DeepMNF: Deep Multimodal Neuroimaging Framework for Diagnosing Autism Spectrum Disorder.

The growing prevalence of neurological disorders, e.g., Autism Spectrum Disorder (ASD), demands robust computer-aided diagnosis (CAD) due to the diverse symptoms which require early intervention, particularly in young children. The absence of a benchmark neuroimaging diagnostics paves the way to study transitions in the brain's anatomical structure and neurological patterns associated with ASD. The existing CADs take advantage of the large-scale baseline dataset from the Autism Brain Imaging Data Exchange (ABIDE) repository to improve diagnostic performance, but the involvement of multisite data also amplifies the variabilities and heterogeneities that hinder satisfactory results. To resolve this problem, we propose a Deep Multimodal Neuroimaging Framework (DeepMNF) that employs Functional Magnetic Resonance Imaging (fMRI) and Structural Magnetic Resonance Imaging (sMRI) to integrate cross-modality spatiotemporal information by exploiting 2-dimensional time-series data along with 3-dimensional images. The purpose is to fuse complementary information that increases group differences and homogeneities. To the best of our knowledge, our DeepMNF achieves superior validation performance than the best reported result on the ABIDE-1 repository involving datasets from all available screening sites. In this work, we also demonstrate the performance of the studied modalities in a single model as well as their possible combinations to develop the multimodal framework.

Correlation-Aware Spatial-Temporal Graph Learning for Multivariate Time-Series Anomaly Detection.

Multivariate time-series anomaly detection is critically important in many applications, including retail, transportation, power grid, and water treatment plants. Existing approaches for this problem mostly employ either statistical models which cannot capture the nonlinear relations well or conventional deep learning (DL) models e.g., convolutional neural network (CNN) and long short-term memory (LSTM) that do not explicitly learn the pairwise correlations among variables. To overcome these limitations, we propose a novel method, correlation-aware spatial-temporal graph learning (termed ), for time-series anomaly detection. explicitly captures the pairwise correlations via a correlation learning (MTCL) module based on which a spatial-temporal graph neural network (STGNN) can be developed. Then, by employing a graph convolution network (GCN) that exploits one-and multihop neighbor information, our STGNN component can encode rich spatial information from complex pairwise dependencies between variables. With a temporal module that consists of dilated convolutional functions, the STGNN can further capture long-range dependence over time. A novel anomaly scoring component is further integrated into to estimate the degree of an anomaly in a purely unsupervised manner. Experimental results demonstrate that can detect and diagnose anomalies effectively in general settings as well as enable early detection across different time delays. Our code is available at https://github.com/huankoh/CST-GL.

Decaffeinated green tea extract as a nature-derived antibiotic alternative: An application in antibacterial nano-thin coating on medical implants.

Plant-derived polyphenols have emerged as molecular building blocks for biomedical architectures. However, the isolation of polyphenols from other components requires labor-intensive procedures, which increases costs and often raises environmental concerns. Here, we suggest that decaffeination can be a convenient and cost-effective method for enhancing the antibacterial performance of polyphenol-rich tea extracts. As a demonstration, we compared the properties of a nano-thin coating made of decaffeinated (dGT coating) and raw green tea extract (GT coating). The dGT coating exhibited enhanced antibacterial performance with regard to bacterial killing and prevention of bacterial attachment compared with the GT coating. Moreover, the chemical reactivity of the dGT coating was further utilized for secondary modifications, which enhanced the overall antibacterial performance of the modified surface. Given its intrinsic low toxicity, we envision that the developed antibacterial coating is ready for the next steps toward application in real clinical settings.

An Interpretable Deep Embedding Model for Few and Imbalanced Biomedical Data.

In healthcare, training examples are usually hard to obtain (e.g., cases of a rare disease), or the cost of labelling data is high. With a large number of features ( p) be measured in a relatively small number of samples ( N), the "big p, small N" problem is an important subject in healthcare studies, especially on the genomic data. Another major challenge of effectively analyzing medical data is the skewed class distribution caused by the imbalance between different class labels. In addition, feature importance and interpretability play a crucial role in the success of solving medical problems. Therefore, in this paper, we present an interpretable deep embedding model (IDEM) to classify new data having seen only a few training examples with highly skewed class distribution. IDEM model consists of a feature attention layer to learn the informative features, a feature embedding layer to directly deal with both numerical and categorical features, a siamese network with contrastive loss to compare the similarity between learned embeddings of two input samples. Experiments on both synthetic data and real-world medical data demonstrate that our IDEM model has better generalization power than conventional approaches with few and imbalanced training medical samples, and it is able to identify which features contribute to the classifier in distinguishing case and control.

Mortality Prediction of Patients With Cardiovascular Disease Using Medical Claims Data Under Artificial Intelligence Architectures: Validation Study.

BACKGROUND: Cardiovascular disease (CVD) is the greatest health problem in Australia, which kills more people than any other disease and incurs enormous costs for the health care system. In this study, we present a benchmark comparison of various artificial intelligence (AI) architectures for predicting the mortality rate of patients with CVD using structured medical claims data. Compared with other research in the clinical literature, our models are more efficient because we use a smaller number of features, and this study could help health professionals accurately choose AI models to predict mortality among patients with CVD using only claims data before a clinic visit. OBJECTIVE: This study aims to support health clinicians in accurately predicting mortality among patients with CVD using only claims data before a clinic visit. METHODS: The data set was obtained from the Medicare Benefits Scheme and Pharmaceutical Benefits Scheme service information in the period between 2004 and 2014, released by the Department of Health Australia in 2016. It included 346,201 records, corresponding to 346,201 patients. A total of five AI algorithms, including four classical machine learning algorithms (logistic regression [LR], random forest [RF], extra trees [ET], and gradient boosting trees [GBT]) and a deep learning algorithm, which is a densely connected neural network (DNN), were developed and compared in this study. In addition, because of the minority of deceased patients in the data set, a separate experiment using the Synthetic Minority Oversampling Technique (SMOTE) was conducted to enrich the data. RESULTS: Regarding model performance, in terms of discrimination, GBT and RF were the models with the highest area under the receiver operating characteristic curve (97.8% and 97.7%, respectively), followed by ET (96.8%) and LR (96.4%), whereas DNN was the least discriminative (95.3%). In terms of reliability, LR predictions were the least calibrated compared with the other four algorithms. In this study, despite increasing the training time, SMOTE was proven to further improve the model performance of LR, whereas other algorithms, especially GBT and DNN, worked well with class imbalanced data. CONCLUSIONS: Compared with other research in the clinical literature involving AI models using claims data to predict patient health outcomes, our models are more efficient because we use a smaller number of features but still achieve high performance. This study could help health professionals accurately choose AI models to predict mortality among patients with CVD using only claims data before a clinic visit.

Identification of novel CD44v6-binding peptides that block CD44v6 and deliver a pro-apoptotic peptide to tumors to inhibit tumor growth and metastasis in mice.

CD44v6, a splice variant of the cell surface glycoprotein CD44, acts as a co-receptor for c-Met and is upregulated in tumors with high metastatic potential. Methods: We screened a phage-displayed peptide library for peptides that selectively bind to CD44v6-overexpressing cells and exploited them to block CD44v6 and deliver a pro-apoptotic peptide to tumors for cancer therapy. Results: CNLNTIDTC (NLN) and CNEWQLKSC (NEW) peptides bound preferentially to CD44v6-high cells than to CD44v6-low cells. The binding affinities of NLN and NEW to CD44v6 protein were 253 ± 79 and 85 ± 18 nM, respectively. Peptide binding to CD44v6-high cells was inhibited by the knockdown of CD44v6 gene expression and competition with an anti-CD44v6 antibody. A pull-down assay with biotin-labeled peptides enriched CD44v6 from cell lysates. NLN and NEW induced CD44v6 internalization and inhibited hepatocyte growth factor-induced c-Met internalization, c-Met and Erk phosphorylation, and cell migration and invasion. In mice harboring tumors, intravenously administered NLN and NEW homed to the tumors and inhibited metastasis to the lungs. When combined with crizotinib, a c-Met inhibitor, treatment with each peptide inhibited metastatic growth more efficiently than each peptide or crizotinib alone. In addition, KLAKLAKKLAKLAK pro-apoptotic peptide guided by NLN (NLN-KLA) or NEW (NEW-KLA) killed tumor cells and inhibited tumor growth and metastasis. No significant systemic side effects were observed after treatments. Conclusions: These results suggest that NLN and NEW are promising metastasis-inhibiting peptide therapeutics and targeting moieties for CD44v6-expressing metastases.

Personalized Monitoring Model for Electrocardiogram Signals: Diagnostic Accuracy Study.

BACKGROUND: Due to the COVID-19 pandemic, the demand for remote electrocardiogram (ECG) monitoring has increased drastically in an attempt to prevent the spread of the virus and keep vulnerable individuals with less severe cases out of hospitals. Enabling clinicians to set up remote patient ECG monitoring easily and determining how to classify the ECG signals accurately so relevant alerts are sent in a timely fashion is an urgent problem to be addressed for remote patient monitoring (RPM) to be adopted widely. Hence, a new technique is required to enable routine and widespread use of RPM, as is needed due to COVID-19. OBJECTIVE: The primary aim of this research is to create a robust and easy-to-use solution for personalized ECG monitoring in real-world settings that is precise, easily configurable, and understandable by clinicians. METHODS: In this paper, we propose a Personalized Monitoring Model (PMM) for ECG data based on motif discovery. Motif discovery finds meaningful or frequently recurring patterns in patient ECG readings. The main strategy is to use motif discovery to extract a small sample of personalized motifs for each individual patient and then use these motifs to predict abnormalities in real-time readings of that patient using an artificial logical network configured by a physician. RESULTS: Our approach was tested on 30 minutes of ECG readings from 32 patients. The average diagnostic accuracy of the PMM was always above 90% and reached 100% for some parameters, compared to 80% accuracy for the Generalized Monitoring Models (GMM). Regardless of parameter settings, PMM training models were generated within 3-4 minutes, compared to 1 hour (or longer, with increasing amounts of training data) for the GMM. CONCLUSIONS: Our proposed PMM almost eliminates many of the training and small sample issues associated with GMMs. It also addresses accuracy and computational cost issues of the GMM, caused by the uniqueness of heartbeats and training issues. In addition, it addresses the fact that doctors and nurses typically do not have data science training and the skills needed to configure, understand, and even trust existing black box machine learning models.

Peptide-based targeted therapeutics and apoptosis imaging probes for cancer therapy.

Peptides have advantages over antibodies in terms of deep tissue penetration, low immunogenicity, and cost-effective production, but they have short circulation time and poor stability in vivo. Peptides have been extensively used as targeting moieties for the delivery of drug-loaded nanoparticles and function as targeted therapeutics in cancer treatment. Here, we review peptides that are exploited as targeted therapeutics in cancer therapy and apoptosis imaging probes for the monitoring of treatment responses.

A Phage Display-Identified Peptide Selectively Binds to Kidney Injury Molecule-1 (KIM-1) and Detects KIM-1-Overexpressing Tumors in vivo.

PURPOSE: This study was carried out to identify a peptide that selectively binds to kidney injury molecule-1 (KIM-1) by screening a phage-displayed peptide library and to use the peptide for the detection of KIM-1overexpressing tumors in vivo. MATERIALS AND METHODS: Biopanning of a phage-displayed peptide library was performed on KIM-1-coated plates. The binding of phage clones, peptides, and a peptide multimer to the KIM-1 protein and KIM-1-overexpressing and KIM-1-low expressing cells was examined by enzyme-linked immunosorbent assay, fluorometry, and flow cytometry. A biotin-peptide multimer was generated using NeutrAvidin. In vivo homing of the peptide to KIM-1-overexpressing and KIM1-low expressing tumors in mice was examined by whole-body fluorescence imaging. RESULTS: A phage clone displaying the CNWMINKEC peptide showed higher binding affinity to KIM-1 and KIM-1-overexpressing 769-P renal tumor cells compared to other phage clones selected after biopanning. The CNWMINKEC peptide and a NeutrAvidin/biotin-CNWMINKEC multimer selectively bound to KIM-1 over albumin and to KIM-1-overexpressing 769-P cells and A549 lung tumor cells compared to KIM-1-low expressing HEK293 normal cells. Co-localization and competition assays using an anti-KIM-1 antibody demonstrated that the binding of the CNWMINKEC peptide to 769-P cells was specifically mediated by KIM-1. The CNWMINKEC peptide was not cytotoxic to cells and was stable for up to 24 hours in the presence of serum. Whole-body fluorescence imaging demonstrated selective homing of the CNWM-INKEC peptide to KIM-1-overexpressing A498 renal tumor compared to KIM1-low expressing HepG2 liver tumor in mice. CONCLUSION: The CNWMINKEC peptide is a promising probe for in vivo imaging and detection of KIM-1‒overexpressing tumors.

Hashing for Adaptive Real-Time Graph Stream Classification With Concept Drifts.

Many applications involve processing networked streaming data in a timely manner. Graph stream classification aims to learn a classification model from a stream of graphs with only one-pass of data, requiring real-time processing in training and prediction. This is a nontrivial task, as many existing methods require multipass of the graph stream to extract subgraph structures as features for graph classification which does not simultaneously satisfy "one-pass" and "real-time" requirements. In this paper, we propose an adaptive real-time graph stream classification method to address this challenge. We partition the unbounded graph stream data into consecutive graph chunks, each consisting of a fixed number of graphs and delivering a corresponding chunk-level classifier. We employ a random hashing function to compress the original node set of graphs in each chunk for fast feature detection when training chunk-level classifiers. Furthermore, a differential hashing strategy is applied to map unlimited increasing features (i.e., cliques) into a fixed-size feature space which is then used as a feature vector for stochastic learning. Finally, the chunk-level classifiers are weighted in an ensemble learning model for graph classification. The proposed method substantially speeds up the graph feature extraction and avoids unbounded graph feature growth. Moreover, it effectively offsets concept drifts in graph stream classification. Experiments on real-world and synthetic graph streams demonstrate that our method significantly outperforms existing methods in both classification accuracy and learning efficiency.

Non-genetic engineering of cytotoxic T cells to target IL-4 receptor enhances tumor homing and therapeutic efficacy against melanoma.

Adoptive transfer of cytotoxic T lymphocytes (CTLs) has been used as an immunotherapy in melanoma. However, the tumor homing and therapeutic efficacy of transferred CTLs against melanoma remain unsatisfactory. Interleukin-4 receptor (IL-4R) is commonly up-regulated in tumors including melanoma. Here, we studied whether IL-4R-targeted CTLs exhibit enhanced tumor homing and therapeutic efficacy against melanoma. CTLs isolated from mice bearing melanomas were non-genetically engineered with IL4RPep-1, an IL-4R-binding peptide, using a membrane anchor composed of dioleylphosphatidylethanolamine. Compared to control CTLs, IL-4R-targeted CTLs showed higher binding to melanoma cells and in vivo tumor homing. They also exerted a more rapid and robust effector response, including increased cytokine secretion and cytotoxicity against melanoma cells and enhanced reprogramming of M2-type macrophages to M1-type macrophages. Moreover, IL-4R-targeted CTLs efficiently inhibited melanoma growth and reversed the immunosuppressive tumor microenvironment. These results suggest that non-genetically engineered CTLs targeting IL-4R have potential as an adoptive T cell therapy against melanoma.

IL4 Receptor-Targeted Proapoptotic Peptide Blocks Tumor Growth and Metastasis by Enhancing Antitumor Immunity.

Cellular cross-talk between tumors and M2-polarized tumor-associated macrophages (TAM) favors tumor progression. Upregulation of IL4 receptor (IL4R) is observed in diverse tumors and TAMs. We tested whether an IL4R-targeted proapoptotic peptide could inhibit tumor progression. The IL4R-binding peptide (IL4RPep-1) preferentially bound to IL4R-expressing tumor cells and M2-polarized macrophages both in vitro and in 4T1 breast tumors in vivo To selectively kill IL4R-expressing cells, we designed an IL4R-targeted proapoptotic peptide, IL4RPep-1-K, by adding the proapoptotic peptide (KLAKLAK)2 to the end of IL4RPep-1. IL4RPep-1-K exerted selective cytotoxicity against diverse IL4R-expressing tumor cells and M2-polarized macrophages. Systemic administration of IL4RPep-1-K inhibited tumor growth and metastasis in 4T1 breast tumor-bearing mice. Interestingly, IL4RPep-1-K treatment increased the number of activated cytotoxic CD8+ T cells while reducing the numbers of immunosuppressive regulatory T cells and M2-polarized TAMs. No significant systemic side effects were observed. These results suggest that IL4R-targeted proapoptotic peptide has potential for treating diverse IL4R-expressing cancers. Mol Cancer Ther; 16(12); 2803-16. ©2017 AACR.

Interleukin-4 receptor-targeted delivery of Bcl-xL siRNA sensitizes tumors to chemotherapy and inhibits tumor growth.

IL-4 receptor (IL-4R) is commonly up-regulated on tumor cells, and interactions between the receptor and Interleukin-4 (IL-4) can induce the expression of anti-apoptotic proteins, including Bcl-xL. This contributes to tumor cell survival and their resistance to chemotherapy. In this study, we exploited IL-4R-targeted delivery of Bcl-xL siRNA to IL-4R-expressing tumor cells in order to sensitize them to chemotherapy. To target IL-4R, an IL-4R-binding peptide, IL4RPep-1, was attached to branched polyethyleneimine-superparamagnetic iron oxide nanoparticles (BPEI-SPION). These nanoparticles were then complexed with Bcl-xL-targeting siRNA. IL-4R-targeted BPEI-SPION/Bcl-xL siRNA more efficiently reduced Bcl-xL gene expression and enhanced cytotoxicity of doxorubicin in MDA-MB231 breast tumor cells compared to untargeted BPEI-SPION/Bcl-xL siRNA. The siRNA was released from the complexes after 15 h of incubation at pH 5.5 and was stable in the complexes up to 72 h in the serum. The IL-4R-targeted BPEI-SPION/siRNA was internalized by cells through IL-4R, successfully escaped the endosomes, and was dispersed into the cytoplasm. Near-infrared fluorescence and magnetic resonance imaging demonstrated that in vivo tumor homing and accumulation of IL-4R-targeted BPEI-SPION/siRNA were both higher than untargeted BPEI-SPION/siRNA. The IL-4R-targeted BPEI-SPION/Bcl-xL siRNA, in combination with doxorubicin, significantly inhibited tumor growth in mice compared to untargeted BPEI-SPION/Bcl-xL siRNA. These results suggest that the IL-4R-targeted delivery of Bcl-xL siRNA to IL-4R-expressing tumors can sensitize tumors to chemotherapy and enhance the efficacy of anti-tumor therapeutics.

A novel method to detect articular chondrocyte death during early stages of osteoarthritis using a non-invasive ApoPep-1 probe.

INTRODUCTION: Current methods for early diagnosis of osteoarthritis (OA) are limited. We assessed whether in vivo detection of chondrocyte death by ApoPep-1 (CQRPPR), a peptide that binds to histone H1 of apoptotic and necrotic cells, could be used to detect the initiation of OA. METHODS: Apoptosis-induced ATDC5 cells were labeled with Annexin V and ApoPep-1. Surgical destabilization of the medial meniscus (DMM) was performed on both knees of 12-week-old male mice and severity of OA was determined by histological analysis according to the Osteoarthritis Research Society International (OARSI) guidelines. At 1, 2, 4, and 8 weeks post-surgery, mice were intravenously injected with fluorescence-labeled ApoPep-1 or control peptide and in vivo imaging was performed within 30 minutes of injection by near-infrared fluorescence (NIRF). Binding of ApoPep-1 to OA joints was demonstrated by ex vivo imaging and immunofluorescent staining using TUNEL and histone H1 and type II collagen antibodies. RESULTS: Strong signals of ApoPep-1 were observed on the apoptotic ATDC5 cells. Knees corresponded to grade II, III, and V OA at 2, 4, and 8 weeks after DMM, respectively. Between 2 and 8 weeks after surgery, the in vivo NIRF signal at OA-ApoPep1-injected joints was consistently stronger than sham-operated or OA-control peptide-injected joints. ApoPep-1, TUNEL, and histone H1 signals were stronger in grade II OA cartilage than sham-operated cartilage when detected by immunofluorescent staining. Type II collagen expression was similar between grade II OA and sham group. CONCLUSION: ApoPep-1 can be used to detect OA in vivo by binding to apoptotic chondrocytes. This is a novel, sensitive, and rapid method which can detect apoptotic cells in OA rodent models soon after its onset.

Enhanced delivery of liposomes to lung tumor through targeting interleukin-4 receptor on both tumor cells and tumor endothelial cells.

A growing body of evidence suggests that pathological lesions express tissue-specific molecular targets or biomarkers within the tissue. Interleukin-4 receptor (IL-4R) is overexpressed in many types of cancer cells, including lung cancer. Here we investigated the properties of IL-4R-binding peptide-1 (IL4RPep-1), a CRKRLDRNC peptide, and its ability to target the delivery of liposomes to lung tumor. IL4RPep-1 preferentially bound to H226 lung tumor cells which express higher levers of IL-4R compared to H460 lung tumor cells which express less IL-4R. Mutational analysis revealed that C1, R2, and R4 residues of IL4RPep-1 were the key binding determinants. IL4RPep-1-labeled liposomes containing doxorubicin were more efficiently internalized in H226 cells and effectively delivered doxorubicin into the cells compared to unlabeled liposomes. In vivo fluorescence imaging of nude mice subcutaneously xenotransplanted with H226 tumor cells indicated that IL4RPep-1-labeled liposomes accumulate more efficiently in the tumor and inhibit tumor growth more effectively compared to unlabeled liposomes. Interestingly, expression of IL-4R was high in vascular endothelial cells of tumor, while little was detected in vascular endothelial cells of control organs including the liver. IL-4R expression in cultured human vascular endothelial cells was also up-regulated when activated by a pro-inflammatory cytokine tumor necrosis factor-α. Moreover, the up-regulation of IL-4R expression was observed in primary human lung cancer tissues. These results indicate that IL-4R-targeting nanocarriers may be a useful strategy to enhance drug delivery through the recognition of IL-4R in both tumor cells and tumor endothelial cells.

Enhanced delivery of T cells to tumor after chemotherapy using membrane-anchored, apoptosis-targeted peptide.

Chemotherapy-induced apoptosis of tumor cells enhances the antigen presentation and sensitizes tumor cells to T cell-mediated cytotoxicity. Here we harnessed the apoptosis of tumor cells as a homing signal for the delivery of T cells to tumor. Jurkat T cells were anchored with ApoPep-1, an apoptosis-targeted peptide ligand, using the biocompatible anchor for membrane (BAM), an oleyl acid derivative. The ApoPep-1-BAM conjugate was efficiently anchored to cell membrane, while little anchoring was obtained with ApoPep-1 alone. The retention period of the ApoPep-1-BAM conjugate on cell membrane was approximately 80 and 40 min in the absence and presence of serum, respectively. ApoPep-1 was resistant to degradation in serum until 2h. The apoptosis-targeted T cells that were anchored with the ApoPep-1-BAM preferentially bound to apoptotic tumor cells over living cells. When intravenously injected into tumor-bearing mice, the number of apoptosis-targeted T cells and in vivo fluorescence signals by the homing of the cells to doxorubicin-treated tumor were higher than those of untargeted T cells. Accumulation of apoptosis-targeted T cells at other organs such as liver was not detected. These results suggest that the chemotherapy-induced apoptosis and subsequent enhancement of T cell delivery to tumor by the membrane anchoring of the apoptosis-targeted peptide could be a novel strategy for cancer immunotherapy.