Practical Recommendations for Monitoring Response to Amyloid-Targeting Treatments for Early Alzheimer Disease

Clinical trials of amyloid-targeting treatments (ATTs) used a variety of primary and secondary endpoints to assess efficacy of treatment in slowing Alzheimer disease (AD) progression.^1,2 These assessments included various clinical assessments as well as imaging and fluid biomarkers of amyloid, tau, and neurodegeneration. Compared with placebo, individuals receiving clinically approved ATTs demonstrated modest slowing on cognitive assessments, robust reduction in amyloid plaque burden measured using amyloid positron emission tomography (PET) tests, and decreases in cerebrospinal fluid (CSF) and blood-based biomarkers of amyloid pathology.^1,2 Markers of tau pathology and neurodegeneration have shown mixed results. Although each of these measures provides valuable information, the individual variability in these measurements and practicality of implementation must be considered when applying them to clinical practice. This article includes an overview of the utility of clinical assessment, amyloid PET scanning, fluid-based biomarkers, and tau and neurodegeneration markers in clinical practice and provides guidance for their use in monitoring the efficacy of ATTs.

Clinical Assessment
Research measures of cognitive and functional status (Table), such as the Clinical Dementia Rating scale, Integrated Alzheimer’s Disease Rating Scale, and Alzheimer’s Disease Assessment Scale–Cognitive subscale, were used in ATT trials, but these scales are less practical clinically due to their prolonged time of administration and level of expertise required to administer and score them appropriately.^3-6 A major value of both the Clinical Dementia Rating and Integrated Alzheimer’s Disease Rating Scale is the incorporation of both cognitive performance and functional status in day-to-day life, providing a more holistic assessment of clinical severity and progression. Therefore, a combination of brief cognitive tests and functional questionnaires are likely to provide similar information in a more practical manner. The Mini-Mental State Examination (MMSE) and Montreal Cognitive Assessment (MoCA) are the most broadly used and validated assessments of global cognition.^7,8 Functional assessments, such as the Functional Assessment Questionnaire (FAQ) and Dementia Severity Rating Scale (DSRS), can be performed in clinic or as part of preclinical questionnaires.^9,10 We recommend completing at least 1 of these cognitive (MMSE or MoCA) and functional (FAQ or DSRS) assessments at baseline and every 6 months during ATT treatment. More detailed assessment of cognition, including domain-specific measures, is appropriate when available, and other assessments of global cognition or functional impairment may also be considered.

Collecting this information is valuable for monitoring clinical progression, but these measures are difficult to interpret as a measure of individual treatment response due to significant heterogeneity in disease trajectory in the natural history of disease.^11,12 On average, scores worsen by 2 to 3 points annually on the MoCA and MMSE, by ~4.5 points annually on the DSRS, and by 2.6 points annually on the FRS.^11-13 However, it is difficult to know whether rates slower than this seen in individuals taking ATT represent a treatment effect or a naturally slower course of disease. For individuals with previous longitudinal assessments, slowing decline compared with previous rates may provide some measure of response to ATT. Conversely, some individuals taking ATT may show persistent decline and progress to moderate dementia. In these cases, we recommend re-evaluating the benefit of continued treatment with patients and caregivers with an emphasis on whether maintaining their functional status outweighs the risks and burdens of continued treatment. Rates based on a single assessment should be interpreted with caution, because there are significant test-to-test fluctuations in MMSE and MoCA.

Amyloid PET
Amyloid PET was used to determine inclusion status and monitor treatment response to ATTs in clinical trials. Clearance of amyloid plaque measured by PET was used as a stopping criterion for donanemab (Kisunla; Eli Lilly, Indianapolis, IN).¹ In phase 3 trials (TRAILBLAZER-ALZ2 [NCT04437511] and Clarity AD [NCT03887455]), 68% of individuals receiving lecanemab (Leqembi; Eisai, Nutley, NJ) and 75% of individuals receiving donanemab had transitioned from amyloid-β positivity (Aβ+) at baseline to amyloid-β negativity (Aβ−) after 18 months of ATT.^1,2 This transition is termed “treatment-related amyloid clearance” (TRAC). In both trials amyloid PET scanning was performed every 6 months and TRAC increased with duration of therapy. We recommend amyloid PET scanning as the primary measure of ATT response and target engagement with amyloid plaques for ATT.

Amyloid status can be determined by visual read performed by a trained radiologist, by quantification of tracer binding, or by combining visual and quantitative approaches.^14-17 In clinical trials, quantitative measures were used to define both Aβ+ and TRAC using the Centiloid (CL) scale, but the specific cutoffs for Aβ+ and TRAC differed by trial.^1,2 The CL scale provides a quantitative assessment of amyloid burden that is tracer-independent, and there are multiple tools that have received Food and Drug Administration and Conformité Européene approval for measurement of CL.^16,18 Given the availability of these tools, a combined visual read and CL quantification approach for monitoring ATT response is recommended (Figure). Cut points for amyloid status vary, but it is generally accepted that CL >30 is consistent with Aβ+ and CL <10 is consistent with Aβ−, with values in between these cutoffs reflecting an intermediate zone.¹⁸ In general, it is recommended to perform visual read without CL quantification, followed by CL quantification, and then assessment of concordance or discordance. In cases of discordance between visual read and CL quantification, specific causes of discordance should be evaluated (eg, true borderline, motion, registration errors, atypical binding pattern). The final assessment of amyloid status should be made by combining visual and quantitative data.

Figure. Recommended timeline to obtain amyloid positron emission tomography for donanemab and lecanemab. Demonstration of TRAC with a positive scan at baseline and negative scan at follow-up. Note the lack of gray–white differentiation at baseline, indicating a positive scan, and white matter only binding at follow-up, indicating a negative scan (top). Recommended combined visual and quantitative approach for assessing amyloid status (bottom). Visual read is performed first, followed by Centiloid quantification. When discordant (bottom row), the image should be re-reviewed to assess for a cause of discordance (registration error, abnormal binding pattern, true borderline case), and additional opinions can be sought when needed. Note that the recommended Centiloid cutoffs are >30 Aβ+ and <10 Aβ−, with values in between considered intermediate.

Appropriate timing of amyloid PET tests to evaluate treatment efficacy should vary by ATT and may also vary depending on patient-specific factors. For lecanemab, amyloid PET scanning should be performed after ~18 months of treatment, which may aid in decisions around maintenance therapy based on treatment response. In contrast, amyloid PET scanning for donanemab should be performed after ~12 months of treatment for donanemab, as 66% of clinical trial participants achieved TRAC by this point and could stop therapy.¹ In those who do not meet criteria for TRAC at 12 months, repeat amyloid PET scanning should be repeated at 18 months to assess treatment efficacy. There are also data which suggest that earlier amyloid PET scanning after 6 months of treatment with donanemab may be reasonable in individuals with baseline CL <60, but further study is needed to determine appropriate timing of PET scanning in this group. Although the primary utility for follow-up amyloid PET scanning is to assess for TRAC, relative stability in amyloid burden after 18 months is likely to suggest treatment nonresponse. As such, additional ATT with the same agent is unlikely to be beneficial. Patient-specific factors may arise during treatment (eg, new atrial fibrillation, malignancy, or immune condition; worsening clinical functional status; change in caregiver circumstances) that warrant amyloid PET scanning outside of these prescribed time points, particularly when other conditions arise that may change the risk–benefit ratio of ATT.

Other Measures of Amyloid Burden
In addition to amyloid PET scanning, CSF and blood-based biomarkers can be used to detect amyloid-β pathology. Each of these measures show strong correlation with amyloid PET scans in the natural history of disease, but data regarding the association between amyloid PET tests and fluid-based biomarkers following ATT are limited.^19-21 In clinical trials, fluid-based biomarkers of amyloid showed less abnormality after ATT compared with baseline and placebo.^1,2 However, these markers did not consistently become negative, and correlation with amyloid PET scans at the individual level has been weak.²² Therefore, CSF and blood-based biomarkers are not recommended to monitor ATT response. However, blood-based biomarkers represent a highly accessible and cost-effective method for assessing changes in AD pathology, and recent advances in robustness of measures of plasma p-tau217 and p-tau217/Aβ1-42 ratios may lead to improvement in the ability of these measures to assess treatment response.²⁰ Further research is needed, but change in plasma biomarkers may one day be used as a trigger for amyloid PET tests to assess for TRAC, thus reducing the number of required PET scans and treatment time for individuals receiving donanemab. 

Measures of Tau and Neurodegeneration
Whereas data suggest that ATT may slow tau accumulation, high inter-individual heterogeneity in regional tau patterns and accumulation, lack of consensus methods for measuring global tau burden, and relative lack of availability of tau PET clinically make clinical assessment of tau change with ATT unfeasible.^2,23,24 Markers of neurodegeneration also show inconsistent patterns following ATT, with fluid-based biomarkers of neurofilament light showing no change with treatment and MRI markers showing accelerated global atrophy with preservation of hippocampal volume.^1,2 The prevailing interpretation of these MRI findings is amyloid-β clearance–related pseudoatrophy driven by reduction in amyloid-β and associated inflammatory products leading to global brain volume loss that spares the hippocampus due to minimal amyloid pathology in this region.²⁵ However, given the lack of clarity around the clinical significance of these findings, MRI-based measures are not recommended for monitoring ATT efficacy.

Summary of Recommendations for Clinical Monitoring of AAT
We recommend monitoring response to ATT with a combination of cognitive tests, functional assessment, and amyloid PET scanning (Box). Global cognitive batteries, such as the MMSE or MoCA, paired with functional assessments, such as the DSRS or FAQ, should be obtained every 6 months to monitor clinical progression during ATT use. However, attributing changes in these measures to treatment response should be done cautiously. Amyloid PET scanning should be performed after 18 months of lecanemab and 12 months of donanemab treatment to assess for amyloid-lowering as the primary measure of ATT response. A combination of visual read and CL quantification should be used to assess the degree of amyloid-lowering and TRAC. In the future, plasma-based measures may aid in patient-specific, lower-cost monitoring of ATT, but these measures require further research and are not currently recommended for clinical monitoring. Tau and neurodegeneration measures are not recommended for monitoring ATT efficacy.