Blood tests for Alzheimer's disease: The impact of disease prevalence on test performance

Abstract

Regulatory agencies, including the US Food and Drug Administration, have recently approved Alzheimer's disease (AD) therapies targeting amyloid-beta (Aβ) pathology. This has placed a spotlight on the need for accurate and accessible diagnostic tools for AD pathology. Currently, AD cerebrospinal fluid (CSF) biomarkers and amyloid imaging have regulatory approval in many countries and are routinely used in individuals presenting with cognitive decline. Emerging blood-based biomarkers have the potential to be more widely accessible than these existing diagnostic tools, but are they ready for widespread uptake in medical care? Like others in the field, we are excited by the technological progress that has enabled the measurement of AD biomarkers in plasma. In particular, plasma phosphorylated-tau (pTau) and Aβ42/40 assays—predominantly evaluated in retrospective studies—have shown good diagnostic accuracy for AD.1, 2 However, our collective excitement for this new era in AD biomarkers has been tempered by efforts to quickly and broadly push these tests into clinical use. This has led to testing, in some cases, being marketed in a manner that lacks transparency related to the assay's performance (e.g., performance data not readily accessible to consumers) and/or targeting populations where diagnostic performance has not been adequately tested (e.g., healthy young adults). Increased transparency and information on diagnostic performance data from laboratories and manufacturers offering and developing AD blood tests are needed. Specifically, reporting of diagnostic performance data in the context of disease prevalence in the population in which the test is (to be) marketed/used. An assay's clinical sensitivity and specificity are helpful in describing the performance of a test; however, this description is incomplete without incorporating disease prevalence in the intended use population. Figure 1 demonstrates the dramatically changing diagnostic performance of Aβ42/40 and pTau blood tests with changes in AD prevalence. This change in performance has significant consequences for medical care. Substantially different medical follow-up is anticipated for a blood test where a positive result is wrong 40% of the time (e.g., a modestly performing Aβ42/40 assay used in a low prevalence setting) versus one that is wrong only 5% of the time (e.g., a high performing pTau assay used in a high prevalence setting) (Figure 1). The fervent push for testing to be offered in lower prevalence populations—such as to healthy persons over the age of 18 via the direct-to-consumer model or the more general call for access at the primary care level—needs to be counter-balanced by the need to first collect (and report) diagnostic performance data in these populations. As a community, we also need to better understand the impact of testing on patients in populations that have not been, generally, part of the fields focus of study. With the introduction of AD CSF biomarkers into clinical practice, there was international consensus and clear guidance regarding appropriate use scenarios for testing.3 This guidance is just starting to be developed for blood-based testing, and consensus has not yet been reached.4 For AD blood biomarkers, preliminary recommendations advise implementation only within specialized memory clinics and only for patients with cognitive symptoms.5 Implementation is not recommended in primary care, or as a stand-alone diagnostic tool.5 AD blood tests have the potential to deliver greater equity into the healthcare system via improved ease of access. From the clinical implementation of AD CSF and amyloid imaging testing—only for specialists in dementia care and individuals with cognitive symptoms—we have learned that AD biomarker testing can be leveraged to improve medical care such as optimizing pharmacotherapy decisions, and reducing the number of diagnostic procedures needed to arrive at a diagnosis.6, 7 For individuals living with AD and their family members, CSF testing is valued for the greater diagnostic certainty it brings, and because it empowers individuals with knowledge about their brain health, helping them prepare for the future.8 We cannot, however, assume that benefits from CSF testing and amyloid imaging will directly translate to the new settings where blood-based testing may be deployed. As we work toward incorporating blood tests for AD into clinical practice, the medical community, including clinical specialties, clinical laboratories, general practitioners, and industry, need to work together to fill in current knowledge gaps, and deliver transparency on test performance and utility to consumers. The authors have nothing to report. No funding was received for this work. M.L.D. reports consulting for Siemens and Eisai, and consulting and lecturing fees from Roche. A.A.S. reports advisory board participation for Fujirebio Diagnostics, Roche Diagnostics and Siemens Healthineers; honoraria for lectures from Roche Diagnostics. M.M.B. reports receiving travel support and lecture fees from Roche Diagnostics, licensing income from technology licensed by Washington University to C2N diagnostics, and is a co-inventor on the following patents related to Alzheimer's disease testing: 018941/US, PCT/US2022/015998. Author disclosures are available in the Supporting information. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.

Topics & Concepts