A hierarchical multilabel graph attention network method to predict the deterioration paths of chronic hepatitis B patients.

OBJECTIVE: Estimating the deterioration paths of chronic hepatitis B (CHB) patients is critical for physicians' decisions and patient management. A novel, hierarchical multilabel graph attention-based method aims to predict patient deterioration paths more effectively. Applied to a CHB patient data set, it offers strong predictive utilities and clinical value. MATERIALS AND METHODS: The proposed method incorporates patients' responses to medications, diagnosis event sequences, and outcome dependencies to estimate deterioration paths. From the electronic health records maintained by a major healthcare organization in Taiwan, we collect clinical data about 177 959 patients diagnosed with hepatitis B virus infection. We use this sample to evaluate the proposed method's predictive efficacy relative to 9 existing methods, as measured by precision, recall, F-measure, and area under the curve (AUC). RESULTS: We use 20% of the sample as holdouts to test each method's prediction performance. The results indicate that our method consistently and significantly outperforms all benchmark methods. It attains the highest AUC, with a 4.8% improvement over the best-performing benchmark, as well as 20.9% and 11.4% improvements in precision and F-measures, respectively. The comparative results demonstrate that our method is more effective for predicting CHB patients' deterioration paths than existing predictive methods. DISCUSSION AND CONCLUSION: The proposed method underscores the value of patient-medication interactions, temporal sequential patterns of distinct diagnosis, and patient outcome dependencies for capturing dynamics that underpin patient deterioration over time. Its efficacious estimates grant physicians a more holistic view of patient progressions and can enhance their clinical decision-making and patient management.

Predictive Analytics for Care and Management of Patients With Acute Diseases: Deep Learning-Based Method to Predict Crucial Complication Phenotypes.

BACKGROUND: Acute diseases present severe complications that develop rapidly, exhibit distinct phenotypes, and have profound effects on patient outcomes. Predictive analytics can enhance physicians' care and management of patients with acute diseases by predicting crucial complication phenotypes for a timely diagnosis and treatment. However, effective phenotype predictions require several challenges to be overcome. First, patient data collected in the early stages of an acute disease (eg, clinical data and laboratory results) are less informative for predicting phenotypic outcomes. Second, patient data are temporal and heterogeneous; for example, patients receive laboratory tests at different time intervals and frequencies. Third, imbalanced distributions of patient outcomes create additional complexity for predicting complication phenotypes. OBJECTIVE: To predict crucial complication phenotypes among patients with acute diseases, we propose a novel, deep learning-based method that uses recurrent neural network-based sequence embedding to represent disease progression while considering temporal heterogeneities in patient data. Our method incorporates a latent regulator to alleviate data insufficiency constraints by accounting for the underlying mechanisms that are not observed in patient data. The proposed method also includes cost-sensitive learning to address imbalanced outcome distributions in patient data for improved predictions. METHODS: From a major health care organization in Taiwan, we obtained a sample of 10,354 electronic health records that pertained to 6545 patients with peritonitis. The proposed method projects these temporal, heterogeneous, and clinical data into a substantially reduced feature space and then incorporates a latent regulator (latent parameter matrix) to obviate data insufficiencies and account for variations in phenotypic expressions. Moreover, our method employs cost-sensitive learning to further increase the predictive performance. RESULTS: We evaluated the efficacy of the proposed method for predicting two hepatic complication phenotypes in patients with peritonitis: acute hepatic encephalopathy and hepatorenal syndrome. The following three benchmark techniques were evaluated: temporal multiple measurement case-based reasoning (MMCBR), temporal short long-term memory (T-SLTM) networks, and time fusion convolutional neural network (CNN). For acute hepatic encephalopathy predictions, our method attained an area under the curve (AUC) value of 0.82, which outperforms temporal MMCBR by 64%, T-SLTM by 26%, and time fusion CNN by 26%. For hepatorenal syndrome predictions, our method achieved an AUC value of 0.64, which is 29% better than that of temporal MMCBR (0.54). Overall, the evaluation results show that the proposed method significantly outperforms all the benchmarks, as measured by recall, F-measure, and AUC while maintaining comparable precision values. CONCLUSIONS: The proposed method learns a short-term temporal representation from patient data to predict complication phenotypes and offers greater predictive utilities than prevalent data-driven techniques. This method is generalizable and can be applied to different acute disease (illness) scenarios that are characterized by insufficient patient clinical data availability, temporal heterogeneities, and imbalanced distributions of important patient outcomes.

A Deep Learning-Based Unsupervised Method to Impute Missing Values in Patient Records for Improved Management of Cardiovascular Patients.

Physicians increasingly depend on electronic health records (EHRs) to manage their patients. However, many patient records have substantial missing values that pose a fundamental challenge to their clinical use. To address this prevailing challenge, we propose an unsupervised deep learning-based method that can facilitate physicians' use of EHRs to improve their management of cardiovascular patients. By building on the deep autoencoder framework, we develop a novel method to impute missing values in patient records. To demonstrate its clinical applicability and values, we use data from cardiovascular patients and evaluate the proposed method's imputation effectiveness and predictive efficacy, in comparison with six prevalent benchmark techniques. The proposed method can impute missing values and predict important patient outcomes more effectively than all the benchmark techniques. This study reinforces the importance of adequately addressing missing values in patient records. It further illustrates how effective imputations can enable greater predictive efficacy with regard to important patient outcomes, which are crucial to the use of EHRs and health analytics for improved patient management. Supported by the complete data imputed by the proposed method, physicians can make timely patient outcome estimations (predictions) and therapeutic treatment assessments.

A deep learning-based, unsupervised method to impute missing values in electronic health records for improved patient management.

Electronic health records (EHRs) often suffer missing values, for which recent advances in deep learning offer a promising remedy. We develop a deep learning-based, unsupervised method to impute missing values in patient records, then examine its imputation effectiveness and predictive efficacy for peritonitis patient management. Our method builds on a deep autoencoder framework, incorporates missing patterns, accounts for essential relationships in patient data, considers temporal patterns common to patient records, and employs a novel loss function for error calculation and regularization. Using a data set of 27,327 patient records, we perform a comparative evaluation of the proposed method and several prevalent benchmark techniques. The results indicate the greater imputation performance of our method relative to all the benchmark techniques, recording 5.3-15.5% lower imputation errors. Furthermore, the data imputed by the proposed method better predict readmission, length of stay, and mortality than those obtained from any benchmark techniques, achieving 2.7-11.5% improvements in predictive efficacy. The illustrated evaluation indicates the proposed method's viability, imputation effectiveness, and clinical decision support utilities. Overall, our method can reduce imputation biases and be applied to various missing value scenarios clinically, thereby empowering physicians and researchers to better analyze and utilize EHRs for improved patient management.

Predicting hepatocellular carcinoma recurrences: A data-driven multiclass classification method incorporating latent variables.

Hepatocellular carcinoma (HCC), a malignant form of cancer, is frequently treated with surgical resections, which have relatively high recurrence rates. Effective recurrence predictions enable physicians' timely detections and adequate therapeutic measures that can greatly improve patient care and outcomes. Toward that end, predictions of early versus late HCC recurrences should be considered separately to reflect their distinct onset time horizons, clinical causes, underlying clinical etiology, and pathogenesis. We propose a novel Bayesian network-based method to predict different HCC recurrence outcomes by considering the respective recurrence evolution paths. Typical patient information obtained in early stages is insufficiently informative to predict recurrence outcomes accurately, due to the lack of subsequent patient progression information. Our method alleviates such information deficiency constraints by incorporating an independent latent variable, dominant recurrence type, to regulate recurrence outcome predictions (early, late, or no recurrence). We use a real-world HCC data set to evaluate the proposed method, relative to three prevalent benchmark techniques. Overall, the results show that our method consistently and significantly outperforms all the benchmark techniques in terms of accuracy, precision, recall, and F-measures. For increased robustness, we use another data set to perform an out-of-sample evaluation and obtain similar results. This study thus contributes to HCC recurrence research and offers several implications for clinical practice.

A preclustering-based ensemble learning technique for acute appendicitis diagnoses.

OBJECTIVE: Acute appendicitis is a common medical condition, whose effective, timely diagnosis can be difficult. A missed diagnosis not only puts the patient in danger but also requires additional resources for corrective treatments. An acute appendicitis diagnosis constitutes a classification problem, for which a further fundamental challenge pertains to the skewed outcome class distribution of instances in the training sample. A preclustering-based ensemble learning (PEL) technique aims to address the associated imbalanced sample learning problems and thereby support the timely, accurate diagnosis of acute appendicitis. MATERIALS AND METHODS: The proposed PEL technique employs undersampling to reduce the number of majority-class instances in a training sample, uses preclustering to group similar majority-class instances into multiple groups, and selects from each group representative instances to create more balanced samples. The PEL technique thereby reduces potential information loss from random undersampling. It also takes advantage of ensemble learning to improve performance. We empirically evaluate this proposed technique with 574 clinical cases obtained from a comprehensive tertiary hospital in southern Taiwan, using several prevalent techniques and a salient scoring system as benchmarks. RESULTS: The comparative results show that PEL is more effective and less biased than any benchmarks. The proposed PEL technique seems more sensitive to identifying positive acute appendicitis than the commonly used Alvarado scoring system and exhibits higher specificity in identifying negative acute appendicitis. In addition, the sensitivity and specificity values of PEL appear higher than those of the investigated benchmarks that follow the resampling approach. Our analysis suggests PEL benefits from the more representative majority-class instances in the training sample. According to our overall evaluation results, PEL records the best overall performance, and its area under the curve measure reaches 0.619. CONCLUSION: The PEL technique is capable of addressing imbalanced sample learning associated with acute appendicitis diagnosis. Our evaluation results suggest PEL is less biased toward a positive or negative class than the investigated benchmark techniques. In addition, our results indicate the overall effectiveness of the proposed technique, compared with prevalent scoring systems or salient classification techniques that follow the resampling approach.

Human CD14hi monocytes and myeloid dendritic cells provide a cell contact-dependent costimulatory signal for early CD40 ligand expression.

CD40L on CD4(+) T cells plays a vital role in the activation of antigen-presenting cells, thus catalyzing a positive feedback loop for T-cell activation. Despite the pivotal juxtaposition of CD40L between antigen-presenting cells and T-cell activation, only a T-cell receptor stimulus is thought to be required for early CD40L surface expression. We show, for the first time, that CD40L expression on peripheral blood CD4(+) T cells is highly dependent on a cell-cell interaction with CD14(hi)CD16(-) monocytes. Interactions with ICAM-1, LFA-3, and to a lesser extent CD80/CD86 contribute to this enhancement of CD40L expression but are not themselves sufficient. The contact-mediated increase in CD40L expression is dependent on new mRNA and protein synthesis. Circulating myeloid dendritic cells also possess this costimulatory activity. By contrast, CD14(lo)CD16(+) monocytes, plasmacytoid dendritic cells, B-cell lymphoma lines, and resting, activated, and Epstein-Barr virus-immortalized primary B cells all lack the capacity to up-regulate early CD40L. The latter indicates that a human B cell cannot activate its cognate T cell to deliver CD40L-mediated help. This finding has functional implications for the role of biphasic CD40L expression, suggesting that the early phase is associated with antigen-presenting cell activation, whereas the late phase is related to B-cell activation.

Immunization with DNA topoisomerase I induces autoimmune responses but not scleroderma-like pathologies in mice.

OBJECTIVE: Anti-DNA topoisomerase I (anti-topo-I) antibody is a marker of systemic sclerosis (SSc). Anti-topo-I antibody levels are positively correlated with both disease severity and activity. However, its pathogenic role in SSc remains unclear. We investigated whether induction of an autoreactive antibody response is directly pathogenic in mice. METHODS: Autoimmune responses were induced in mice immunized with human recombinant topo-I (rhutopo-I). Both humoral and T cell-mediated autoimmune responses were assessed. Necropsy analyses were performed to determine pathologic changes in immunized mice. RESULTS: Autoimmune prone SJL and non-obese diabetic mice developed higher humoral autoreactive responses against mouse topo-I than did BALB/c and C56BL/6 mice. Splenic T cells also showed proliferative responses and interferon-gamma secretion in response to rhutopo-I. However, serum anti-topo-I antibody levels declined 2 months after the initial immunization. Neither weight loss nor dermal thickening was observed in mice during a followup period of 9 months. Whole-body necropsy analyses, including skin, lung, heart, kidney, gastrointestinal tract, and joints, showed no typical findings of human SSc. Coadministration of anti-CD25 and anti-CTLA-4 antibody with the initial immunization resulted in higher titers of anti-topo-I antibody, but these mice also did not develop SSc-like pathologic features. Development of an anti-topo-I response was not associated with acceleration of the recognized abnormalities in tight-skin mice. CONCLUSION: Although tolerance was broken and anti-topo-I antibody was induced by immunization with rhutopo-I in mice, induction of this antibody was not sufficient to induce SSc-like disease.

Clinical subsets, skin thickness progression rate, and serum antibody levels in systemic sclerosis patients with anti-topoisomerase I antibody.

OBJECTIVE: To describe the clinical and laboratory features and natural history of the disease in systemic sclerosis (SSc; scleroderma) patients with anti-topoisomerase I (anti-topo I) antibody who have different skin thickness progression rates (STPRs). METHODS: SSc patients (n = 212) who were anti-topo I antibody positive were divided into 5 subgroups based on STPRs. Skin thickness was measured using the modified Rodnan skin thickness score (MRSS). Anti-topo I IgG antibody levels were determined. RESULTS: Sixty patients who were anti-topo I antibody positive had diffuse cutaneous SSc (dcSSc) with rapid progression, 82 had dcSSC with intermediate progression, and 29 had dcSSc with slow progression, 14 had limited cutaneous SSc (lcSSc) that became dcSSc, and 27 had lcSSc that did not change throughout. Patients beginning with lcSSc were younger at disease onset and had longer disease duration when diagnosed as having SSc. Interstitial lung disease was common and was equally distributed across the subgroups. Renal crisis occurred most often in patients with rapid progression (22%) and was absent in lcSSc patients. Cardiac involvement was most frequent in the dcSSc subgroups. Both kidney and heart disease occurred most often within 3 years after the onset of skin thickening. The 10-year cumulative survival rate was <40% for patients with rapid and intermediate progression. Renal and cardiac causes of death were disproportionately frequent in these 2 subgroups. Anti-topo I antibody levels correlated with the STPR and the MRSS. CONCLUSION: Anti-topo I antibody-positive patients with SSc with a rapid STPR have reduced survival rates, primarily due to early and often fatal renal and cardiac involvement. Anti-topo I antibody levels parallel the MRSS at the first visit and the STPR. This information is important for managing physicians and researchers planning clinical trials involving patients with early dcSSc.

System for infectious disease information sharing and analysis: design and evaluation.

Motivated by the importance of infectious disease informatics (IDI) and the challenges to IDI system development and data sharing, we design and implement BioPortal, a Web-based IDI system that integrates cross-jurisdictional data to support information sharing, analysis, and visualization in public health. In this paper, we discuss general challenges in IDI, describe BioPortal's architecture and functionalities, and highlight encouraging evaluation results obtained from a controlled experiment that focused on analysis accuracy, task performance efficiency, user information satisfaction, system usability, usefulness, and ease of use.

T cell lines from systemic sclerosis patients and healthy controls recognize multiple epitopes on DNA topoisomerase I.

Anti-DNA topoisomerase I (topo-I) antibodies are exclusively detected in patients with systemic sclerosis (SSc). Participation of this topo-I-specific autoimmune response in the pathogenesis of SSc has been actively investigated, but remains unproven. Here we characterized the peripheral T cell proliferative response to recombinant topo-I (rtopo-I) in 16 SSc patients with circulating anti-topo-I antibody. A low level (cpm < 2000) T cell proliferation (SI > 3) was detected in 6 (38%) of 16 patients. This low level response was similar to those previously observed in healthy controls. We established 56 topo-I-specific T cell lines recognizing 13 distinct T cell epitopes on topo-I from 4 SSc patients and 2 healthy controls. These T cell lines were established from in vitro activated PBMC (CD25(+)) by rtopo-I antigen. However they did not have the phenotype of regulatory T cells. Notably, 40 (71%) of the 56 T cell lines recognizing a common epitope were established from one patient. DNA sequencing of the T cell receptor cDNA produced an identical sequence indicating these T cells were from a single topo-I-specific T cell precursor. These results suggest that topo-I-specific T cells can become clonally expanded in some patients and may contribute to the pathogenesis of this disease.

Condom use by women recently diagnosed with a sexually transmitted infection: effects of study methodology on the apparent influence of hormonal/surgical contraception.

OBJECTIVE/GOAL: The objective of this study was to investigate potential predictors of consistent condom use (CCU), including the influence of hormonal contraception/surgical sterilization (HC/SS). STUDY: Regression methods were used to predict CCU and other measures of CU among 214 sexually active, 18- to 45-year-old women previously diagnosed with a sexually transmitted infection. RESULTS: CCU was significantly associated with younger age, African American ethnicity, having casual partners, recent HIV testing, condom use self-efficacy, and concern about partner relationship. HC/SS was not significantly associated with the likelihood of CCU, before (HC/SS, 21.3%, non-HC/SS, 25.3%; odds ratio [OR], 0.798; P=0.4914) or after (OR, 1.209; P=0.5995) controlling for confounders (age, ethnicity, casual partners). Controlling for age and ethnicity eliminated initial significant or near-significant inverse associations between HC/SS and 3 alternative measures of interval condom use ("any use," "number of unprotected acts," "proportion condom-protected contacts") and substantially diminished the association between HC/SS and "condom use at last sex." CONCLUSIONS: Choice of condom use measure and control of confounding variables can substantially affect results when studying potential predictors of condom use such as HC/SS.

Molecular recognition patterns of serum anti-DNA topoisomerase I antibody in systemic sclerosis.

Autoreactive anti-DNA topoisomerase I (anti-Topo I) Abs are commonly detected in sera of systemic sclerosis (SSc) patients. Our studies have established a positive correlation between the levels of serum anti-Topo I Abs and both disease severity and activity of SSc. The molecular targets of anti-Topo I Ab on Topo I domains remain to be further defined. In this report, we studied the molecular recognition pattern of serum anti-Topo I Ab in 52 SSc patients. The highest reactivity of serum anti-Topo I Abs was against the core subdomains I and II (aa 207-441) and, to a lesser extent, against the core subdomain III (aa 433-636) of Topo I. The linker domain (aa 636-712) and the C-terminal domain (aa 713-765) had much less reactivity than the core domain (aa 207-636). Strikingly, very little reactivity was directed against the N-terminal domain (aa 1-213) by serum anti-Topo I Ab. This molecular recognition pattern was consistent among all SSc serum samples studied. Results from patients with serial serum samples indicated that this pattern remained unchanged over time. Interestingly, some naive B cells from healthy controls, upon transformation by EBV, produced IgM Abs against Topo I. These Abs had low affinity for Topo I and reacted equally to all domains of Topo I. The molecular recognition pattern of serum anti-Topo I Ab in SSc suggests the presence of a unique antigenic stimulation in vivo in this disease.

Escherichia coli expressing recombinant antigen and listeriolysin O stimulate class I-restricted CD8+ T cells following uptake by human APC.

Vaccination against cancer or intracellular pathogens requires stimulation of class I-restricted CD8(+) T cells. It is therefore important to develop Ag delivery vectors that will promote cross-presentation by APCs and stimulate appropriate inflammatory responses. Toward this goal, we tested the potential of Escherichia coli as an Ag delivery vector in in vitro human culture. Bacteria expressing enhanced green fluorescent protein were internalized efficiently by dendritic cells, as shown by flow cytometry and fluorescence microscopy. Phenotypic changes in DC were observed, including up-regulation of costimulatory molecules and IL-12p40 production. We tested whether bacteria expressing recombinant Ags could stimulate human T cells using the influenza matrix protein as a model Ag. Specific responses against an immunodominant epitope were seen using IFN-gamma ELISPOT assays when the matrix protein was coexpressed with listeriolysin O, but not when expressed alone. THP-1 macrophages were also capable of stimulating T cells after uptake of bacteria, but showed slower kinetics and lower overall levels of T cell stimulation than dendritic cells. Increased phagocytosis of bacteria induced by differentiation of THP-1 increased their ability to stimulate T cells, as did opsonization. Presentation was blocked by proteasome inhibitors, but not by lysosomal protease inhibitors leupeptin and E64. These results demonstrate that recombinant E. coli can be engineered to direct Ags to the cytosol of human phagocytic APCs, and suggest possible vaccine strategies for generating CD8(+) T cell responses against pathogens or tumors.

Rapid and extensive membrane reorganization by dendritic cells following exposure to bacteria revealed by high-resolution imaging.

Using live cell imaging, we demonstrate that immature dendritic cells (DC) derived from human peripheral blood monocytes undergo pronounced morphologic changes in vitro within minutes of exposure to unopsonized Escherichia coli, developing extensive membrane veils that efficiently capture additional bacteria. Internalization does not occur in the veils, but instead, bacteria are transported to the central region of the cell, where they sink directly into the plasma membrane. In contrast, exposure to polystyrene beads does not induce notable changes in cell morphology, and DC do not efficiently capture beads when introduced alone or mixed with bacteria. Long dendritic processes were also visualized in some cells that allowed capture of clumps of bacteria at a distance of more than 100 microm. These results demonstrate that immature DC can distinguish between inert particles and bacteria and alter their shape and phagocytic capacity in response to the latter.

Correlation of serum anti-DNA topoisomerase I antibody levels with disease severity and activity in systemic sclerosis.

OBJECTIVE: To investigate correlations between serum levels of topoisomerase I-specific antibody (anti-topo I) and clinical features of systemic sclerosis (SSc), including disease severity (the total skin score [TSS]) and disease activity. METHODS: Using highly sensitive enzyme-linked immunosorbent assays, we measured the levels of anti-topo I antibody, including total IgG, individual IgG subclasses, and IgA, and analyzed their correlations with the TSS in 59 patients with SSc, all of whom had diffuse cutaneous involvement. Serial serum samples were obtained from 11 of these patients. RESULTS: The titers of anti-topo I antibody, including IgG and IgA, were positively correlated with the TSS, a measure of SSc disease severity. In 8 of the 11 patients from whom serial serum samples were obtained, changes in the levels of both IgG and IgA, when detectable, paralleled changes in the TSS. In 3 patients, an increasing anti-topo I IgG level preceded an increase in the TSS. The level of each IgG subclass also correlated with and tended to parallel the TSS. The patients with very active disease had higher mean IgG (P < 0.001) and IgA (P < 0.05) titers than did those with inactive disease. CONCLUSION: Serum levels of anti-topo I antibody correlate positively with disease severity and disease activity in SSc.

The male silkworm moth (Antheraea pernyi) is a key ingredient in hu-bao and sheng-bao for specific prolongation of the life-span of the male fruit fly (Drosophila melanogaster).

It is well established in Traditional Chinese Medicine that certain natural products, such as male silkworm moths, have different therapeutic effects on men than on women. These natural products have been used as dietary supplements specifically formulated for men or for women. However, this presumed sex-specific effect of certain natural products has not yet been confirmed experimentally with animal models or in human clinical trials. Here, using the fruit fly (Drosophila melanogaster) as a longevity model, we examined the effect of hu-bao (HB) and seng-bao (SB), two marketed health products made from a mixture of natural ingredients. Our results convincingly demonstrate that the effect of HB and SB are indeed specific for the male fly. The life-span of the male was significantly increased when HB or SB was added to the culture medium. In contrast, neither HB nor SB had much effect on the female fly. Upon removal of the male silkworm moth ingredient from HB or SB, the life-span prolongation effect of HB and SB was drastically diminished. Only with the addition of the male silkworm moth did the culture medium show a statistically significant life-span prolongation effect. This result suggests that the male silkworm moth is a key ingredient, in combination with other components, for specific prolongation of the life-span of male flies.