Development and validation of a nutrition-related genetic–clinical–radiological nomogram associated with behavioral and psychological symptoms in Alzheimer’s disease

Abstract

Abstract Background: Few evidence is available in the early prediction models of behavioral and psychological symptoms of dementia (BPSD) in Alzheimer’s disease (AD). This study aimed to develop and validate a novel genetic–clinical–radiological nomogram for evaluating BPSD in patients with AD and explore its underlying nutritional mechanism. Methods: This retrospective study included 165 patients with AD from the Chinese Imaging, Biomarkers, and Lifestyle (CIBL) cohort between June 1, 2021, and March 31, 2022. Data on demoimagedatas, neuropsychological assessments, single-nucleotide polymorphisms of AD risk genes, and regional brain volumes were collected. A multivariate logistic regression model identified BPSD-associated factors, for subsequently constructing a diagnostic nomogram. This nomogram was internally validated through 1000-bootstrap resampling and externally validated using a time-series split based on the CIBL cohort data between June 1, 2022, and February 1, 2023. Area under receiver operating characteristic (ROC) curves, calibration curves, and decision curve analysis (DCA) were used to assess the discrimination, calibration, and clinical applicability of the nomogram. Results: Factors independently associated with BPSD were: CETP rs1800775 (odds ratio [OR] = 4.137, 95% confidence interval [CI]: 1.276–13.415, P = 0.018), decreased Mini Nutritional Assessment score (OR = 0.187, 95% CI: 0.086–0.405, P <0.001), increased caregiver burden inventory score (OR = 8.993, 95% CI: 3.830–21.119, P <0.001), and decreased brain stem volume (OR = 0.006, 95% CI: 0.001–0.191, P = 0.004). These variables were incorporated into the nomogram. The area under the ROC curve was 0.925 (95% CI: 0.884–0.967, P <0.001) in the internal validation and 0.791 (95% CI: 0.686–0.895, P <0.001) in the external validation. The calibration plots showed favorable consistency between the prediction of nomogram and actual observations, and the DCA showed that the model was clinically useful in both validations. Conclusion: A novel nomogram was established and validated based on lipid metabolism-related genes, nutritional status, and brain stem volumes, which may allow patients with AD to benefit from early triage and more intensive monitoring of BPSD. Registration: Chictr.org.cn, ChiCTR2100049131.