Artificial neural network prediction of clozapine response with combined pharmacogenetic and clinical data

Development and Validation of a Deep Learning Model for Predicting Treatment Response in Patients With Newly Diagnosed Epilepsy

Importance

Selection of antiseizure medications (ASMs) for epilepsy remains largely a trial-and-error approach. Under this approach, many patients have to endure sequential trials of ineffective treatments until the "right drugs" are prescribed.

Objective

To develop and validate a deep learning model using readily available clinical information to predict treatment success with the first ASM for individual patients.

Design, Setting, and Participants

This cohort study developed and validated a prognostic model. Patients were treated between 1982 and 2020. All patients were followed up for a minimum of 1 year or until failure of the first ASM. A total of 2404 adults with epilepsy newly treated at specialist clinics in Scotland, Malaysia, Australia, and China between 1982 and 2020 were considered for inclusion, of whom 606 (25.2%) were excluded from the final cohort because of missing information in 1 or more variables.

Exposures

One of 7 antiseizure medications.

Main Outcomes and Measures

With the use of the transformer model architecture on 16 clinical factors and ASM information, this cohort study first pooled all cohorts for model training and testing. The model was trained again using the largest cohort and externally validated on the other 4 cohorts. The area under the receiver operating characteristic curve (AUROC), weighted balanced accuracy, sensitivity, and specificity of the model were all assessed for predicting treatment success based on the optimal probability cutoff. Treatment success was defined as complete seizure freedom for the first year of treatment while taking the first ASM. Performance of the transformer model was compared with other machine learning models.

Results

The final pooled cohort included 1798 adults (54.5% female; median age, 34 years [IQR, 24-50 years]). The transformer model that was trained using the pooled cohort had an AUROC of 0.65 (95% CI, 0.63-0.67) and a weighted balanced accuracy of 0.62 (95% CI, 0.60-0.64) on the test set. The model that was trained using the largest cohort only had AUROCs ranging from 0.52 to 0.60 and a weighted balanced accuracy ranging from 0.51 to 0.62 in the external validation cohorts. Number of pretreatment seizures, presence of psychiatric disorders, electroencephalography, and brain imaging findings were the most important clinical variables for predicted outcomes in both models. The transformer model that was developed using the pooled cohort outperformed 2 of the 5 other models tested in terms of AUROC.

Conclusions and Relevance

In this cohort study, a deep learning model showed the feasibility of personalized prediction of response to ASMs based on clinical information. With improvement of performance, such as by incorporating genetic and imaging data, this model may potentially assist clinicians in selecting the right drug at the first trial.

Machine learning for prediction of schizophrenia using genetic and demographic factors in the UK biobank

Machine learning (ML) holds promise for precision psychiatry, but its predictive performance is unclear. We assessed whether ML provided added value over logistic regression for prediction of schizophrenia, and compared models built using polygenic risk scores (PRS) or clinical/demographic factors. LASSO and ridge-penalised logistic regression, support vector machines (SVM), random forests, boosting, neural networks and stacked models were trained to predict schizophrenia, using PRS for schizophrenia (PRS_SZ), sex, parental depression, educational attainment, winter birth, handedness and number of siblings as predictors. Models were evaluated for discrimination using area under the receiver operator characteristic curve (AUROC) and relative importance of predictors using permutation feature importance (PFI). In a secondary analysis, fitted models were tested for association with schizophrenia-related traits which had not been used in model development. Following learning curve analysis, 738 cases and 3690 randomly sampled controls were selected from the UK Biobank. ML models combining all predictors showed the highest discrimination (linear SVM, AUROC = 0.71), but did not significantly outperform logistic regression. AUROC was robust over 100 random resamples of controls. PFI identified PRS_SZ as the most important predictor. Highest variance in fitted models was explained by schizophrenia-related traits including fluid intelligence (most associated: linear SVM), digit symbol substitution (RBF SVM), BMI (XGBoost), smoking status (XGBoost) and deprivation (linear SVM). In conclusion, ML approaches did not provide substantial added value for prediction of schizophrenia over logistic regression, as indexed by AUROC; however, risk scores derived with different ML approaches differ with respect to association with schizophrenia-related traits.

Tardive dyskinesia in Chinese patients with schizophrenia: Prevalence, clinical correlates and relationship with cognitive impairment

OBJECTIVE

Tardive dyskinesia (TD) has a high prevalence and is one of the distressing side effects of antipsychotic medications. Few studies have explored the relationship between TD, clinical correlates, and cognition. The aim of this study was to assess the prevalence, clinical correlates and cognitive impairment of co-occurring TD in Chinese patients with schizophrenia.

METHODS

We recruited 655 patients with chronic schizophrenia who met the DSM-IV diagnostic criteria for schizophrenia and collected clinical and demographic data. All patients were assessed using the Abnormal Involuntary Movement Scale (AIMS) for the severity of TD, Positive and Negative Syndrome Scale (PANSS) for psychopathological symptoms, and Repeatable Battery for the Assessment of Neuropsychological Status (RBANS) for cognition.

RESULTS

The overall TD prevalence was 41.1%, 42.9% (246/574) in men and 28.4% (23/81) in women (χ2 = 6.1 df = 1, p < 0.05). There were significant differences in age, sex, duration of illness, number of hospitalizations, drug type, smoking and PANSS negative symptom subscore between TD and non-TD groups (all p < 0.05). Moreover, patients with TD scored lower for immediate memory, attention, delayed memory, and RBANS total scores (all p < 0.05). Logistic regression showed a significant correlation between TD and age, sex, drug type and attention subscore.

CONCLUSION

Our results suggest that multiple demographic and clinical variables may be associated with the development of TD. Moreover, TD patients may exhibit more cognitive impairment than non-TD patients.

CT-Angiography-Based Outcome Prediction on Diabetic Foot Ulcer Patients: A Statistical Learning Approach

The purpose of our study is to predict the occurrence and prognosis of diabetic foot ulcers (DFUs) by clinical and lower extremity computed tomography angiography (CTA) data of patients using the artificial neural networks (ANN) model. DFU is a common complication of diabetes that severely affects the quality of life of patients, leading to amputation and even death. There are a lack of valid predictive techniques for the prognosis of DFU. In clinical practice, the use of scales alone has a large subjective component, leading to significant bias and heterogeneity. Currently, there is a lack of evidence-based support for patients to develop clinical strategies before reaching end-stage outcomes. The present study provides a novel technical tool for predicting the prognosis of DFU. After screening the data, 203 patients with diabetic foot ulcers (DFUs) were analyzed and divided into two subgroups based on their Wagner Score (138 patients in the low Wagner Score group and 65 patients in the high Wagner Score group). Based on clinical and lower extremity CTA data, 10 predictive factors were selected for inclusion in the model. The total dataset was randomly divided into the training sample, testing sample and holdout sample in ratio of 3:1:1. After the training sample and testing sample developing the ANN model, the holdout sample was utilized to assess the accuracy of the model. ANN model analysis shows that the sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and area under the curve (AUC) of the overall ANN model were 92.3%, 93.5%, 87.0%, 94.2% and 0.955, respectively. We observed that the proposed model performed superbly on the prediction of DFU with a 91.6% accuracy. Evaluated with the holdout sample, the model accuracy, sensitivity, specificity, PPV and NPV were 88.9%, 90.0%, 88.5%, 75.0% and 95.8%, respectively. By contrast, the logistic regression model was inferior to the ANN model. The ANN model can accurately and reliably predict the occurrence and prognosis of a DFU according to clinical and lower extremity CTA data. We provided clinicians with a novel technical tool to develop clinical strategies before end-stage outcomes.

The relationship between grey matter volume and striatal dopamine function in psychosis: a multimodal 18F-DOPA PET and voxel-based morphometry study

A leading hypothesis for schizophrenia and related psychotic disorders proposes that cortical brain disruption leads to subcortical dopaminergic dysfunction, which underlies psychosis in the majority of patients who respond to treatment. Although supported by preclinical findings that prefrontal cortical lesions lead to striatal dopamine dysregulation, the relationship between prefrontal structural volume and striatal dopamine function has not been tested in people with psychosis. We therefore investigated the in vivo relationship between striatal dopamine synthesis capacity and prefrontal grey matter volume in treatment-responsive patients with psychosis, and compared them to treatment non-responsive patients, where dopaminergic mechanisms are not thought to be central. Forty patients with psychosis across two independent cohorts underwent ¹⁸F-DOPA PET scans to measure dopamine synthesis capacity (indexed as the influx rate constant K_i^cer) and structural 3T MRI. The PET, but not MR, data have been reported previously. Structural images were processed using DARTEL-VBM. GLM analyses were performed in SPM12 to test the relationship between prefrontal grey matter volume and striatal K_i^cer. Treatment responders showed a negative correlation between prefrontal grey matter and striatal dopamine synthesis capacity, but this was not evident in treatment non-responders. Specifically, we found an interaction between treatment response, whole striatal dopamine synthesis capacity and grey matter volume in left (pFWE corr. = 0.017) and right (pFWE corr. = 0.042) prefrontal cortex. We replicated the finding in right prefrontal cortex in the independent sample (pFWE corr. = 0.031). The summary effect size was 0.82. Our findings are consistent with the long-standing hypothesis of dysregulation of the striatal dopaminergic system being related to prefrontal cortex pathology in schizophrenia, but critically also extend the hypothesis to indicate it can be applied to treatment-responsive schizophrenia only. This suggests that different mechanisms underlie the pathophysiology of treatment-responsive and treatment-resistant schizophrenia.

Pattern classification as decision support tool in antipsychotic treatment algorithms

Pattern classification aims to establish a new approach in personalized treatment. The scope is to tailor treatment on individual characteristics during all phases of care including prevention, diagnosis, treatment, and clinical outcome. In psychotic disorders, this need results from the fact that a third of patients with psychotic symptoms do not respond to antipsychotic treatment and are described as having treatment-resistant disorders. This, in addition to the high variability of treatment responses among patients, enhances the need of applying advanced classification algorithms to identify antipsychotic treatment patterns. This review comprehensively summarizes advancements and challenges of pattern classification in antipsychotic treatment response to date and aims to introduce clinicians and researchers to the challenges of including pattern classification into antipsychotic treatment decision algorithms.

Comparison of the predictive outcomes for anti-tuberculosis drug-induced hepatotoxicity by different machine learning techniques

BACKGROUND

The study compared the predictive outcomes of artificial neural network, support vector machine and random forest on the occurrence of anti-tuberculosis drug-induced hepatotoxicity.

METHODS

The clinical and genomic data of patients treated with anti-tuberculosis drugs at Taipei Medical University-Wanfang Hospital were used as training sets, and those at Taipei Medical University-Shuang Ho Hospital served as test sets. Features were selected through a univariate risk factor analysis and literature evaluation. The accuracy, sensitivity, specificity, and the area under the receiver operating characteristic curve were calculated to compare the traditional, genomic, and combined models of the three techniques.

RESULTS

Nine models were created with 7 clinical factors and 4 genotypes. Artificial neural network with clinical and genomic factors exhibited the best performance, with an accuracy of 88.67%, a sensitivity of 80%, and a specificity of 90.4% for the test set. The area under the receiver operating characteristic curve of this best model reached 0.894 for training set and 0.898 for test set, which was significantly better than 0.801 for training set and 0.728 for test set by support vector machine and 0.724 for training set and 0.718 for test set by random forest.

CONCLUSIONS

Artificial neural network with clinical and genomic data can become a clinical useful tool in predicting anti-tuberculosis drug-induced hepatotoxicity. The machine learning technique can be an innovation to predict and prevent adverse drug reaction.

Development of artificial neural network model for prediction of post-streptococcus mutans in dental caries

BACKGROUND AND OBJECTIVE

Streptococcus mutans is the primary initiator and most common organism associated with dental caries. Prediction of post-Streptococcus mutans favours in the selection of appropriate caries excavation method which eventually results in meliorate caries-free cavity preparation for restoration. The objective of this study is to predict the post-Streptococcus mutans prior to dental caries excavation based on pre- Streptococcus mutans using iOS App developed on Artificial Neural Network (ANN) model.

METHODS

For the current research work, children with occlusal dentinal caries lesion were chosen, 45 primary molar teeth cases were studied. Caries excavation was done with carbide bur, polymer bur and spoon excavator. The colony forming units for pre and post-Streptococcus mutans were recorded, data emanating from clinical trials was employed to develop the ANN models. ANN models were trained, validated and tested with the registered clinical data using different ANN architectures.

RESULTS

Feedforward backpropagation ANN model with an architecture of 4-5-1, predicts post-Streptococcus mutans with an efficiency of 0.99033, mean squared error and mean absolute percentage error for testing cases were 0.2341 and 4.967 respectively.

CONCLUSIONS

Caries excavation methods and pre-Streptococcus mutans are feed as inputs, while post-Streptococcus mutans as targets to develop ANN model. Based on the developed ANN model, an ingenious iOS App was developed, the global clinician may utilize the App to meticulously predict post-Streptococcus mutans on iPhone based on pre-Streptococcus mutans, which in turn aids in decision making for the selection of caries excavation method. This study manifests the potential application of iOS App with built-in ANN model in efficiently predicting the post-Streptococcus mutans. Also, the study extends scope for applications of iOS App with built-in ANN models in clinical medicine.

From cognitive and clinical substrates to functional profiles: Disentangling heterogeneity in schizophrenia

The relationship between neurocognition and functioning among patients with schizophrenia is well documented. However, integrating neuropsychological, clinical and psychopathological data to better investigate functional outcome still constitutes a challenge. Artificial neural network-based modeling might help to better capture clinical heterogeneity by analyzing the non-linear relationships among multiple variables. Two hundred and fourteen clinically stabilized patients with schizophrenia were recruited and assessed for neurocognition, psychopathology and functioning. Artificial neural network analyses were conducted to yield significant predictors of functional outcome among clinical and cognitive variables and to build distinct functional Profiles, each characterized by a different medley of cognitive and clinical features. Twenty-two key predictors of daily functioning emerged, encompassing neurocognitive and clinical domains, with major roles for processing speed and attention. Four Profiles were constructed based on specific levels of functioning, each characterized by a distinct distribution of key clinical and neurocognitve measures. This study highlights the importance of a more in-depth investigation of cognitive and clinical heterogeneity. A better understanding of the building blocks of these Profiles would lead to more individualized rehabilitation treatments.

Integrating Space, Time, and Orientation in Spiking Neural Networks: A Case Study on Multimodal Brain Data Modeling

Recent progress in a noninvasive brain data sampling technology has facilitated simultaneous sampling of multiple modalities of brain data, such as functional magnetic resonance imaging, electroencephalography, diffusion tensor imaging, and so on. In spite of the potential benefits from integrating predictive modeling of multiple modality brain data, this area of research remains mostly unexplored due to a lack of methodological advancements. The difficulty in fusing multiple modalities of brain data within a single model lies in the heterogeneous temporal and spatial characteristics of the data sources. Recent advances in spiking neural network systems, however, provide the flexibility to incorporate multidimensional information within the model. This paper proposes a novel, unsupervised learning algorithm for fusing temporal, spatial, and orientation information in a spiking neural network architecture that could potentially be used to understand and perform predictive modeling using multimodal data. The proposed algorithm is evaluated both qualitatively and quantitatively using synthetically generated data to characterize its behavior and its ability to utilize spatial, temporal, and orientation information within the model. This leads to improved pattern recognition capabilities and performance along with robust interpretability of the brain data. Furthermore, a case study is presented, which aims to build a computational model that discriminates between people with schizophrenia who respond or do not respond to monotherapy with the antipsychotic clozapine.

Biological Predictors of Clozapine Response: A Systematic Review

Background: Clozapine is the recommended antipsychotic for treatment-resistant schizophrenia (TRS) but there is significant variability between patients in the degree to which clozapine will improve symptoms. The biological basis of this variability is unknown. Although clozapine has efficacy in TRS, it can elicit adverse effects and initiation is often delayed. Identification of predictive biomarkers of clozapine response may aid initiation of clozapine treatment, as well as understanding of its mechanism of action. In this article we systematically review prospective or genetic studies of biological predictors of response to clozapine. Methods: We searched the PubMed database until 20th January 2018 for studies investigating "clozapine" AND ("response" OR "outcome") AND "schizophrenia." Inclusion required that studies examined a biological variable in relation to symptomatic response to clozapine. For all studies except genetic-studies, inclusion required that biological variables were measured before clozapine initiation. Results: Ninety-eight studies met the eligibility criteria and were included in the review, including neuroimaging, blood-based, cerebrospinal fluid (CSF)-based, and genetic predictors. The majority (70) are genetic studies, collectively investigating 379 different gene variants, however only three genetic variants (DRD3 Ser9Gly, HTR2A His452Tyr, and C825T GNB3) have independently replicated significant findings. Of the non-genetic variables, the most consistent predictors of a good response to clozapine are higher prefrontal cortical structural integrity and activity, and a lower ratio of the dopamine and serotonin metabolites, homovanillic acid (HVA): 5-hydroxyindoleacetic acid (5-HIAA) in CSF. Conclusions: Recommendations include that future studies should ensure adequate clozapine trial length and clozapine plasma concentrations, and may include multivariate models to increase predictive accuracy.

The Group of Treatment Resistant Schizophrenias. Heterogeneity in Treatment Resistant Schizophrenia (TRS)

Schizophrenia is composed of a heterogeneous group of patient segments. Our current notion of the heterogeneity in schizophrenia is based on patients presenting with diverse disease symptom phenotypes, risk factors, structural and functional neuropathology, and a mixed range of expressed response to treatment. It is important for clinicians to recognize the various clinical presentations of resistance to treatment in schizophrenia and to understand how heterogeneity across treatment resistant patient segments may potentially inform new strategies for the development of effective treatments for Treatment Resistant Schizophrenia (TRS). The heterogeneity of schizophrenia may be reduced by parsing patient segments based on whether patients demonstrate an adequate or inadequate response to treatment. In our current concept of TRS, TRS is defined as non-response to at least two adequate trials of antipsychotic medication and is estimated to affect about 30% of all patients with schizophrenia. In this narrative review, the author discusses that the demonstration of inadequate response to antipsychotic drugs (APDs) may infer that some TRS patients may be suffering from a non-dopamine pathophysiology since D2 receptor antagonist-based treatment is ineffective. Preliminary neurobiological findings may further support the pathophysiologic distinction of TRS from that of general schizophrenia. Investigation of the basis for heterogeneity in TRS through the systematic investigation of relevant "clusters" of similarly at risk individuals may hopefully bring us closer to realize a precision medicine approach for developing effective therapies for TRS patient segments.

Two distinct patterns of treatment resistance: clinical predictors of treatment resistance in first-episode schizophrenia spectrum psychoses

BACKGROUND

Clozapine remains the only evidence-based antipsychotic for treatment-resistant schizophrenia (TRS). The ability to predict which patients with their first onset of schizophrenia would subsequently meet criteria for treatment resistance (TR) could help to diminish the severe functional disability which may ensue if TR is not recognized and correctly treated.

METHOD

This is a 5-year longitudinal assessment of clinical outcomes in a cohort of 246 first-episode schizophrenia spectrum patients recruited as part of the NIHR Genetics and Psychosis (GAP) study conducted in South London from 2005 to 2010. We examined the relationship between baseline demographic and clinical measures and the emergence of TR. TR status was determined from a review of electronic case records. We assessed for associations with early-, and late-onset TR, and non-TR, and differences between those TR patients treated with clozapine and those who were not.

RESULTS

Seventy per cent (n = 56) of TR patients, and 23% of the total study population (n = 246) were treatment resistant from illness onset. Those who met criteria for TR during the first 5 years of illness were more likely to have an early age of first contact for psychosis (<20 years) [odds ratio (OR) 2.49, 95% confidence interval (CI) 1.25-4.94] compared to those with non-TR. The relationship between an early age of first contact (<20 years) and TR was significant in patients of Black ethnicity (OR 3.71, 95% CI 1.44-9.56); and patients of male gender (OR 3.13 95% CI 1.35-7.23).

CONCLUSIONS

For the majority of the TR group, antipsychotic TR is present from illness onset, necessitating increased consideration for the earlier use of clozapine.

Applicability of gene expression and systems biology to develop pharmacogenetic predictors; antipsychotic-induced extrapyramidal symptoms as an example

Pharmacogenetics has been driven by a candidate gene approach. The disadvantage of this approach is that is limited by our current understanding of the mechanisms by which drugs act. Gene expression could help to elucidate the molecular signatures of antipsychotic treatments searching for dysregulated molecular pathways and the relationships between gene products, especially protein-protein interactions. To embrace the complexity of drug response, machine learning methods could help to identify gene-gene interactions and develop pharmacogenetic predictors of drug response. The present review summarizes the applicability of the topics presented here (gene expression, network analysis and gene-gene interactions) in pharmacogenetics. In order to achieve this, we present an example of identifying genetic predictors of extrapyramidal symptoms induced by antipsychotic.

Clearance rate and BP-ANN model in paraquat poisoned patients treated with hemoperfusion

In order to investigate the effect of hemoperfusion (HP) on the clearance rate of paraquat (PQ) and develop a clearance model, 41 PQ-poisoned patients who acquired acute PQ intoxication received HP treatment. PQ concentrations were determined by high performance liquid chromatography (HPLC). According to initial PQ concentration, study subjects were divided into two groups: Low-PQ group (0.05–1.0 μg/mL) and High-PQ group (1.0–10 μg/mL). After initial HP treatment, PQ concentrations decreased in both groups. However, in the High-PQ group, PQ levels remained in excess of 0.05 μg/mL and increased when the second HP treatment was initiated. Based on the PQ concentrations before and after HP treatment, the mean clearance rate of PQ calculated was 73 ± 15%. We also established a backpropagation artificial neural network (BP-ANN) model, which set PQ concentrations before HP treatment as input data and after HP treatment as output data. When it is used to predict PQ concentration after HP treatment, high prediction accuracy (R = 0.9977) can be obtained in this model. In conclusion, HP is an effective way to clear PQ from the blood, and the PQ concentration after HP treatment can be predicted by BP-ANN model.

Pharmacogenetic predictor of extrapyramidal symptoms induced by antipsychotics: multilocus interaction in the mTOR pathway

Antipsychotic (AP) treatment-emergent extrapyramidal symptoms (EPS) are acute adverse reactions of APs. The aim of the present study is to analyze gene-gene interactions in nine genes related to the mTOR pathway, in order to develop genetic predictors of the appearance of EPS. 243 subjects (78 presenting EPS: 165 not) from three cohorts participated in the present study: Cohort 1, patients treated with risperidone, (n=114); Cohort 2, patients treated with APs other than risperidone (n=102); Cohort 3, AP-naïve patients with first-episode psychosis treated with risperidone, paliperidone or amisulpride, n=27. We analyzed gene-gene interactions by multifactor dimensionality reduction assay (MDR). In Cohort 1, we identified a four-way interaction, including the rs1130214 (AKT1), rs456998 (FCHSD1), rs7211818 (Raptor) and rs1053639 (DDIT4), that correctly predicted 97 of the 114 patients (85% accuracy). We validated the predictive power of the four-way interaction in Cohort 2 and in Cohort 3 with 86% and 88% accuracy respectively. We develop and validate a powerful pharmacogenetic predictor of AP-induced EPS. For the first time, the mTOR pathway has been related to EPS susceptibility and AP response. However, validation in larger and independent populations will be necessary for optimal generalization.

Opening Pandora’s box in the UK: a hypothetical pharmacogenetic test for clozapine

Clozapine is a uniquely efficacious antipsychotic drug in treatment-resistant schizophrenia. Its use is restricted due to adverse effects including a rare but dangerous reduction in neutrophils (agranulocytosis) and the mandatory hematological monitoring this entails in many countries. We review the statistical, ethical and legal issues arising from a hypothetical pharmacogenetic test for clozapine, using the UK as an exemplary case for consideration. Our key findings include: a consideration of the probabilistic results that a pharmacogenetic test may return; the impact on drug licensing; and the potential for pharmacogenetic tests for clozapine being used without consent under the UK’s legal framework. We make recommendations regarding regulatory changes applicable to the special case of pharmacogenetic testing in clozapine treatment.

Predicting hospital-acquired infections by scoring system with simple parameters

Background

Hospital-acquired infections (HAI) are associated with increased attributable morbidity, mortality, prolonged hospitalization, and economic costs. A simple, reliable prediction model for HAI has great clinical relevance. The objective of this study is to develop a scoring system to predict HAI that was derived from Logistic Regression (LR) and validated by Artificial Neural Networks (ANN) simultaneously.

Methodology/Principal Findings

A total of 476 patients from all the 806 HAI inpatients were included for the study between 2004 and 2005. A sample of 1,376 non-HAI inpatients was randomly drawn from all the admitted patients in the same period of time as the control group. External validation of 2,500 patients was abstracted from another academic teaching center. Sixteen variables were extracted from the Electronic Health Records (EHR) and fed into ANN and LR models. With stepwise selection, the following seven variables were identified by LR models as statistically significant: Foley catheterization, central venous catheterization, arterial line, nasogastric tube, hemodialysis, stress ulcer prophylaxes and systemic glucocorticosteroids. Both ANN and LR models displayed excellent discrimination (area under the receiver operating characteristic curve [AUC]: 0.964 versus 0.969, p = 0.507) to identify infection in internal validation. During external validation, high AUC was obtained from both models (AUC: 0.850 versus 0.870, p = 0.447). The scoring system also performed extremely well in the internal (AUC: 0.965) and external (AUC: 0.871) validations.

Conclusions

We developed a scoring system to predict HAI with simple parameters validated with ANN and LR models. Armed with this scoring system, infectious disease specialists can more efficiently identify patients at high risk for HAI during hospitalization. Further, using parameters either by observation of medical devices used or data obtained from EHR also provided good prediction outcome that can be utilized in different clinical settings.

Systematic quantitative characterization of cellular responses induced by multiple signals

Background

Cells constantly sense many internal and environmental signals and respond through their complex signaling network, leading to particular biological outcomes. However, a systematic characterization and optimization of multi-signal responses remains a pressing challenge to traditional experimental approaches due to the arising complexity associated with the increasing number of signals and their intensities.

Results

We established and validated a data-driven mathematical approach to systematically characterize signal-response relationships. Our results demonstrate how mathematical learning algorithms can enable systematic characterization of multi-signal induced biological activities. The proposed approach enables identification of input combinations that can result in desired biological responses. In retrospect, the results show that, unlike a single drug, a properly chosen combination of drugs can lead to a significant difference in the responses of different cell types, increasing the differential targeting of certain combinations. The successful validation of identified combinations demonstrates the power of this approach. Moreover, the approach enables examining the efficacy of all lower order mixtures of the tested signals. The approach also enables identification of system-level signaling interactions between the applied signals. Many of the signaling interactions identified were consistent with the literature, and other unknown interactions emerged.

Conclusions

This approach can facilitate development of systems biology and optimal drug combination therapies for cancer and other diseases and for understanding key interactions within the cellular network upon treatment with multiple signals.

Do we have any solid evidence of clinical utility about the pathophysiology of schizophrenia?

A diagnosis of schizophrenia, as in most of psychiatric practice, is made largely by eliciting symptoms with reference to subjective, albeit operationalized, criteria. This diagnosis then provides some rationale for management. Objective diagnostic and therapeutic tests are much more desirable, provided they are reliably measured and interpreted. Definite advances have been made in our understanding of schizophrenia in recent decades, but there has been little consideration of how this information could be used in clinical practice. We review here the potential utility of the strongest and best replicated risk factors for and manifestations of schizophrenia within clinical, epidemiological, cognitive, blood biomarker and neuroimaging domains. We place particular emphasis on the sensitivity, specificity and predictive power of pathophysiological indices for making a diagnosis, establishing an early diagnosis or predicting treatment response in schizophrenia. We conclude that a number of measures currently available have the potential to increase the rigour of clinical assessments in schizophrenia. We propose that the time has come to more fully evaluate these and other well replicated abnormalities as objective potential diagnostic and prognostic guides, and to steer future clinical, therapeutic and nosological research in this direction.

Where to position clozapine: re-examining the evidence

OBJECTIVE

To review clozapine's position in treatment algorithms for schizophrenia.

METHOD

Clozapine's status is reviewed in the context of its initial discovery and unique clinical and (or) pharmacological profile, withdrawal and link with hematologic concerns, reintroduction with monitoring guidelines, prototype for atypicality, positioning in treatment algorithms, and current evidence regarding efficacy, effectiveness, and side effects.

RESULTS

The hematologic monitoring implemented with clozapine's reintroduction here in North America has proven successful in preventing clozapine-related deaths secondary to agranulocytosis. While its other side effects are not without concern, present evidence does not link clozapine to increased mortality rates; indeed, it appears better than other antipsychotics in this regard. Moreover, its clinical superiority compared with all other antipsychotics has been confirmed both in efficacy and in effectiveness trials.

CONCLUSIONS

Schizophrenia continues to represent a treatment challenge, with many people demonstrating suboptimal response and poor functional outcome. Clozapine is routinely positioned as a third-line treatment in schizophrenia, but in light of existing evidence this warrants re-examination.

Application of an artificial neural network to predict postinduction hypotension during general anesthesia

BACKGROUND

Perioperative hypotension is associated with adverse outcomes in patients undergoing surgery. A computer-based model that integrates related factors and predicts the risk of hypotension would be helpful in clinical anesthesia. The purpose of this study was to develop artificial neural network (ANN) models to identify patients at high risk for postinduction hypotension during general anesthesia.

METHODS

Anesthesia records for March through November 2007 were reviewed, and 1017 records were analyzed. Eleven patient-related, 2 surgical, and 5 anesthetic variables were used to develop the ANN and logistic regression (LR) models. The quality of the models was evaluated by an external validation data set. Three clinicians were asked to make predictions of the same validation data set on a case-by-case basis.

RESULTS

The ANN model had an accuracy of 82.3%, sensitivity of 76.4%, and specificity of 85.6%. The accuracy of the LR model was 76.5%, the sensitivity was 74.5%, and specificity was 77.7%. The area under the receiver operating characteristic curve for the ANN and LR models was 0.893 and 0.840. The clinicians had the lowest predictive accuracy and sensitivity compared with the ANN and LR models.

CONCLUSIONS

The ANN model developed in this study had good discrimination and calibration and would provide decision support to clinicians and increase vigilance for patients at high risk of postinduction hypotension during general anesthesia.

Novel solutions for an old disease: diagnosis of acute appendicitis with random forest, support vector machines, and artificial neural networks

BACKGROUND

Diagnosing acute appendicitis clinically is still difficult. We developed random forests, support vector machines, and artificial neural network models to diagnose acute appendicitis.

METHODS

Between January 2006 and December 2008, patients who had a consultation session with surgeons for suspected acute appendicitis were enrolled. Seventy-five percent of the data set was used to construct models including random forest, support vector machines, artificial neural networks, and logistic regression. Twenty-five percent of the data set was withheld to evaluate model performance. The area under the receiver operating characteristic curve (AUC) was used to evaluate performance, which was compared with that of the Alvarado score.

RESULTS

Data from a total of 180 patients were collected, 135 used for training and 45 for testing. The mean age of patients was 39.4 years (range, 16-85). Final diagnosis revealed 115 patients with and 65 without appendicitis. The AUC of random forest, support vector machines, artificial neural networks, logistic regression, and Alvarado was 0.98, 0.96, 0.91, 0.87, and 0.77, respectively. The sensitivity, specificity, positive, and negative predictive values of random forest were 94%, 100%, 100%, and 87%, respectively. Random forest performed better than artificial neural networks, logistic regression, and Alvarado.

CONCLUSION

We demonstrated that random forest can predict acute appendicitis with good accuracy and, deployed appropriately, can be an effective tool in clinical decision making.

Predicting hypotensive episodes during spinal anesthesia with the application of artificial neural networks

Hypotension is one of the most frequent adverse effects of spinal anesthesia. Several factors might be related to the occurrence of hypotension. Predictions of the hypotensive event, however, had been addressed by only a few authors using logistic regression (LR) models. Artificial neural networks (ANN) are pattern-recognition tools that can be used to detect complex patterns within data sets. The purpose of this study was to develop the ANN-based predictive model to identify patients with high risk of hypotension during spinal anesthesia. From September 2004 to December 2006, the anesthesia records of 1501 patients receiving surgery under spinal anesthesia were used to develop the ANN and LR models. By random selection 75% of data were used for training and the remaining 25% of data were used as test set for validating the predictive performance. Five senior anesthesiologists were asked to review the data of test set and to make predictions of hypotensive event during spinal anesthesia by clinical experience. The ANN model had a sensitivity of 75.9% and specificity of 76.0%. The LR model had a sensitivity of 68.1% and specificity of 73.5%. The area under receiver operating characteristic curves were 0.796 and 0.748. The ANN model performed significantly better than the LR model. The prediction of clinicians had the lowest sensitivity of 28.7%, 22.2%, 21.3%, 16.1%, and 36.1%, and specificity of 76.8%, 84.3%, 83.1%, 87.0%, and 64.0%. The computer-based predictive model should be useful in increasing vigilance in those patients most at risk for hypotension during spinal anesthesia, in allowing for patient-specific therapeutic intervention, or even in suggesting the use of alternative methods of anesthesia.