An official website of the United States government
Abstract INTRODUCTIONFactors underlying discordant visual and quantitative amyloid beta–positron emission tomography (Aβ‐PET) results and their clinical implications are not well understood. METHODSParticipants from the 1Florida Alzheimer's Disease Research Center (1FLADRC) underwent Aβ‐PET, blood draw, brain magnetic resonance imaging (MRI), and neuropsychological testing. We evaluated differences in demographics, apolipoprotein E (APOE) status, biomarkers, and cognition among older adults with concordant and discordant visual‐quantitative Aβ‐PET. Discordance was defined as positive visual read (V) of Aβ‐PET with below‐threshold Centiloid quantification (Q; CL <25; V+/Q–) or negative visual read with CL ≥25 (V–/Q+). RESULTSWe studied 386 participants (mean age ± SD: 70.7 ± 7.8, 55.2% female, 44.6% Hispanic White). Compared to V+/Q–, V–/Q+ had a higher frequency ofAPOEε4 carriers (40%). Black/African American participants were overrepresented in V–/Q+ (40.9%). Both discordant groups had higher plasma phosphorylated tau 217 (p‐tau217) and glial fibrillary acidic protein (GFAP) than V–/Q– but lower than V+/Q+. Discordant groups had greater gray matter volume and better cognitive performance than V+/Q+. DISCUSSIONDiscordant Aβ‐PET findings likely hold clinical significance and may reflect early stages of neuropathological progression. HighlightsGroups with concordant/discordant visual‐quantitative amyloid beta–positron emission tomography (Aβ‐PET) results were compared.Visual–/quant+ were more likely than visual+/quant– to be apolipoprotein E (APOE) ε4 carriers and Black/African American.Discordant groups had higher plasma phosphorylated tau 217 (p‐tau217) and glial fibrillary acidic protein (GFAP) than concordant negative.Discordant groups had less atrophy and better cognition than concordant positive.Centiloid quantification should supplement visual reads in clinical settings.
more »
« less
This content will become publicly available on August 17, 2027
Machine learning for the prediction of amyloid PET positivity using plasma biomarkers, cognition, APOE genotype, and structural imaging
PurposeMachine learning to enable precise, non-invasive detection of cerebral amyloid-beta (Aβ) pathology by integrating cognitive assessments, plasma biomarkers, and structural neuroimaging. MethodsWe developed an explainable multimodal machine-learning framework to predict amyloid PET visual read status using plasma biomarkers, cognitive assessments, APOE genotype, demographic variables, and structural MRI measurements. The development cohort consisted of 170 participants from the 1Florida Alzheimer’s Disease Research Center (ADRC) and 129 participants from the ADNI4 cohort. Machine-learning pipelines were evaluated using combinations of five classifiers and multiple feature-selection approaches with Bayesian hyperparameter optimization and stratified five-fold cross-validation. Model interpretability was assessed using SHapley Additive exPlanations (SHAP). ResultsEnsemble methods consistently outperformed linear and distance-based classifiers. The optimal pipeline, XGBoost with mutual-information feature selection, achieved a mean AUC of 0.891 ± 0.048 and a recall of 0.84. SHAP analyses identified the plasma p-tau217/Aβ42 ratio as the most influential predictor, followed by plasma p-tau217, p-tau181, MMSE, and APOE ε4 status. Structural MRI variables provided complementary predictive information. ADNI4 evaluation yielded a mean AUC of 0.545, while training on ADRC and testing on ADNI4 achieved an AUC of 0.636 and an overall accuracy of 63%. Importantly, plasma p-tau217/Aβ42 and p-tau217 remained the dominant predictors across cohorts. ConclusionMultimodal machine learning can predict amyloid PET status while providing biologically interpretable explanations. Plasma tau and amyloid-related biomarkers carried the strongest predictive signal, while cognition, APOE genotype, and MRI refined classification decisions. External validation highlighted the challenges of cohort heterogeneity and domain shift but demonstrated preservation of biologically meaningful biomarker relationships.
more »
« less
- PAR ID:
- 10711033
- Publisher / Repository:
- Frontiers Media SA
- Date Published:
- Journal Name:
- Frontiers in Neuroscience
- Volume:
- 20
- ISSN:
- 1662-453X
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
IntroductionEarly detection of Alzheimer's disease (AD) requires models that combine brain structure changes with genetic risk, but existing methods struggle to align these different data types. MethodsWe present R-GenIMA, an interpretable multimodal large language model that pairs a region-of-interest vision transformer with genetic prompting to jointly analyze structural MRI and single nucleotide polymorphisms (SNPs). Each brain region becomes a visual token and SNP profiles are encoded as structured text, letting the model link regional atrophy to genetic factors through cross-modal attention. Tested on the ADNI cohort, R-GenIMA performs well in classifying four groups: normal cognition, subjective memory concerns, mild cognitive impairment, and AD. ResultsBeyond accuracy, it produces biologically meaningful explanations, identifying stage-specific brain regions and genes. The model consistently highlighted known AD risk genes (APOE, BIN1, CLU, RBFOX1) and revealed stage-specific patterns: striatal involvement in subjective decline, frontotemporal changes in early impairment, and broad network disruption in AD. DiscussionThese results show that interpretable multimodal AI can integrate imaging and genetics to reveal disease mechanisms, providing a foundation for clinical tools that enable earlier risk assessment and inform precision treatment in Alzheimer's disease.more » « less
-
BackgroundSemantic intrusion errors (SIEs) are associated with mild cognitive impairment (MCI) due to Alzheimer's disease (AD). It is unknown whether accounting for maximum learning capacity still leads to an increase in SIEs when elevated plasma p-tau217, a biological indicator of underlying AD, is present. MethodsOne hundred fifty-eight older adult participants completed the Loewenstein-Acevedo Scales for Semantic Interference and Learning (LASSI-L), a sensitive cognitive challenge test designed to elicit SIEs. Of these, 108 were clinically diagnosed with amnestic MCI (aMCI). Fifty-eight individuals met or exceeded a plasma p-tau217positivity of >0.55 pg/ml, while 50 individuals scored below this threshold. ResultsAfter adjusting for demographic covariates and maximum learning capacity, the aMCI p-tau217+ group evidenced more SIEs compared to aMCI p-tau217- on the first (list B1;p= 0.035) and second trials of the competing list (list B2;p= 0.006). Biological predictors such asApoEε4 status, higher p-tau217, and older age were predictors of an elevated number of SIEs [list B2:F(3,104) = 10.92;p= 0.001;R= 0.489)]. ConclusionsUnlike previous studies that used amyloid PET or other plasma biomarkers, individuals with aMCI p-tau217+ evidenced more SIEs, even after adjusting for their initial learning capacity, a covariate that has not been studied previously. These findings support that SIEs are more prevalent in the presence of underlying AD pathology and occur independent of learning deficits.more » « less
-
Background: Machine learning is a promising tool for biomarker-based diagnosis of Alzheimer’s disease (AD). Performing multimodal feature selection and studying the interaction between biological and clinical AD can help to improve the performance of the diagnosis models. Objective: This study aims to formulate a feature ranking metric based on the mutual information index to assess the relevance and redundancy of regional biomarkers and improve the AD classification accuracy. Methods: From the Alzheimer’s Disease Neuroimaging Initiative (ADNI), 722 participants with three modalities, including florbetapir-PET, flortaucipir-PET, and MRI, were studied. The multivariate mutual information metric was utilized to capture the redundancy and complementarity of the predictors and develop a feature ranking approach. This was followed by evaluating the capability of single-modal and multimodal biomarkers in predicting the cognitive stage. Results: Although amyloid-β deposition is an earlier event in the disease trajectory, tau PET with feature selection yielded a higher early-stage classification F1-score (65.4%) compared to amyloid-β PET (63.3%) and MRI (63.2%). The SVC multimodal scenario with feature selection improved the F1-score to 70.0% and 71.8% for the early and late-stage, respectively. When age and risk factors were included, the scores improved by 2 to 4%. The Amyloid-Tau-Neurodegeneration [AT(N)] framework helped to interpret the classification results for different biomarker categories. Conclusion: The results underscore the utility of a novel feature selection approach to reduce the dimensionality of multimodal datasets and enhance model performance. The AT(N) biomarker framework can help to explore the misclassified cases by revealing the relationship between neuropathological biomarkers and cognition.more » « less
-
BackgroundMetastatic cancer remains one of the leading causes of cancer-related mortality worldwide. Yet, the prediction of survivability in this population remains limited by heterogeneous clinical presentations and high-dimensional molecular features. Advances in machine learning (ML) provide an opportunity to integrate diverse patient- and tumor-level factors into explainable predictive ML models. Leveraging large real-world datasets and modern ML techniques can enable improved risk stratification and precision oncology. ObjectiveThis study aimed to develop and interpret ML models for predicting overall survival in patients with metastatic cancer using the Memorial Sloan Kettering-Metastatic (MSK-MET) dataset and to identify key prognostic biomarkers through explainable artificial intelligence techniques. MethodsWe performed a retrospective analysis of the MSK-MET cohort, comprising 25,775 patients across 27 tumor types. After data cleaning and balancing, 20,338 patients were included. Overall survival was defined as deceased versus living at last follow-up. Five classifiers (extreme gradient boosting [XGBoost], logistic regression, random forest, decision tree, and naive Bayes) were trained using an 80/20 stratified split and optimized via grid search with 5-fold cross-validation. Model performance was assessed using accuracy, area under the curve (AUC), precision, recall, and F1-score. Model explainability was achieved using Shapley additive explanations (SHAP). Survival analyses included Kaplan-Meier estimates, Cox proportional hazards models, and an XGBoost-Cox model for time-to-event prediction. The positive predictive value and negative predictive value were calculated at the Youden index–optimal threshold. ResultsXGBoost achieved the highest performance (accuracy=0.74; AUC=0.82), outperforming other classifiers. In survival analyses, the XGBoost-Cox model with a concordance index (C-index) of 0.70 exceeded the traditional Cox model (C-index=0.66). SHAP analysis and Cox models consistently identified metastatic site count, tumor mutational burden, fraction of genome altered, and the presence of distant liver and bone metastases as among the strongest prognostic factors, a pattern that held at both the pan-cancer level and recurrently across cancer-specific models. At the cancer-specific level, performance varied; prostate cancer achieved the highest predictive accuracy (AUC=0.88), while pancreatic cancer was notably more challenging (AUC=0.68). Kaplan-Meier analyses demonstrated marked survival separation between patients with and without metastases (80-month survival: approximately 0.80 vs 0.30). At the Youden-optimal threshold, positive predictive value and negative predictive value were approximately 70% and 80%, respectively, supporting clinical use for risk stratification. ConclusionsExplainable ML models, particularly XGBoost combined with SHAP, can strongly predict survivability in metastatic cancers while highlighting clinically meaningful features. These findings support the use of ML-based tools for patient counseling, treatment planning, and integration into precision oncology workflows. Future work should include external validation on independent cohorts, integration with electronic health records via Fast Healthcare Interoperability Resources–based dashboards, and prospective clinician-in-the-loop evaluation to assess real-world use.more » « less
