scholarly journals Random forest model for feature-based Alzheimer’s disease conversion prediction from early mild cognitive impairment subjects

PLoS ONE ◽  
2021 ◽  
Vol 16 (4) ◽  
pp. e0244773
Author(s):  
Matthew Velazquez ◽  
Yugyung Lee ◽  
Keyword(s):  
Alzheimer’S Disease ◽  
Alzheimer's Disease ◽  
Cognitive Impairment ◽  
Mild Cognitive Impairment ◽  
Random Forest ◽  
Clinical Features ◽  
Random Forest Model ◽  
Clinical Feature ◽  
Forest Model ◽  
Early Mild Cognitive Impairment

Alzheimer’s Disease (AD) conversion prediction from the mild cognitive impairment (MCI) stage has been a difficult challenge. This study focuses on providing an individualized MCI to AD conversion prediction using a balanced random forest model that leverages clinical data. In order to do this, 383 Early Mild Cognitive Impairment (EMCI) patients were gathered from the Alzheimer’s Disease Neuroimaging Initiative (ADNI). Of these patients, 49 would eventually convert to AD (EMCI_C), whereas the remaining 334 did not convert (EMCI_NC). All of these patients were split randomly into training and testing data sets with 95 patients reserved for testing. Nine clinical features were selected, comprised of a mix of demographic, brain volume, and cognitive testing variables. Oversampling was then performed in order to balance the initially imbalanced classes prior to training the model with 1000 estimators. Our results showed that a random forest model was effective (93.6% accuracy) at predicting the conversion of EMCI patients to AD based on these clinical features. Additionally, we focus on explainability by assessing the importance of each clinical feature. Our model could impact the clinical environment as a tool to predict the conversion to AD from a prodromal stage or to identify ideal candidates for clinical trials.

scholarly journals Random forest model for feature-based Alzheimer’s disease conversion prediction from early mild cognitive impairment subjects

2020 ◽  
Author(s):  
Matthew Velazquez ◽  
Yugyung Lee ◽  
Keyword(s):  
Alzheimer’S Disease ◽  
Alzheimer's Disease ◽  
Cognitive Impairment ◽  
Mild Cognitive Impairment ◽  
Random Forest ◽  
Clinical Features ◽  
Random Forest Model ◽  
Clinical Feature ◽  
Forest Model ◽  
Early Mild Cognitive Impairment

AbstractAlzheimer’s Disease (AD) conversion prediction from the mild cognitive impairment (MCI) stage has been a difficult challenge. This study focuses on providing an individualized MCI to AD conversion prediction using a balanced random forest model that leverages clinical data. In order to do this, 383 Early Mild Cognitive Impairment (EMCI) patients were gathered from the Alzheimer’s Disease Neuroimaging Initiative (ADNI). Of these patients, 49 would eventually convert to AD (EMCI_C), whereas the remaining 335 did not convert (EMCI_NC). All of these patients were split into training and testing data sets with 95 patients reserved for testing. Nine clinical features were selected, comprised of a mix of demographic, brain volume, and cognitive testing variables. Oversampling was then performed in order to balance the initially imbalanced classes. Our results showed that a random forest model was effective (93.6% accuracy) at predicting the conversion of EMCI patients to AD based on these clinical features. Additionally, we assessed the importance of each clinical feature at both the individual and model level for interpretation of the prediction itself. Our model could impact the clinical environment as a tool to predict the conversion to AD from a prodromal stage or to identify ideal candidates for clinical trials.

scholarly journals Recruitment and Retention of Participant and Study Partner Dyads in Two Multinational Alzheimer’s Disease Registration Trials

2020 ◽  
Author(s):  
Olivia M Bernstein ◽  
Joshua D. Grill ◽  
Daniel L. Gillen
Keyword(s):  
Alzheimer’S Disease ◽  
Alzheimer's Disease ◽  
Random Forest ◽  
Random Forest Model ◽  
P Value ◽  
Confounding Factors ◽  
Forest Model ◽  
Partner Type ◽  
Study Partner ◽  
The Relationship

Abstract Background: Early study exit is detrimental to statistical power and increases the risk for bias in Alzheimer’s disease clinical trials. Previous analyses in early phase academic trials demonstrated associations between rates of trial incompletion and participants’ study partner type, with participants enrolling with non-spouse study partners being at greater risk.Methods: We conducted secondary analyses of two multinational phase III trials of semagacestat, an oral gamma secretase inhibitor, for mild-to-moderate AD dementia. Cox’s proportional hazards regression model was used to estimate the relationship between study partner type and the risk of early exit from the trial after adjustment for a priori identified potential confounding factors. Additionally, we used a random forest model to identify top predictors of dropout.Results: Among participants with spousal, adult child, and other study partners, respectively, 35%, 38%, and 36% dropped out or died prior to protocol-defined study completion, respectively. In unadjusted models, the risk of trial incompletion differed by study partner type (unadjusted p-value=0.027 for test of differences by partner type), but in models adjusting for potential confounding factors the differences were not statistically significant (p-value=0.928). In exploratory modeling, participant age was identified as the primary characteristic to explain the relationship between study partner type and the risk of failing to complete the trial. Participant age was also the strongest predictor of trial incompletion in the random forest model.Conclusions: After adjustment for age, no qualitative differences in the risk of incompletion were observed when comparing participants with different study partner types in these trials. Differences between our findings and the findings of previous studies may be explained by differences in trial phase, size, geographic regions, or the composition of academic and non-academic sites.

Increase of EEG Alpha3/Alpha2 Power Ratio Detects Inferior Parietal Lobule Atrophy in Mild Cognitive Impairment

2018 ◽  
Vol 15 (5) ◽  
pp. 443-451
Author(s):  
Davide V. Moretti
Keyword(s):  
Alzheimer’S Disease ◽  
Alzheimer's Disease ◽  
Cognitive Impairment ◽  
Mild Cognitive Impairment ◽  
Clinical Feature ◽  
Power Ratio ◽  
Inferior Parietal Lobule ◽  
Mri Scans ◽  
Parietal Lobule ◽  
Early Alzheimer’S Disease

Objective: The inferior parietal lobule (IPL) has been implicate in many higher cognitive processes, as visuo-motor transformations, tool use or tool making. In subjects with mild cognitive impairment (MCI) at major risk to develop Alzheimer' Disease (AD) an impairment of subtle visuomotor or praxic abilities is a well-known clinical feature. Enhance of the ratio of EEG alpha3/alpha2 frequency power was detected in subjects with MCI who will transform in Alzheimer's disease (AD). Methods: We explored of the association of alpha3/alpha2 power ratio with cortical size of IPL in patients with MCI. 74 subjects with MCI undergone EEG recording and MRI scans. Alpha3/alpha2 power ratio in addition to cortical size had been computed for each patient. Three MCI groups had been acquired in keeping with growing tertile values of alpha3/alpha2 ratio. Huge difference of cortical thickness among the groups was calculated. Higher alpha3/alpha2 power ratio group had broader cortical loss compared to other teams on the IPL, particularly in the Supramarginal Gyrus, and Precuneus on both hemispheres. Results: Our results unveil the possible part that the IPL could play in determining the classic alterations of early Alzheimer's disease (AD). Conclusion: Finally, the rise of alpha3/alpha2 power ratio detected a focused anatomo-functional association that could be a reliable marker of incipient AD.

scholarly journals Recruitment and retention of participant and study partner dyads in two multinational Alzheimer’s disease registration trials

2021 ◽  
Vol 13 (1) ◽  
Author(s):  
Olivia M. Bernstein ◽  
Joshua D. Grill ◽  
Daniel L. Gillen
Keyword(s):  
Alzheimer’S Disease ◽  
Alzheimer's Disease ◽  
Random Forest ◽  
Random Forest Model ◽  
Phase Iii ◽  
Confounding Factors ◽  
Forest Model ◽  
Partner Type ◽  
Study Partner ◽  
The Relationship

Abstract Background Early study exit is detrimental to statistical power and increases the risk for bias in Alzheimer’s disease clinical trials. Previous analyses in early phase academic trials demonstrated associations between rates of trial incompletion and participants’ study partner type, with participants enrolling with non-spouse study partners being at greater risk. Methods We conducted secondary analyses of two multinational phase III trials of semagacestat, an oral gamma secretase inhibitor, for mild-to-moderate AD dementia. Cox’s proportional hazards regression model was used to estimate the relationship between study partner type and the risk of early exit from the trial after adjustment for a priori identified potential confounding factors. Additionally, we used a random forest model to identify top predictors of dropout. Results Among participants with spousal, adult child, and other study partners, respectively, 35%, 38%, and 36% dropped out or died prior to protocol-defined study completion, respectively. In unadjusted models, the risk of trial incompletion differed by study partner type (unadjusted p value = 0.027 for test of differences by partner type), but in models adjusting for potential confounding factors, the differences were not statistically significant (p value = 0.928). In exploratory modeling, participant age was identified as the primary characteristic to explain the relationship between study partner type and the risk of failing to complete the trial. Participant age was also the strongest predictor of trial incompletion in the random forest model. Conclusions After adjustment for age, no differences in the risk of incompletion were observed when comparing participants with different study partner types in these trials. Differences between our findings and the findings of previous studies may be explained by differences in trial phase, size, geographic regions, or the composition of academic and non-academic sites.

scholarly journals Peripheral inflammation in mild cognitive impairment with possible and probable Lewy body disease and Alzheimer’s disease

2019 ◽  
Vol 31 (04) ◽  
pp. 551-560 ◽  
Author(s):  
Eleanor King ◽  
John Tiernan O’Brien ◽  
Paul Donaghy ◽  
Christopher Morris ◽  
Nicola Barnett ◽  
...  
Keyword(s):  
Alzheimer’S Disease ◽  
Alzheimer's Disease ◽  
Cognitive Impairment ◽  
Mild Cognitive Impairment ◽  
Clinical Features ◽  
Lewy Body ◽  
Clinical Symptoms ◽  
Lewy Body Disease ◽  
Peripheral Inflammation ◽  
Protein Levels

ABSTRACTObjectives and design:To Investigate the peripheral inflammatory profile in patients with mild cognitive impairment (MCI) from three subgroups – probable Lewy body disease (probable MCI-LB), possible Lewy body disease, and probable Alzheimer’s disease (probable MCI-AD) – as well as associations with clinical features.Setting:Memory clinics and dementia services.Participants:Patients were classified based on clinical symptoms as probable MCI-LB (n = 38), possible MCI-LB (n = 18), and probable MCI-AD (n = 21). Healthy comparison subjects were recruited (n = 20).Measurements:Ten cytokines were analyzed from plasma samples: interferon (IFN)-gamma, interleukin (IL)-1beta, IL-2, IL-4, IL-6, IL-8, IL-10, IL-12p70, IL-13, and tumor necrosis factor (TNF)-alpha. C-reactive protein levels were investigated.Results:There was a higher level of IL-10, IL-1beta, IL-2, and IL-4 in MCI groups compared to the healthy comparison group (p < 0.0085). In exploratory analyses to understand these findings, the MC-AD group lower IL-1beta (p = 0.04), IL-2 (p = 0.009), and IL-4 (p = 0.012) were associated with increasing duration of memory symptoms, and in the probable MCI-LB group, lower levels of IL-1beta were associated with worsening motor severity (p = 0.002). In the possible MCI-LB, longer duration of memory symptoms was associated with lower levels of IL-1beta (p = 0.003) and IL-4 (p = 0.026).Conclusion:There is increased peripheral inflammation in patients with MCI compared to healthy comparison subjects regardless of the MCI subtype. These possible associations with clinical features are consistent with other work showing that inflammation is increased in early disease but require replication. Such findings have importance for timing of putative therapeutic strategies aimed at lowering inflammation.

Sign in / Sign up

Export Citation Format

Share Document