Prediction Model and Nomogram for Amyloid Positivity Using Clinical and MRI Features in Individuals With Subjective Cognitive Decline

Abstract

There is an urgent need for the precise prediction of cerebral amyloidosis using noninvasive and accessible indicators to facilitate the early diagnosis of individuals with the preclinical stage of Alzheimer's disease (AD). Two hundred and four individuals with subjective cognitive decline (SCD) were enrolled in this study. All subjects completed neuropsychological assessments and underwent 18F-florbetapir PET, structural MRI, and functional MRI. A total of 315 features were extracted from the MRI, demographics, and neuropsychological scales and selected using the least absolute shrinkage and selection operator (LASSO). The logistic regression (LR) model, based on machine learning, was trained to classify SCD as either β-amyloid (Aβ) positive or negative. A nomogram was established using a multivariate LR model to predict the risk of Aβ+. The performance of the prediction model and nomogram was assessed with area under the curve (AUC) and calibration. The final model was based on the right rostral anterior cingulate thickness, the grey matter volume of the right inferior temporal, the ReHo of the left posterior cingulate gyrus and right superior temporal gyrus, as well as MoCA-B and AVLT-R. In the training set, the model achieved a good AUC of 0.78 for predicting Aβ+, with an accuracy of 0.72. The validation of the model also yielded a favorable discriminatory ability with an AUC of 0.88 and an accuracy of 0.83. We have established and validated a model based on cognitive, sMRI, and fMRI data that exhibits adequate discrimination. This model has the potential to predict amyloid status in the SCD group and provide a noninvasive, cost-effective way that might facilitate early screening, clinical diagnosis, and drug clinical trials.

Keywords: Alzheimer's disease; logistic regression; machine learning; nomogram; subjective cognitive decline.

FIGURE 1

Features selection by least absolute shrinkage and selection operator (LASSO). (A) A coefficient profile plot was produced against the log lambda sequence. LASSO coefficients profiles (y‐axis) of the 12 features. (B) Ten‐fold cross‐validation for tuning parameter selection in the LASSO model. The partial likelihood deviance (binomial deviance) curve was plotted versus log(λ). To avoid overfitting, 1 standard error was selected.

FIGURE 2

Nomogram for the prediction model. The nomogram was developed in the training set. For instance, with the help of the nomogram model, we can see that a person with a ReHo_PCG_L value of 0.8, a ReHo_STG_R value of 0.3, an inferiortemporal_R value of 0.84, a rh_rostralanteriorcingulate value of 2.2, an AVLT‐R score of 17, and a MoCA‐B score of 16, might have approximately an 83% chance of Aβ+.

FIGURE 3

The calibration curves of the model in the training set (A) and test set (B), and ROC curves of the prediction model in training (C) and test (D) groups. A: Training set; B: Test set; C: Training set; D: Test set. Calibration curve: The black solid line above the x‐axis represents sample distribution. The dotted lines on the diagonal represent the perfect prediction of the ideal model, and the solid lines represent the performance of the training set and the test set. The closer the solid line is to the dotted line, the better the predictive effect. The x‐axis represents the predicted Aβ positive risk, and the y‐axis represents the actual Aβ status.