Imaging markers for Alzheimer disease: which vs how

Area of Science:

• Neurology
• Medical Imaging
• Biomarkers

Background:

• Revised Alzheimer disease (AD) diagnostic criteria emphasize imaging biomarkers for early detection.
• The accuracy of these biomarkers depends on both the specific marker used and the measurement method (metric).

Purpose of the Study:

• To evaluate the diagnostic accuracy of different imaging markers and metrics in distinguishing AD from healthy individuals.
• To assess the prognostic accuracy of these markers and metrics in predicting disease progression in mild cognitive impairment.

Main Methods:

• Utilized likelihood ratios (LR+ and LR-) to quantify diagnostic and prognostic accuracy.
• Compared the performance of various imaging markers (amyloid PET, FDG-PET, SPECT, MRI) and measurement metrics (visual, manual, automated).

Main Results:

• Marker performance varied, with diagnostic LR+ between 4.4-9.4 and LR- between 0.25-0.08.
• Metric-specific variations in accuracy were substantial, with LR+ ranging from 1 to 100 and LR- from 1.00 to 0.01.
• Measurement metrics accounted for equal or greater variance in accuracy compared to the markers themselves, particularly for ¹⁸F-FDG-PET and MRI.

Conclusions:

• The method of measurement is as critical as the imaging biomarker itself for accurate AD diagnosis and prognosis.
• Standardized operating procedures are essential for the reliable integration of imaging biomarkers into clinical practice and drug development for Alzheimer disease.