scholarly journals Developing an Explainable Machine Learning-Based Personalised Dementia Risk Prediction Model: A Transfer Learning Approach With Ensemble Learning Algorithms

2021 ◽  
Vol 4 ◽  
Author(s):  
Samuel O. Danso ◽  
Zhanhang Zeng ◽  
Graciela Muniz-Terrera ◽  
Craig W. Ritchie
Keyword(s):  
Machine Learning ◽  
Risk Factors ◽  
Transfer Learning ◽  
Risk Prediction ◽  
Clinical Utility ◽  
Prediction Models ◽  
Source Model ◽  
Geometric Accuracy ◽  
Risk Prediction Models ◽  
Dementia Risk

Alzheimer's disease (AD) has its onset many decades before dementia develops, and work is ongoing to characterise individuals at risk of decline on the basis of early detection through biomarker and cognitive testing as well as the presence/absence of identified risk factors. Risk prediction models for AD based on various computational approaches, including machine learning, are being developed with promising results. However, these approaches have been criticised as they are unable to generalise due to over-reliance on one data source, poor internal and external validations, and lack of understanding of prediction models, thereby limiting the clinical utility of these prediction models. We propose a framework that employs a transfer-learning paradigm with ensemble learning algorithms to develop explainable personalised risk prediction models for dementia. Our prediction models, known as source models, are initially trained and tested using a publicly available dataset (n = 84,856, mean age = 69 years) with 14 years of follow-up samples to predict the individual risk of developing dementia. The decision boundaries of the best source model are further updated by using an alternative dataset from a different and much younger population (n = 473, mean age = 52 years) to obtain an additional prediction model known as the target model. We further apply the SHapely Additive exPlanation (SHAP) algorithm to visualise the risk factors responsible for the prediction at both population and individual levels. The best source model achieves a geometric accuracy of 87%, specificity of 99%, and sensitivity of 76%. In comparison to a baseline model, our target model achieves better performance across several performance metrics, within an increase in geometric accuracy of 16.9%, specificity of 2.7%, and sensitivity of 19.1%, an area under the receiver operating curve (AUROC) of 11% and a transfer learning efficacy rate of 20.6%. The strength of our approach is the large sample size used in training the source model, transferring and applying the “knowledge” to another dataset from a different and undiagnosed population for the early detection and prediction of dementia risk, and the ability to visualise the interaction of the risk factors that drive the prediction. This approach has direct clinical utility.

scholarly journals Assessing the Predictive Validity of Simple Dementia Risk Models in Harmonized Stroke Cohorts

Stroke ◽  
2020 ◽  
Vol 51 (7) ◽  
pp. 2095-2102
Author(s):  
Eugene Y.H. Tang ◽  
Christopher I. Price ◽  
Louise Robinson ◽  
Catherine Exley ◽  
David W. Desmond ◽  
...  
Keyword(s):  
Risk Factors ◽  
General Population ◽  
Risk Prediction ◽  
Prediction Models ◽  
Disease Risk ◽  
Risk Index ◽  
Area Under The Curve ◽  
Risk Prediction Models ◽  
C Statistic ◽  
Dementia Risk

Background and Purpose: Stroke is associated with an increased risk of dementia. To assist in the early identification of individuals at high risk of future dementia, numerous prediction models have been developed for use in the general population. However, it is not known whether such models also provide accurate predictions among stroke patients. Therefore, the aim of this study was to determine whether existing dementia risk prediction models that were developed for use in the general population can also be applied to individuals with a history of stroke to predict poststroke dementia with equivalent predictive validity. Methods: Data were harmonized from 4 stroke studies (follow-up range, ≈12–18 months poststroke) from Hong Kong, the United States, the Netherlands, and France. Regression analysis was used to test 3 risk prediction models: the Cardiovascular Risk Factors, Aging and Dementia score, the Australian National University Alzheimer Disease Risk Index, and the Brief Dementia Screening Indicator. Model performance or discrimination accuracy was assessed using the C statistic or area under the curve. Calibration was tested using the Grønnesby and Borgan and the goodness-of-fit tests. Results: The predictive accuracy of the models varied but was generally low compared with the original development cohorts, with the Australian National University Alzheimer Disease Risk Index (C-statistic, 0.66) and the Brief Dementia Screening Indicator (C-statistic, 0.61) both performing better than the Cardiovascular Risk Factors, Aging and Dementia score (area under the curve, 0.53). Conclusions: Dementia risk prediction models developed for the general population do not perform well in individuals with stroke. Their poor performance could have been due to the need for additional or different predictors related to stroke and vascular risk factors or methodological differences across studies (eg, length of follow-up, age distribution). Future work is needed to develop simple and cost-effective risk prediction models specific to poststroke dementia.

Performance Metrics for the Comparative Analysis of Clinical Risk Prediction Models Employing Machine Learning

2021 ◽  
Author(s):  
Chenxi Huang ◽  
Shu-Xia Li ◽  
César Caraballo ◽  
Frederick A. Masoudi ◽  
John S. Rumsfeld ◽  
...  
Keyword(s):  
Machine Learning ◽  
Risk Prediction ◽  
Health Care Professionals ◽  
Clinical Decision Making ◽  
Performance Metrics ◽  
Prediction Models ◽  
Learning Models ◽  
Risk Prediction Models ◽  
Clinical Risk ◽  
Machine Learning Models

Background: New methods such as machine learning techniques have been increasingly used to enhance the performance of risk predictions for clinical decision-making. However, commonly reported performance metrics may not be sufficient to capture the advantages of these newly proposed models for their adoption by health care professionals to improve care. Machine learning models often improve risk estimation for certain subpopulations that may be missed by these metrics. Methods and Results: This article addresses the limitations of commonly reported metrics for performance comparison and proposes additional metrics. Our discussions cover metrics related to overall performance, discrimination, calibration, resolution, reclassification, and model implementation. Models for predicting acute kidney injury after percutaneous coronary intervention are used to illustrate the use of these metrics. Conclusions: We demonstrate that commonly reported metrics may not have sufficient sensitivity to identify improvement of machine learning models and propose the use of a comprehensive list of performance metrics for reporting and comparing clinical risk prediction models.

scholarly journals Polygenic risk prediction models for colorectal cancer: a systematic review

BMC Cancer ◽  
2022 ◽  
Vol 22 (1) ◽  
Author(s):  
Michele Sassano ◽  
Marco Mariani ◽  
Gianluigi Quaranta ◽  
Roberta Pastorino ◽  
Stefania Boccia
Keyword(s):  
Colorectal Cancer ◽  
Risk Factors ◽  
Systematic Review ◽  
Prediction Model ◽  
Risk Prediction ◽  
Genetic Variants ◽  
Prediction Models ◽  
Risk Prediction Models ◽  
Discriminatory Accuracy ◽  
Traditional Risk Factors

Abstract Background Risk prediction models incorporating single nucleotide polymorphisms (SNPs) could lead to individualized prevention of colorectal cancer (CRC). However, the added value of incorporating SNPs into models with only traditional risk factors is still not clear. Hence, our primary aim was to summarize literature on risk prediction models including genetic variants for CRC, while our secondary aim was to evaluate the improvement of discriminatory accuracy when adding SNPs to a prediction model with only traditional risk factors. Methods We conducted a systematic review on prediction models incorporating multiple SNPs for CRC risk prediction. We tested whether a significant trend in the increase of Area Under Curve (AUC) according to the number of SNPs could be observed, and estimated the correlation between AUC improvement and number of SNPs. We estimated pooled AUC improvement for SNP-enhanced models compared with non-SNP-enhanced models using random effects meta-analysis, and conducted meta-regression to investigate the association of specific factors with AUC improvement. Results We included 33 studies, 78.79% using genetic risk scores to combine genetic data. We found no significant trend in AUC improvement according to the number of SNPs (p for trend = 0.774), and no correlation between the number of SNPs and AUC improvement (p = 0.695). Pooled AUC improvement was 0.040 (95% CI: 0.035, 0.045), and the number of cases in the study and the AUC of the starting model were inversely associated with AUC improvement obtained when adding SNPs to a prediction model. In addition, models constructed in Asian individuals achieved better AUC improvement with the incorporation of SNPs compared with those developed among individuals of European ancestry. Conclusions Though not conclusive, our results provide insights on factors influencing discriminatory accuracy of SNP-enhanced models. Genetic variants might be useful to inform stratified CRC screening in the future, but further research is needed.

Abstract 196: Development and Validation of Machine Learning-based Race-specific Models to Predict 10-year Risk of Heart Failure: A Multi-cohort Analysis

Circulation ◽  
2020 ◽  
Vol 142 (Suppl_3) ◽  
Author(s):  
Matthew W Segar ◽  
Byron Jaeger ◽  
Kershaw V Patel ◽  
Vijay Nambi ◽  
Chiadi E Ndumele ◽  
...  
Keyword(s):  
Machine Learning ◽  
Risk Factors ◽  
Heart Failure ◽  
Risk Prediction ◽  
Prediction Models ◽  
External Validation ◽  
Superior Performance ◽  
Risk Models ◽  
Black And White ◽  
White Adults

Introduction: Heart failure (HF) risk and the underlying biological risk factors vary by race. Machine learning (ML) may improve race-specific HF risk prediction but this has not been fully evaluated. Methods: The study included participants from 4 cohorts (ARIC, DHS, JHS, and MESA) aged > 40 years, free of baseline HF, and with adjudicated HF event follow-up. Black adults from JHS and white adults from ARIC were used to derive race-specific ML models to predict 10-year HF risk. The ML models were externally validated in subgroups of black and white adults from ARIC (excluding JHS participants) and pooled MESA/DHS cohorts and compared to prior established HF risk scores developed in ARIC and MESA. Harrell’s C-index and Greenwood-Nam-D’Agostino chi-square were used to assess discrimination and calibration, respectively. Results: In the derivation cohorts, 288 of 4141 (7.0%) black and 391 of 8242 (4.7%) white adults developed HF over 10 years. The ML models had excellent discrimination in both black and white participants (C-indices = 0.88 and 0.89). In the external validation cohorts for black participants from ARIC (excluding JHS, N = 1072) and MESA/DHS pooled cohorts (N = 2821), 131 (12.2%) and 115 (4.1%) developed HF. The ML model had adequate calibration and demonstrated superior discrimination compared to established HF risk models (Fig A). A consistent pattern was also observed in the external validation cohorts of white participants from the MESA/DHS pooled cohorts (N=3236; 100 [3.1%] HF events) (Fig A). The most important predictors of HF in both races were NP levels. Cardiac biomarkers and glycemic parameters were most important among blacks while LV hypertrophy and prevalent CVD and traditional CV risk factors were the strongest predictors among whites (Fig B). Conclusions: Race-specific and ML-based HF risk models that integrate clinical, laboratory, and biomarker data demonstrated superior performance when compared to traditional risk prediction models.

scholarly journals Are existing dementia risk prediction models reliable?

2020 ◽  
Vol 16 (S10) ◽  
Author(s):  
Jantje Goerdten ◽  
Iva Čukić ◽  
Samuel O Danso ◽  
Isabelle Carriere ◽  
Graciela Muniz Terrera
Keyword(s):  
Risk Prediction ◽  
Prediction Models ◽  
Risk Prediction Models ◽  
Dementia Risk

scholarly journals A Review on Automatic Mammographic Density and Parenchymal Segmentation

2015 ◽  
Vol 2015 ◽  
pp. 1-31 ◽  
Author(s):  
Wenda He ◽  
Arne Juette ◽  
Erika R. E. Denton ◽  
Arnau Oliver ◽  
Robert Martí ◽  
...  
Keyword(s):  
Breast Cancer ◽  
Risk Factors ◽  
Risk Assessment ◽  
Risk Prediction ◽  
Prediction Models ◽  
Prevention Measures ◽  
Clinical Environment ◽  
Tissue Segmentation ◽  
Risk Prediction Models ◽  
The Impact

Breast cancer is the most frequently diagnosed cancer in women. However, the exact cause(s) of breast cancer still remains unknown. Early detection, precise identification of women at risk, and application of appropriate disease prevention measures are by far the most effective way to tackle breast cancer. There are more than 70 common genetic susceptibility factors included in the current non-image-based risk prediction models (e.g., the Gail and the Tyrer-Cuzick models). Image-based risk factors, such as mammographic densities and parenchymal patterns, have been established as biomarkers but have not been fully incorporated in the risk prediction models used for risk stratification in screening and/or measuring responsiveness to preventive approaches. Within computer aided mammography, automatic mammographic tissue segmentation methods have been developed for estimation of breast tissue composition to facilitate mammographic risk assessment. This paper presents a comprehensive review of automatic mammographic tissue segmentation methodologies developed over the past two decades and the evidence for risk assessment/density classification using segmentation. The aim of this review is to analyse how engineering advances have progressed and the impact automatic mammographic tissue segmentation has in a clinical environment, as well as to understand the current research gaps with respect to the incorporation of image-based risk factors in non-image-based risk prediction models.

scholarly journals Risk Prediction Models for Erosive Wear in Preschool-aged Children: A Prospective Study

2021 ◽  
Author(s):  
Gabriella Gatt ◽  
Nikolai Attard
Keyword(s):  
Risk Factors ◽  
Risk Prediction ◽  
Prediction Models ◽  
Demographic Factors ◽  
Erosive Wear ◽  
Categorical Variables ◽  
Operating Characteristics ◽  
Risk Prediction Models ◽  
Specific Risk ◽  
Over Time

Abstract BackgroundDespite increasing prevalence, age specific risk predictive models for erosive tooth wear in preschool age children have not been developed. Identification of at risk groups and the timely introduction of behavioural change or treatment will stop the progression of erosive wear in the permanent dentition. The aim of this study was to identify age specific risk factors for erosive wear. Distinct risk prediction models for three year old and five year old children were developed.MethodsA prospective cohort study included clinical examinations and parent administered questionnaires for three and five-year-old children. Chi-square tests explored categorical demographic variables, Spearman Rank Order correlation tests examined changes in BEWE scores with changes in food frequencies while Wilcoxon signed rank tests evaluated the temporal effect of frequencies of consumption of dietary items. Mann-Whitney U tests compared changes in BEWE scores over time for the twenty-six bivariate categorical variables and Kruskall-Wallis tests compared changes in BEWE scores over time across the remaining 55 categorical variables representing demographic factors, oral hygiene habits and dietary habits. Change in BEWE scores for continuous variables was investigated using Spearman Rho correlation coefficient Test. Those variables showing significance with a difference in BEWE cumulative score over time were utilised to develop two risk prediction models. The models were evaluated by Receiver Operating Characteristics (ROC) analysis.ResultsRisk factors for the three-year-old cohort included the erosive wear (χ2 (1, 92) = 12.829, p < 0.001), district (χ2 (5, 92) = 17.032, p = 0.004) and family size (χ2 (1, 92) = 4.547, p = 0.033). Risk factors for the five-year-old cohort also included erosive wear (χ2 (1, 144) = 4.768, p = 0.029) gender (χ2 (1, 144) = 19.399, p <0.001), consumption of iced tea (χ2 (1, 144) = 8.872, p = 0.003) and dry mouth (χ2 (1, 144) = 9.598, p = 0.002).Conclusions: Predictive risk factors for three-year-old children are based on demographic factors and are distinct from those for the five-year-old cohort, which are based on biological and behavioural factors. The presence of erosive wear is a risk factor for further wear in both age cohorts.

scholarly journals Development and Evaluation of Models to Predict Readmission or Death After Discharge from Intensive Care 

2021 ◽  
Author(s):  
Jamie M Boyd ◽  
Matthew T James ◽  
Danny J Zuege ◽  
Henry Thomas Stelfox
Keyword(s):  
Risk Factors ◽  
Systematic Review ◽  
Intensive Care ◽  
Risk Prediction ◽  
Prediction Models ◽  
Meta Analysis ◽  
Local Data ◽  
Predictive Values ◽  
Risk Prediction Models ◽  
Icu Discharge

Abstract Background Patients being discharged from the intensive care unit (ICU) have variable risks of subsequent readmission or death; however, there is limited understanding of how to predict individual patient risk. We sought to derive risk prediction models for ICU readmission or death after ICU discharge to guide clinician decision-making. Methods Systematic review and meta-analysis to identify risk factors. Development and validation of risk prediction models using two retrospective cohorts of patients discharged alive from medical-surgical ICUs (n = 3 ICUs, n = 11,291 patients; n = 14 ICUs, n = 11,400 patients). Models were developed using literature and data-derived weighted coefficients. Results Sixteen variables identified from the systematic review were used to develop four risk prediction models. In the validation cohort there were 795 (7%) patients who were re-admitted to ICU and 703 (7%) patients who died after ICU discharge. The area under the curve (AUROC) for ICU readmission for the literature (0.615 [95%CI: 0.593, 0.637]) and data (0.652 [95%CI: 0.631, 0.674]) weighted models showed poor discrimination. The AUROC for death after ICU discharge for the literature (0.708 [95%CI: 0.687, 0.728]) and local data weighted (0.752 [95%CI: 0.733, 0.770]) models showed good discrimination. The negative predictive values for ICU readmission and death after ICU discharge ranged from 94%-98%. Conclusions Identifying risk factors and weighting coefficients using systematic review and meta-analysis to develop prediction models is feasible and can identify patients at low risk of ICU readmission or death after ICU discharge.

scholarly journals Development and Validation of Machine Learning-Based Race-Specific Models to Predict 10-Year Risk of Heart Failure: A Multi-Cohort Analysis

Circulation ◽  
2021 ◽  
Author(s):  
Matthew W. Segar ◽  
Byron C. Jaeger ◽  
Kershaw V. Patel ◽  
Vijay Nambi ◽  
Chiadi E. Ndumele ◽  
...  
Keyword(s):  
Machine Learning ◽  
Risk Factors ◽  
Heart Failure ◽  
Risk Prediction ◽  
Prediction Models ◽  
Cohort Analysis ◽  
Specific Model ◽  
Superior Performance ◽  
Risk Models ◽  
White Adults

Background: Heart failure (HF) risk and the underlying risk factors vary by race. Traditional models for HF risk prediction treat race as a covariate in risk prediction and do not account for significant parameters such as cardiac biomarkers. Machine learning (ML) may offer advantages over traditional modeling techniques to develop race-specific HF risk prediction models and elucidate important contributors of HF development across races. Methods: We performed a retrospective analysis of four large, community cohort studies (ARIC, DHS, JHS, and MESA) with adjudicated HF events. Participants were aged >40 years and free of HF at baseline. Race-specific ML models for HF risk prediction were developed in the JHS cohort (for Black race-specific model) and White adults from ARIC (for White rate-specific model). The models included 39 candidate variables across demographic, anthropometric, medical history, laboratory, and electrocardiographic domains. The ML models were externally validated and compared with prior established traditional and non-race specific ML models in race-specific subgroups of the pooled MESA/DHS cohort and Black participants of ARIC. Harrell's C-index and Greenwood-Nam-D'Agostino chi-square tests were used to assess discrimination and calibration, respectively. Results: The ML models had excellent discrimination in the derivation cohorts for Black (N=4,141 in JHS, C-index=0.88) and White (N=7,858 in ARIC, C-index=0.89) participants. In the external validation cohorts, the race-specific ML model demonstrated adequate calibration and superior discrimination (C-indices=0.80-0.83 [for Black individuals] and 0.82 [for White individuals]) compared with established HF risk models or with non-race specific ML models derived using race as a covariate. Among the risk factors, natriuretic peptide levels were the most important predictor of HF risk across both races, followed by troponin levels in Black and EKG-based Cornell voltage in White individuals. Other key predictors of HF risk among Black individuals were glycemic parameters and socioeconomic factors. In contrast, prevalent cardiovascular (CV) disease and traditional CV risk factors were stronger predictors of HF risk in White adults. Conclusions: Race-specific and ML-based HF risk models that integrate clinical, laboratory, and biomarker data demonstrated superior performance when compared with traditional HF risk and non-race specific ML models. This approach identifies distinct race-specific contributors of HF.

Sign in / Sign up

Export Citation Format

Share Document