Early detection of slow progressing, long-term diseases such as Alzheimer’s (AD) is preferred among all who are involved – patients, caregivers, clinicians, and physicians alike. The authors involved in this study have a rich publication history of work with Alzheimer’s Disease Neuroimaging Initiative (ADNI) data acquisition and analysis that has provided the field with important contributions that have helped us to better understand this disease’s detection, progression, and recent strides towards treatment and (hopefully) prevention. 

The “gold standard” for AD diagnosis is detection and measurement of the peptide amyloid-beta (Aβ), specifically Aβ42 and Aβ40, which differ by two amino acids in length. Faulty clearance of the extracellular soluble form of Aβ42 is thought to lead to aggregation, eventually forming plaques in the brain; this is thought to be the first of three phases of AD pathology (A/T/N). The second phase is intracellular aggregation of phosphorylated tau (p-tau) tangles (T), followed by neuronal degeneration or atrophy (N). As pathology progresses over time, Ab42 levels measurably decrease, and p-tau increases; the time span can vary among patients, although a lucky few may present as Aβ(+) but show no signs of cognitive impairment.

Aβ-PET imaging or measurement of cerebrospinal fluid (CSF) Aβ42 levels are used to distinguish Aβ positive (Aβ(+)) from Aβ(-) individuals - placing patients either on the AD continuum or not, respectively.  Both methods are highly accurate, but the former method is costly and invasive, while the latter method can have painful side-effects. 

The use of plasma biomarkers has been gaining momentum in this research community because access is minimally invasive (simple blood draw) and low-cost. Additionally, detection methods have proven to be relatively sensitive, accurate, and correlate with imaging and CSF data. This study combined multidisciplinary data from multisite acquisitions to test how well specific plasma biomarkers (Aβ42/Aβ40, p-tau-181, NfL) detect Ab-positivity among ADNI participants who were diagnosed as cognitively normal/unimpaired (CU) or in early phase of pathology (mild cognitive impairment, MCI). Using statistical modeling, they were able to confirm that plasma ratio of Aβ42/Aβ40 is a robust biomarker for detection of Aβ(+), and was better than p-tau-181 or NfL. Further, they were able to increase accuracy by adding relevant variables into the model from clinical information (e.g., demographics, APOE carrier status, global cognition) and imaging (MRI-score calculated by deep learning).

 By combining information from clinical, imaging, and plasma within the same research cohort to model Aβ-positivity, these researchers not only confirmed that Aβ42/Aβ40 from plasma is a robust biomarker of brain Aβ-positivity, but also made its predictive capability stronger. Using this approach for early detection and placement on the AD continuum may help to ease cost and increase participation for future studies.

This paper was published in Brain Communications. Additional authors include Dallas Veitch (UCSF); Paul Aisen (USC); Clifford Jack Jr (Mayo Clinic); William Jagust (UCB); Ronald Petersen (Mayo Clinic); Andrew Saykin (Indiana University); James Bollinger, Vitaliy Ovod, Kwasi Mawuenyega, and Randall Bateman (Washington University), Leslie Shaw and John Trojanowski (University of Pennsylvania); Kaj Blennow, Henrik Zetterberg (University of Gothenburg, Sweden; UCL Institute, London), and Michael Weiner (UCSF).