Dataset Information

Individual prediction of long-term outcome in adolescents at ultra-high risk for psychosis: Applying machine learning techniques to brain imaging data.

ABSTRACT: An important focus of studies of individuals at ultra-high risk (UHR) for psychosis has been to identify biomarkers to predict which individuals will transition to psychosis. However, the majority of individuals will prove to be resilient and go on to experience remission of their symptoms and function well. The aim of this study was to investigate the possibility of using structural MRI measures collected in UHR adolescents at baseline to quantitatively predict their long-term clinical outcome and level of functioning. We included 64 UHR individuals and 62 typically developing adolescents (12-18 years old at recruitment). At six-year follow-up, we determined resilience for 43 UHR individuals. Support Vector Regression analyses were performed to predict long-term functional and clinical outcome from baseline MRI measures on a continuous scale, instead of the more typical binary classification. This led to predictive correlations of baseline MR measures with level of functioning, and negative and disorganization symptoms. The highest correlation (r?=?0.42) was found between baseline subcortical volumes and long-term level of functioning. In conclusion, our results show that structural MRI data can be used to quantitatively predict long-term functional and clinical outcome in UHR individuals with medium effect size, suggesting that there may be scope for predicting outcome at the individual level. Moreover, we recommend classifying individual outcome on a continuous scale, enabling the assessment of different functional and clinical scales separately without the need to set a threshold. Hum Brain Mapp 38:704-714, 2017. © 2016 Wiley Periodicals, Inc.

SUBMITTER: de Wit S

PROVIDER: S-EPMC6866746 | biostudies-literature | 2017 Feb

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Individual prediction of long-term outcome in adolescents at ultra-high risk for psychosis: Applying machine learning techniques to brain imaging data.

de Wit Sanne S Ziermans Tim B TB Nieuwenhuis M M Schothorst Patricia F PF van Engeland Herman H Kahn René S RS Durston Sarah S Schnack Hugo G HG

Human brain mapping 20161004 2

An important focus of studies of individuals at ultra-high risk (UHR) for psychosis has been to identify biomarkers to predict which individuals will transition to psychosis. However, the majority of individuals will prove to be resilient and go on to experience remission of their symptoms and function well. The aim of this study was to investigate the possibility of using structural MRI measures collected in UHR adolescents at baseline to quantitatively predict their long-term clinical outcome ...[more]

PMID: 27699911

Similar Datasets

Project description:ObjectiveMachine learning techniques have demonstrated superior discrimination compared to conventional statistical approaches in predicting trauma death. The objective of this study is to evaluate whether machine learning algorithms can be used to assess risk and dynamically identify patient-specific modifiable factors critical to patient trajectory for multiple key outcomes after severe injury.MethodsSuperLearner, an ensemble machine-learning algorithm, was applied to prospective observational cohort data from 1494 critically-injured patients. Over 1000 agnostic predictors were used to generate prediction models from multiple candidate learners for outcomes of interest at serial time points post-injury. Model accuracy was estimated using cross-validation and area under the curve was compared to select among predictors. Clinical variables responsible for driving outcomes were estimated at each time point.ResultsSuperLearner fits demonstrated excellent cross-validated prediction of death (overall AUC 0.94-0.97), multi-organ failure (overall AUC 0.84-0.90), and transfusion (overall AUC 0.87-0.9) across multiple post-injury time points, and good prediction of Acute Respiratory Distress Syndrome (overall AUC 0.84-0.89) and venous thromboembolism (overall AUC 0.73-0.83). Outcomes with inferior data quality included coagulopathic trajectory (AUC 0.48-0.88). Key clinical predictors evolved over the post-injury timecourse and included both anticipated and unexpected variables. Non-random missingness of data was identified as a predictor of multiple outcomes over time.ConclusionsMachine learning algorithms can be used to generate dynamic prediction after injury while avoiding the risk of over- and under-fitting inherent in ad hoc statistical approaches. SuperLearner prediction after injury demonstrates promise as an adaptable means of helping clinicians integrate voluminous, evolving data on severely-injured patients into real-time, dynamic decision-making support.

Project description:BackgroundExocrine pancreatic insufficiency (EPI) is a serious condition characterized by a lack of functional exocrine pancreatic enzymes and the resultant inability to properly digest nutrients. EPI can be caused by a variety of disorders, including chronic pancreatitis, pancreatic cancer, and celiac disease. EPI remains underdiagnosed because of the nonspecific nature of clinical symptoms, lack of an ideal diagnostic test, and the inability to easily identify affected patients using administrative claims data.ObjectivesTo develop a machine learning model that identifies patients in a commercial medical claims database who likely have EPI but are undiagnosed.MethodsA machine learning algorithm was developed in Scikit-learn, a Python module. The study population, selected from the 2014 Truven MarketScan® Commercial Claims Database, consisted of patients with EPI-prone conditions. Patients were labeled with 290 condition category flags and split into actual positive EPI cases, actual negative EPI cases, and unlabeled cases. The study population was then randomly divided into a training subset and a testing subset. The training subset was used to determine the performance metrics of 27 models and to select the highest performing model, and the testing subset was used to evaluate performance of the best machine learning model.ResultsThe study population consisted of 2088 actual positive EPI cases, 1077 actual negative EPI cases, and 437 530 unlabeled cases. In the best performing model, the precision, recall, and accuracy were 0.91, 0.80, and 0.86, respectively. The best-performing model estimated that the number of patients likely to have EPI was about 12 times the number of patients directly identified as EPI-positive through a claims analysis in the study population. The most important features in assigning EPI probability were the presence or absence of diagnosis codes related to pancreatic and digestive conditions.ConclusionsMachine learning techniques demonstrated high predictive power in identifying patients with EPI and could facilitate an enhanced understanding of its etiology and help to identify patients for possible diagnosis and treatment.

Project description:ImportanceIn coordinated specialty care (CSC) settings for people with a first episode of psychosis, the development of reliable, validated individual-level prediction tools for key outcomes may be informative for shared clinician and client decision-making.ObjectiveTo develop an individual-level prediction tool using machine-learning methods that predicts a trajectory of education/work status or psychiatric hospitalization outcomes over a client's next year of quarterly follow-up assessments. Additionally, to visualize these predictions in a way that is informative to clinicians and clients.Design, setting, and participantsIndividual-level data were collected for all patients enrolled in the OnTrackNY program at enrollment and at quarterly follow-ups using standardized forms. The OnTrackNY program, a network of CSC sites in New York State, provides person-centered, recovery-oriented, and evidence-based psychosocial and pharmaceutical interventions to individuals aged 16 to 30 years with recent-onset (<2 years) nonaffective psychosis. Although data collection is ongoing, data for this study were collected from October 2013 to December 2018, and the time frame for analysis was July 2020 to May 2021. Data were separated into a training/cross-validation set to perform internally validated model development and a separate holdout test set (~20% of the sample) for external validation. Random probability forest models were developed to predict individual-level trajectories of outcomes.ExposuresForty-three individual-level demographic and clinical features collected at enrollment in OnTrackNY, 25 of which were time-varying and updated at quarterly follow-up assessments, and 13 site-level demographic and economic census variables.Main outcomes and measuresIndividual-level education and/or employment status and psychiatric hospitalization trajectories at quarterly follow-up periods across the first 2 years of CSC.ResultsThe total study sample consists of 1298 individuals aged 16 to 30 years and included 341 women (26.3%), 949 men (73.1%), and 8 (<1%) with another gender. Prediction models performed well for 1-year trajectories of education/work across all validation sets, with areas under the receiver operating characteristic curve (AUCs) ranging from 0.68 (95% CI, 0.63-0.74) to 0.88 (95% CI, 0.81-0.96). Predictive accuracy for psychiatric hospitalization 3 months ahead reached AUC above 0.70; moreover, predictions of future psychiatric hospitalizations at 6 months and beyond were consistently poor, with AUCs below 0.60. Given the good externally validated performance for predicting education/work, a prototype interactive visualization tool displaying individual-level education/work trajectories and related features was developed.Conclusions and relevanceThis study suggests that accurate prediction tools can be developed for outcomes in people with first-episode psychosis, which may help inform shared clinician/client decision-making. Future work should study the effectiveness of its deployment, including proper communication to inform shared clinician/client decision-making in the context of a learning health care system. At present, more work is needed to develop better performing prediction models for future psychiatric hospitalizations before any tool is recommended for this outcome.

Project description:Given our aging society and the prevalence of age-related hearing loss that often develops during adulthood, hearing loss is a common public health issue affecting almost all older adults. Moderate-to-moderately severe hearing loss can usually be corrected with hearing aids; however, severe-to-profound hearing loss often requires a cochlear implant (CI). However, post-operative CI results vary, and the performance of the previous prediction models is limited, indicating that a new approach is needed. For postlingually deaf adults (n de120) who received CI with full insertion, we predicted CI outcomes using a Random-Forest Regression (RFR) model and investigated the effect of preoperative factors on CI outcomes. Postoperative word recognition scores (WRS) served as the dependent variable to predict. Predictors included duration of deafness (DoD), age at CI operation (ageCI), duration of hearing-aid use (DoHA), preoperative hearing threshold and sentence recognition score. Prediction accuracy was evaluated using mean absolute error (MAE) and Pearson's correlation coefficient r between the true WRS and predicted WRS. The fitting using a linear model resulted in prediction of WRS with r = 0.7 and MAE = 15.6 ± 9. RFR outperformed the linear model (r = 0.96, MAE = 6.1 ± 4.7, p < 0.00001). Cross-hospital data validation showed reliable performance using RFR (r = 0.91, MAE = 9.6 ± 5.2). The contribution of DoD to prediction was the highest (MAE increase when omitted: 14.8), followed by ageCI (8.9) and DoHA (7.5). After CI, patients with DoD < 10 years presented better WRSs and smaller variations (p < 0.01) than those with longer DoD. Better WRS was also explained by younger age at CI and longer-term DoHA. Machine learning demonstrated a robust prediction performance for CI outcomes in postlingually deaf adults across different institutes, providing a reference value for counseling patients considering CI. Health care providers should be aware that the patients with severe-to-profound hearing loss who cannot have benefit from hearing aids need to proceed with CI as soon as possible and should continue using hearing aids until after CI operation.

Dataset Information

Individual prediction of long-term outcome in adolescents at ultra-high risk for psychosis: Applying machine learning techniques to brain imaging data.

Publications

Individual prediction of long-term outcome in adolescents at ultra-high risk for psychosis: Applying machine learning techniques to brain imaging data.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets