Predicting Amyloid Positivity in Cognitively Unimpaired Older Adults: A Machine Learning Approach Using A4 Data

SymScore: Machine learning accuracy meets transparency in a symbolic regression-based clinical score generator

Self-report questionnaires play a crucial role in healthcare for assessing disease risks, yet their extensive length can be burdensome for respondents, potentially compromising data quality. To address this, machine learning-based shortened questionnaires have been developed. While these questionnaires possess high levels of accuracy, their practical use in clinical settings is hindered by a lack of transparency and the need for specialized machine learning expertise. This makes their integration into clinical workflows challenging and also decreases trust among healthcare professionals who prefer interpretable tools for decision-making. To preserve both predictive accuracy and interpretability, this study introduces the Symbolic Regression-Based Clinical Score Generator (SymScore). SymScore produces score tables for shortened questionnaires, which enable clinicians to estimate the results that reflect those of the original questionnaires. SymScore generates the score tables by optimally grouping responses, assigning weights based on predictive importance, imposing necessary constraints, and fitting models via symbolic regression. We compared SymScore's performance with the machine learning-based shortened questionnaires MCQI-6 (n=310) and SLEEPS (n=4257), both renowned for their high accuracy in assessing sleep disorders. SymScore's questionnaire demonstrated comparable performance (MAE = 10.73, R2 = 0.77) to that of the MCQI-6 (MAE = 9.94, R2 = 0.82) and achieved AUROC values of 0.85-0.91 for various sleep disorders, closely matching those of SLEEPS (0.88-0.94). By generating accurate and interpretable score tables, SymScore ensures that healthcare professionals can easily explain and trust its results without specialized machine learning knowledge. Thus, SymScore advances explainable AI for healthcare by offering a user-friendly and resource-efficient alternative to machine learning-based questionnaires, supporting improved patient outcomes and workflow efficiency.

Machine learning models for dementia screening to classify brain amyloid positivity on positron emission tomography using blood markers and demographic characteristics: a retrospective observational study

BACKGROUND

Intracerebral amyloid β (Aβ) accumulation is considered the initial observable event in the pathological process of Alzheimer's disease (AD). Efficient screening for amyloid pathology is critical for identifying patients for early treatment. This study developed machine learning models to classify positron emission tomography (PET) Aβ-positivity in participants with preclinical and prodromal AD using data accessible to primary care physicians.

METHODS

This retrospective observational study assessed the classification performance of combinations of demographic characteristics, routine blood test results, and cognitive test scores to classify PET Aβ-positivity using machine learning. Participants with mild cognitive impairment (MCI) or normal cognitive function who visited Oita University Hospital or had participated in the USUKI study and met the study eligibility criteria were included. The primary endpoint was assessment of the classification performance of the presence or absence of intracerebral Aβ accumulation using five machine learning models (i.e., five combinations of variables), each constructed with three classification algorithms, resulting in a total of 15 patterns. L2-regularized logistic regression, and kernel Support Vector Machine (SVM) and Elastic Net algorithms were used to construct the classification models using 34 pre-selected variables (12 demographic characteristics, 11 blood test results, 11 cognitive test results).

RESULTS

Data from 262 records (260 unique participants) were analyzed. The mean (standard deviation [SD]) participant age was 73.8 (7.8) years. Using L2-regularized logistic regression, the mean receiver operating characteristic (ROC) area under the curve (AUC) (SD) in Model 0 (basic demographic characteristics) was 0.67 (0.01). Classification performance was similar in Model 1 (basic demographic characteristics and Mini Mental State Examination [MMSE] subscores) and Model 2 (demographic characteristics and blood test results) with a cross-validated mean ROC AUC (SD) of 0.70 (0.01) for both. Model 3 (demographic characteristics, blood test results, MMSE subscores) and Model 4 (Model 3 and ApoE4 phenotype) showed improved performance with a mean ROC AUC (SD) of 0.73 (0.01) and 0.76 (0.01), respectively. In models using blood test results, thyroid-stimulating hormone and mean corpuscular volume tended to be the largest contributors to classification. Classification performances were similar using the SVM and Elastic Net algorithms.

CONCLUSIONS

The machine learning models used in this study were useful for classifying PET Aβ-positivity using data from routine physician visits.

TRIAL REGISTRATION

UMIN Clinical Trials Registry (UMIN000051776, registered on 31/08/2023).

Protocol for a double-blind placebo-controlled randomised controlled trial assessing the impact of oral semaglutide in amyloid positivity (ISAP) in community dwelling UK adults

INTRODUCTION

Glucagon-like peptide-1 receptor agonists (GLP-1 RAs), currently marketed for type 2 diabetes and obesity, may offer novel mechanisms to delay or prevent neurotoxicity associated with Alzheimer's disease (AD). The impact of semaglutide in amyloid positivity (ISAP) trial is investigating whether the GLP-1 RA semaglutide reduces accumulation in the brain of cortical tau protein and neuroinflammation in individuals with preclinical/prodromal AD.

METHODS AND ANALYSIS

ISAP is an investigator-led, randomised, double-blind, superiority trial of oral semaglutide compared with placebo. Up to 88 individuals aged ≥55 years with brain amyloid positivity as assessed by positron emission tomography (PET) or cerebrospinal fluid, and no or mild cognitive impairment, will be randomised. People with the low-affinity binding variant of the rs6971 allele of the Translocator Protein 18 kDa (TSPO) gene, which can interfere with interpreting TSPO PET scans (a measure of neuroinflammation), will be excluded.At baseline, participants undergo tau, TSPO PET and MRI scanning, and provide data on physical activity and cognition. Eligible individuals are randomised in a 1:1 ratio to once-daily oral semaglutide or placebo, starting at 3 mg and up-titrating to 14 mg over 8 weeks. They will attend safety visits and provide blood samples to measure AD biomarkers at weeks 4, 8, 26 and 39. All cognitive assessments are repeated at week 26. The last study visit will be at week 52, when all baseline measurements will be repeated. The primary end point is the 1-year change in tau PET signal.

ETHICS AND DISSEMINATION

The study was approved by the West Midlands-Edgbaston Research Ethics Committee (22/WM/0013). The results of the study will be disseminated through scientific presentations and peer-reviewed publications.

TRIAL REGISTRATION NUMBER

ISRCTN71283871.

Predicting Brain Amyloid Status Using the National Institute of Health Toolbox (NIHTB) for Assessment of Neurological and Behavioral Function

BACKGROUND

Amyloid-beta (Aβ) plaque is a neuropathological hallmark of Alzheimer's disease (AD). As anti-amyloid monoclonal antibodies enter the market, predicting brain amyloid status is critical to determine treatment eligibility.

OBJECTIVE

To predict brain amyloid status utilizing machine learning approaches in the Advancing Reliable Measurement in Alzheimer's Disease and Cognitive Aging (ARMADA) study.

DESIGN

ARMADA is a multisite study that implemented the National Institute of Health Toolbox for Assessment of Neurological and Behavioral Function (NIHTB) in older adults with different cognitive ability levels (normal, mild cognitive impairment, early-stage dementia of the AD type).

SETTING

Participants across various sites were involved in the ARMADA study for validating the NIHTB.

PARTICIPANTS

199 ARMADA participants had either PET or CSF information (mean age 76.3 ± 7.7, 51.3% women, 42.3% some or complete college education, 50.3% graduate education, 88.9% White, 33.2% with positive AD biomarkers).

MEASUREMENTS

We used cognition, emotion, motor, sensation scores from NIHTB, and demographics to predict amyloid status measured by PET or CSF. We applied LASSO and random forest models and used the area under the receiver operating curve (AUROC) to evaluate the ability to identify amyloid positivity.

RESULTS

The random forest model reached AUROC of 0.74 with higher specificity than sensitivity (AUROC 95% CI:0.73 - 0.76, Sensitivity 0.50, Specificity 0.88) on the held-out test set; higher than the LASSO model (0.68 (95% CI:0.68 - 0.69)). The 10 features with the highest importance from the random forest model are: picture sequence memory, cognition total composite, cognition fluid composite, list sorting working memory, words-in-noise test (hearing), pattern comparison processing speed, odor identification, 2-minutes-walk endurance, 4-meter walk gait speed, and picture vocabulary. Overall, our model revealed the validity of measurements in cognition, motor, and sensation domains, in associating with AD biomarkers.

CONCLUSION

Our results support the utilization of the NIH toolbox as an efficient and standardizable AD biomarker measurement that is better at identifying amyloid negative (i.e., high specificity) than positive cases (i.e., low sensitivity).

Predicting amyloid-beta pathology in the general population

INTRODUCTION

Reliable models to predict amyloid beta (Aβ) positivity in the general aging population are lacking but could become cost-efficient tools to identify individuals at risk of developing Alzheimer's disease.

METHODS

We developed Aβ prediction models in the clinical Anti-Amyloid Treatment in Asymptomatic Alzheimer's (A4) Study (n = 4,119) including a broad range of easily ascertainable predictors (demographics, cognition and daily functioning, health and lifestyle factors). Importantly, we determined the generalizability of our models in the population-based Rotterdam Study (n = 500).

RESULTS

The best performing model in the A4 Study (area under the curve [AUC] = 0.73 [0.69-0.76]), including age, apolipoprotein E (APOE) ε4 genotype, family history of dementia, and subjective and objective measures of cognition, walking duration and sleep behavior, was validated in the independent Rotterdam Study with higher accuracy (AUC = 0.85 [0.81-0.89]). Yet, the improvement relative to a model including only age and APOE ε4 was marginal.

DISCUSSION

Aβ prediction models including inexpensive and non-invasive measures were successfully applied to a general population-derived sample more representative of typical older non-demented adults.

Identification of preclinical dementia according to ATN classification for stratified trial recruitment: A machine learning approach

INTRODUCTION

The Amyloid/Tau/Neurodegeneration (ATN) framework was proposed to identify the preclinical biological state of Alzheimer's disease (AD). We investigated whether ATN phenotype can be predicted using routinely collected research cohort data.

METHODS

927 EPAD LCS cohort participants free of dementia or Mild Cognitive Impairment were separated into 5 ATN categories. We used machine learning (ML) methods to identify a set of significant features separating each neurodegeneration-related group from controls (A-T-(N)-). Random Forest and linear-kernel SVM with stratified 5-fold cross validations were used to optimize model whose performance was then tested in the ADNI database.

RESULTS

Our optimal results outperformed ATN cross-validated logistic regression models by between 2.2% and 8.3%. The optimal feature sets were not consistent across the 4 models with the AD pathologic change vs controls set differing the most from the rest. Because of that we have identified a subset of 10 features that yield results very close or identical to the optimal.

DISCUSSION

Our study demonstrates the gains offered by ML in generating ATN risk prediction over logistic regression models among pre-dementia individuals.

MRI-based Deep Learning Assessment of Amyloid, Tau, and Neurodegeneration Biomarker Status across the Alzheimer Disease Spectrum

Background PET can be used for amyloid-tau-neurodegeneration (ATN) classification in Alzheimer disease, but incurs considerable cost and exposure to ionizing radiation. MRI currently has limited use in characterizing ATN status. Deep learning techniques can detect complex patterns in MRI data and have potential for noninvasive characterization of ATN status. Purpose To use deep learning to predict PET-determined ATN biomarker status using MRI and readily available diagnostic data. Materials and Methods MRI and PET data were retrospectively collected from the Alzheimer's Disease Imaging Initiative. PET scans were paired with MRI scans acquired within 30 days, from August 2005 to September 2020. Pairs were randomly split into subsets as follows: 70% for training, 10% for validation, and 20% for final testing. A bimodal Gaussian mixture model was used to threshold PET scans into positive and negative labels. MRI data were fed into a convolutional neural network to generate imaging features. These features were combined in a logistic regression model with patient demographics, APOE gene status, cognitive scores, hippocampal volumes, and clinical diagnoses to classify each ATN biomarker component as positive or negative. Area under the receiver operating characteristic curve (AUC) analysis was used for model evaluation. Feature importance was derived from model coefficients and gradients. Results There were 2099 amyloid (mean patient age, 75 years ± 10 [SD]; 1110 male), 557 tau (mean patient age, 75 years ± 7; 280 male), and 2768 FDG PET (mean patient age, 75 years ± 7; 1645 male) and MRI pairs. Model AUCs for the test set were as follows: amyloid, 0.79 (95% CI: 0.74, 0.83); tau, 0.73 (95% CI: 0.58, 0.86); and neurodegeneration, 0.86 (95% CI: 0.83, 0.89). Within the networks, high gradients were present in key temporal, parietal, frontal, and occipital cortical regions. Model coefficients for cognitive scores, hippocampal volumes, and APOE status were highest. Conclusion A deep learning algorithm predicted each component of PET-determined ATN status with acceptable to excellent efficacy using MRI and other available diagnostic data. © RSNA, 2023 Supplemental material is available for this article.

Association of Stages of Objective Memory Impairment With Incident Symptomatic Cognitive Impairment in Cognitively Normal Individuals

BACKGROUND AND OBJECTIVES

Increasing evidence indicates that a subset of cognitively normal individuals has subtle cognitive impairment at baseline. We sought to identify them using the Stages of Objective Memory Impairment (SOMI) system. Symptomatic cognitive impairment was operationalized by a Clinical Dementia Rating (CDR) ≥0.5. We hypothesized that incident impairment would be higher for participants with subtle retrieval impairment (SOMI-1), higher still for those with moderate retrieval impairment (SOMI-2), and highest for those with storage impairment (SOMI-3/4) after adjusting for demographics and APOE ε4 status. A secondary objective was to determine whether including biomarkers of β-amyloid, tau pathology, and neurodegeneration in the models affect prediction. We hypothesized that even after adjusting for in vivo biomarkers, SOMI would remain a significant predictor of time to incident symptomatic cognitive impairment.

METHODS

Among 969 cognitively normal participants, defined by a CDR = 0, from the Knight Alzheimer Disease Research Center, SOMI stage was determined from their baseline Free and Cued Selective Reminding Test scores, 555 had CSF and structural MRI measures and comprised the biomarker subgroup, and 144 of them were amyloid positive. Cox proportional hazards models tested associations of SOMI stages at baseline and biomarkers with time to incident cognitive impairment defined as the transition to CDR ≥0.5.

RESULTS

Among all participants, the mean age was 69.35 years, 59.6% were female, and mean follow-up was 6.36 years. Participants in SOMI-1-4 had elevated hazard ratios for the transition from normal to impaired cognition in comparison with those who were SOMI-0 (no memory impairment). Individuals in SOMI-1 (mildly impaired retrieval) and SOMI-2 (moderately impaired retrieval) were at nearly double the risk of clinical progression compared with persons with no memory problems. When memory storage impairment emerges (SOMI-3/4), the hazard ratio for clinical progression increased approximately 3 times. SOMI stage remained an independent predictor of incident cognitive impairment after adjusting for all biomarkers.

DISCUSSION

SOMI predicts the transition from normal cognition to incident symptomatic cognitive impairment (CDR ≥0.5). The results support the use of SOMI to identify those cognitively normal participants most likely to develop incident cognitive impairment who can then be referred for biomarker screening.

A universal AutoScore framework to develop interpretable scoring systems for predicting common types of clinical outcomes

The AutoScore framework can automatically generate data-driven clinical scores in various clinical applications. Here, we present a protocol for developing clinical scoring systems for binary, survival, and ordinal outcomes using the open-source AutoScore package. We describe steps for package installation, detailed data processing and checking, and variable ranking. We then explain how to iterate through steps for variable selection, score generation, fine-tuning, and evaluation to generate understandable and explainable scoring systems using data-driven evidence and clinical knowledge. For complete details on the use and execution of this protocol, please refer to Xie et al. (2020),1 Xie et al. (2022)2, Saffari et al. (2022)3 and the online tutorial https://nliulab.github.io/AutoScore/.

AutoScore-Ordinal: an interpretable machine learning framework for generating scoring models for ordinal outcomes

Background

Risk prediction models are useful tools in clinical decision-making which help with risk stratification and resource allocations and may lead to a better health care for patients. AutoScore is a machine learning–based automatic clinical score generator for binary outcomes. This study aims to expand the AutoScore framework to provide a tool for interpretable risk prediction for ordinal outcomes.

Methods

The AutoScore-Ordinal framework is generated using the same 6 modules of the original AutoScore algorithm including variable ranking, variable transformation, score derivation (from proportional odds models), model selection, score fine-tuning, and model evaluation. To illustrate the AutoScore-Ordinal performance, the method was conducted on electronic health records data from the emergency department at Singapore General Hospital over 2008 to 2017. The model was trained on 70% of the data, validated on 10% and tested on the remaining 20%.

Results

This study included 445,989 inpatient cases, where the distribution of the ordinal outcome was 80.7% alive without 30-day readmission, 12.5% alive with 30-day readmission, and 6.8% died inpatient or by day 30 post discharge. Two point-based risk prediction models were developed using two sets of 8 predictor variables identified by the flexible variable selection procedure. The two models indicated reasonably good performance measured by mean area under the receiver operating characteristic curve (0.758 and 0.793) and generalized c-index (0.737 and 0.760), which were comparable to alternative models.

Conclusion

AutoScore-Ordinal provides an automated and easy-to-use framework for development and validation of risk prediction models for ordinal outcomes, which can systematically identify potential predictors from high-dimensional data.