The sources focus on a study aimed at evaluating the accuracy of tau PET in predicting the progression of MCI to dementia, comparing its performance to Aβ PET and MRI. Researchers studied 448 participants with MCI from various centers in South Korea, Sweden, the US, and Switzerland, with a mean follow-up of two years. The study included a discovery cohort of 331 participants and a validation cohort of 117 participants. The researchers assessed the prognostic value of tau PET, Aβ PET, and MRI using quantitative thresholds and visual readings. The primary outcomes were the progression from MCI to all-cause dementia (regardless of suspected etiology) or to AD dementia (AD as the suspected etiology).

The results of the study indicate that tau PET was the most accurate standalone marker in predicting the progression to dementia among individuals with MCI. Notably, only the model with tau PET was superior in predicting all-cause dementia compared to a base model considering age, sex, education, and MMSE score. This finding was further validated in the independent cohort. For predicting AD dementia specifically, tau PET (both quantitative and visual reads) and Aβ PET (Centiloids) showed improved prediction compared to the base model in the discovery cohort. However, only the improved predictive power of tau PET (quantitative and visual reads) was replicated in the validation cohort.

In both the discovery and validation cohorts, LASSO logistic regression analyses revealed that the optimal model for predicting all-cause dementia included tau PET and MRI measures. Specifically, the optimal models included predictors for temporal meta-ROI tau, AD signature cortical thickness, and MTA visual read in the discovery cohort, and temporal meta-ROI tau, tau PET visual read, and AD signature cortical thickness in the validation cohort. The study also found that the combination of positive biomarkers (Aβ PET, tau PET, and MRI) was associated with a higher risk of dementia progression than any single positive biomarker. Notably, the A+T+N+ biomarker profile (positive for all three markers) carried the highest risk for AD dementia progression.

Sensitivity analyses revealed that tau PET performed better in younger individuals and that all neuroimaging markers performed better in APOEε4 non-carriers in the discovery cohort. However, these results were not replicated in the validation cohort, suggesting further investigation is necessary. Overall, the study suggests that tau PET holds significant promise as a prognostic marker for dementia progression in individuals with MCI, highlighting its potential clinical utility in predicting disease course and informing treatment strategies.