Dataset Information

Screening and predicting progression from high-risk mild cognitive impairment to Alzheimer's disease.

ABSTRACT: Individuals with mild cognitive impairment (MCI) are clinically heterogeneous, with different risks of progression to Alzheimer's disease. Regular follow-up and examination may be time-consuming and costly, especially for MRI and PET. Therefore, it is necessary to identify a more precise MRI population. In this study, a two-stage screening frame was proposed for evaluating the predictive utility of additional MRI measurements among high-risk MCI subjects. In the first stage, the K-means cluster was performed for trajectory-template based on two clinical assessments. In the second stage, high-risk individuals were filtered out and imputed into prognosis models with varying strategies. As a result, the ADAS-13 was more sensitive for filtering out high-risk individuals among patients with MCI. The optimal model included a change rate of clinical assessments and three neuroimaging measurements and was significantly associated with a net reclassification improvement (NRI) of 0.246 (95% CI 0.021, 0.848) and integrated discrimination improvement (IDI) of 0.090 (95% CI - 0.062, 0.170). The ADAS-13 longitudinal models had the best discrimination performance (Optimism-corrected concordance index = 0.830), as validated by the bootstrap method. Considering the limited medical and financial resources, our findings recommend follow-up MRI examination 1 year after identification for high-risk individuals, while regular clinical assessments for low-risk individuals.

SUBMITTER: Ge XY

PROVIDER: S-EPMC8413294 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:IntroductionHeterogeneity in risk of conversion to Alzheimer's disease (AD) among individuals with mild cognitive impairment (MCI) is well known. Novel statistical methods that are based on partially ordered set (poset) models can be used to create models that provide detailed and accurate information about performance with specific cognitive functions. This approach allows for the study of direct links between specific cognitive functions and risk of conversion to AD from MCI. It also allows for further delineation of multi-domain amnestic MCI, in relation to specific non-amnestic cognitive deficits, and the modeling of a range of episodic memory functioning levels.MethodsFrom the Alzheimer's Disease Neuroimaging Initiative (ADNI) study, conversion at 24 months of 268 MCI subjects was analyzed. It was found that 101 of those subjects (37.7%) converted to AD within that time frame. Poset models were then used to classify cognitive performance for MCI subjects. Respective observed conversion rates to AD were calculated for various cognitive subgroups, and by APOE e4 allele status. These rates were then compared across subgroups.ResultsThe observed conversion rate for MCI subjects with a relatively lower functioning with a high level of episodic memory at baseline was 61.2%. In MCI subjects who additionally also had relatively lower perceptual motor speed functioning and at least one APOE e4 allele, the conversion rate was 84.2%. In contrast, the observed conversion rate was 9.8% for MCI subjects with a relatively higher episodic memory functioning level and no APOE e4 allele. Relatively lower functioning with cognitive flexibility and perceptual motor speed by itself also appears to be associated with higher conversion rates.ConclusionsAmong MCI subjects, specific baseline cognitive profiles that were derived through poset modeling methods, are clearly associated with differential rates of conversion to AD. More precise delineation of MCI by such cognitive functioning profiles, including notions such as multidomain amnestic MCI, can help in gaining further insight into how heterogeneity arises in outcomes. Poset-based modeling methods may be useful for providing more precise classification of cognitive subgroups among MCI for imaging and genetics studies, and for developing more efficient and focused cognitive test batteries.

Project description:Random Survival Forests (RSF) has recently showed better performance than statistical survival methods as Cox proportional hazard (CPH) in predicting conversion risk from mild cognitive impairment (MCI) to Alzheimer's disease (AD). However, RSF application in real-world clinical setting is still limited due to its black-box nature.For this reason, we aimed at providing a comprehensive study of RSF explainability with SHapley Additive exPlanations (SHAP) on biomarkers of stable and progressive patients (sMCI and pMCI) from Alzheimer's Disease Neuroimaging Initiative. We evaluated three global explanations-RSF feature importance, permutation importance and SHAP importance-and we quantitatively compared them with Rank-Biased Overlap (RBO). Moreover, we assessed whether multicollinearity among variables may perturb SHAP outcome. Lastly, we stratified pMCI test patients in high, medium and low risk grade, to investigate individual SHAP explanation of one pMCI patient per risk group.We confirmed that RSF had higher accuracy (0.890) than CPH (0.819), and its stability and robustness was demonstrated by high overlap (RBO > 90%) between feature rankings within first eight features. SHAP local explanations with and without correlated variables had no substantial difference, showing that multicollinearity did not alter the model. FDG, ABETA42 and HCI were the first important features in global explanations, with the highest contribution also in local explanation. FAQ, mPACCdigit, mPACCtrailsB and RAVLT immediate had the highest influence among all clinical and neuropsychological assessments in increasing progression risk, as particularly evident in pMCI patients' individual explanation. In conclusion, our findings suggest that RSF represents a useful tool to support clinicians in estimating conversion-to-AD risk and that SHAP explainer boosts its clinical utility with intelligible and interpretable individual outcomes that highlights key features associated with AD prognosis.

Dataset Information

Screening and predicting progression from high-risk mild cognitive impairment to Alzheimer's disease.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets