-neoplastic lesions, using the support vector machine classifier and the merit of area under the curve (AUC) of the receiver operating characteristics.

RESULTS

The AUC of classification was improved from 0.74 (using the image intensity alone) to 0.85 (by also considering the gradient and curvature images) in differentiating the neoplastic lesions from non-neoplastic ones, e.g., hyperplastic polyps from tubular adenomas, tubulovillous adenomas and adenocarcinomas.

CONCLUSIONS

The experimental results demonstrated that texture features from higher-order images can significantly improve the classification accuracy in pathological differentiation of colorectal lesions. The gain in differentiation capability shall increase the potential of computed tomography colonography for colorectal cancer screening by not only detecting polyps but also classifying them for optimal polyp management for the best outcome in personalized medicine.