Although the underlying neuropathology is different, Alzheimer's disease (AD) and frontotemporal lobe degeneration (FTLD) may have overlapping clinical manifestations, and the differential diagnosis is difficult, often requiring biopsy for pathological diagnosis or lumbar puncture for cerebrospinal fluid examination. For this reason, Corey T. McMillan and others from the Department of Neurology at the University of Pennsylvania in the United States conducted a study to try to use MRI instead of cerebrospinal fluid examination to identify neurodegenerative diseases. The results of the study were published online on December 27, 2012 in the journal Neurology. . The results of the study found that MRI can replace cerebrospinal fluid biomarkers as a non-invasive procedure for screening the pathological mechanisms of AD and FTLD.
Researchers included 185 patients with AD or FTLD manifestations who were clinically diagnosed with neurodegenerative diseases. They underwent lumbar puncture or volumetric MRI. In the subgroup, 32 patients were genetically or biopsy confirmed as AD or FTLD. We use the monovalent decomposition technique to decompose the MRI volume, and use linear review and cross-validation to predict the ratio (tt/Aβ) of CSF total tau (tt) and β amyloid (Aβ1-42) in patients with AD and FTLD. Subsequently, four convergent resources (including neuroanatomical visualization and subgroup classification of genetic or biopsy-proven AD or FTLD) were used to evaluate the accuracy of MRI-based prediction of tt/Aβ.
The results of the study show that: regression analysis shows that MRI predicted tt/Aβ is highly correlated with the real cerebrospinal fluid tt/Aβ. In each group, the predicted cerebrospinal fluid tt/Aβ and the real cerebrospinal fluid tt/Aβ overlap extensively in neuroanatomy: low tt/Aβ in line with FTLD is related to the median area of the forehead and abdomen; high tt/Aβ in line with AD is related to Related to the posterior cortex. For patients with known pathology and known cerebrospinal fluid tt/Aβ levels for clinical diagnosis, the accuracy of MRI predicting tt/Aβ is 75%.