Precision diagnosis of cognitive impairment due to Alzheimer's disease for therapeutic interventions

Abstract

With the advent of anti‐amyloid monoclonal antibody (AAMA) therapy, precision diagnosis is necessary for identifying appropriate patients with cognitive disorders due to Alzheimer's disease. Therapy with AAMAs requires that candidates be diagnosed with mild cognitive impairment or mild dementia, have elevated brain amyloid‐β, have good physical, psychiatric, and medical health, and lack clinical or biomarker evidence of potentially impactful non‐Alzheimer brain disorders. The first three diagnostic activities are the core of the Clinical Practice Guidelines, but the last element of the precision diagnosis requires new decision‐making tools for recognizing multi‐etiology cognitive impairment. Within the context of shared decision‐making between clinician, patient, and family, proper diagnosis is essential. In addition to discussing the benefits and risks of AAMA therapy, the experienced clinician must empathetically assist in bridging the gap between expectations of benefit and the patient's overall diagnostic suitability for AAMA therapy.

Highlights

• In order to prescribe an anti‐amyloid monoclonal antibody (AAMA) to the right patients, those selected for treatment should be diagnosed with mild cognitive impairment or mild dementia, have elevated brain amyloid‐β (Aβ), and have good physical, psychiatric, or medical health.

• Persons with Alzheimer's biology as the primary etiology are the ideal AAMA treatment recipients. A novel activity necessary to optimize therapeutic response is to exclude persons with clinical or biomarker evidence of alternative contributory brain disorders.

• While mild clinical severity and elevated brain amyloid‐β are essential elements for selecting patients for AAMA treatment, clinical judgment is required to weigh the implications of more advanced disease severity, other medical co‐morbidities, and the presence of other contributory neuropathologies.

1. INTRODUCTION

In this issue of Alzheimer's & Dementia, a workgroup (of which I was a member) assembled by the Alzheimer's Association presents a comprehensive approach to the diagnosis of later‐life cognitive disorders and their underlying causes. The Clinical Practice Guidelines (CPG) ¹ , ² rest on an unhurried, skillful history, a neurological examination, and a thoughtful synthesis of clinical and biomarker evidence to formulate diagnoses. Anti‐amyloid monoclonal antibody (AAMA) therapy for the treatment of cognitive disorders due to Alzheimer's disease (AD) has become a reality, and precision diagnosis for later life cognitive disorders has never been more important.

Therapy with lecanemab ³ , ⁴ and donanemab (expected) ⁵ requires that candidates have mild cognitive impairment (MCI) or mild dementia and have elevated brain amyloid‐β (Aβ). In addition, in order to improve the likelihood of achieving benefits from therapy, treatment candidates should have good physical, psychiatric, and medical health. These diagnostic considerations are the core of the CPG. ¹ , ² In addition, precision diagnosis requires attention to the possibility that individuals meeting the required criteria for AAMA treatment may also have co‐existing non‐AD pathologies that will be unresponsive to the AD‐directed treatment.

2. MCI OR MILD DEMENTIA COMPATIBLE WITH AD AS AN ETIOLOGY

Precision diagnosis is needed to identify patients for AAMA therapy whose cognitive impairment is in the appropriate severity range. AAMA therapy is currently intended only for persons who have objective evidence of cognitive impairment ⁴ , ⁶ in the mild cognitive impairment to mild dementia range. Neither lecanemab nor donanemab are indicated for persons outside of this range. The approach to a comprehensive medical and neurological history that includes an interview of a knowledgeable informant, a neurological examination, a cognitive assessment including neuropsychological testing are key recommendations in the CPG. ¹ , ² Clinical judgment is required to integrate information from the patient examination, the informant's history, and the neuropsychologist's impressions to arrive at an estimate of severity independent of biomarker evidence.

3. WEIGHING THE BURDEN OF MEDICAL, PSYCHIATRIC, AND OTHER COMORBID CONDITIONS

Precise diagnosis is also necessary to appreciate the threats to therapeutic benefits posed by co‐morbid conditions in patients who are being considered for AAMA therapy. Knowledge of the multi‐dimensional health status of patients, an overarching theme of the CPG, ¹ , ² is critical to judging the likelihood that a patient will be able to follow the rigorous demands of AAMA treatment as well as for estimating the likelihood of achieving benefits from the treatment. Substantial impairments in vision, hearing, or mobility may themselves compromise the quality of life in a way that would overshadow any cognitive delays. Cardiac, neoplastic, or other non‐neurological diseases are common in elderly cognitively impaired cohorts. ⁷ , ⁸ Systemic co‐morbid illnesses could have a major bearing on both quality of life and longevity that would neutralize the expected benefits of therapy. The presence of clinically important anxiety, depression, or psychotic symptoms might also attenuate therapeutic benefits. Sensori‐motor, medical, and neuropsychiatric comorbidities do not automatically rule out suitability for AAMA therapy, but their presence should be part of any discussion about expectations of benefits from AAMA treatment.

4. BIOMARKER PROOF OF ELEVATED BRAIN Ab AND AD‐RELATED REGIONAL NEURODEGENERATION

Treating patients with an AAMA who do not have truly elevated brain Aβ is not acceptable. The laboratory determination of elevated brain Aβ to be employed in selecting patients for AAMA therapy should ideally have a zero false positive rate. There are potentially four approaches to establishing elevated brain Aβ. With rapid advances in plasma biomarkers, there will be pressure to use a marker such as plasma ptau‐217 as a stand‐alone test, ⁹ but in one study, the false‐positive rate was 11%, ¹⁰ suggesting that for now, plasma biomarkers alone are not indicated for diagnosing elevated brain Aβ in the context of consideration for AAMA therapy. The second approach is cerebrospinal fluid (CSF) biomarker determinations that proxy elevated brain Aβ. This approach is acceptable and comparable to positron emission tomography (PET) from a diagnostic perspective, ¹¹ but CSF studies occasionally produce false positive diagnoses compared to amyloid PET. Further, if CSF studies alone are used to determine elevated brain Aβ levels, they will be of limited value as a baseline for subsequent determinations of therapeutic effectiveness. The third approach, used in the lecanemab clinical trial ³ was a visual read of amyloid PET for “not elevated” versus “elevated” Aβ. It may be difficult to distinguish between marginal and substantial burdens of Aβ with visual reading alone. Therefore, quantitative amyloid PET should be the preferred precision‐medicine approach. The quantitative burden of brain Aβ contributes to an understanding of the duration of the AD neurodegenerative process ¹² and makes a prediction of treatment response to lecanemab ¹³ and donanemab. ⁵ Brain Aβ values in the range near the minimum threshold level in a person with mild dementia (i.e., when the burden of Aβ is too modest for the degree of cognitive impairment) raise a concern about alternative neurodegenerative or cerebrovascular contributory etiologies. In contrast, higher brain Aβ levels are more convincing for a primary role in AD pathology.

If available, [18]fluorodeoxyglucose (FDG) PET can make unique contributions to diagnosis in persons being considered for AAMA therapy, beyond that offered by clinical examinations, magnetic resonance imaging (MRI) scans, and amyloid PET. FDG PET provides information supporting the presence of AD pathology because the topography of AD‐related hypometabolism is relatively specific and sensitive. ¹⁴ , ¹⁵ , ¹⁶ The topography of FDG PET hypometabolism is an excellent indicator of the extent of neocortical AD neurodegeneration and can be paired with clinical diagnoses to judge (a) the likelihood that AD is the underlying dominant etiology, and (b) whether AD neurodegeneration is at a level where AAMA benefits could be expected. For example, in persons with mild dementia clinically who have minimal neocortical hypometabolism, the possibility of a non‐AD pathology becomes more likely. ¹⁷ In an Aβ‐positive patient with MCI or mild dementia with extensive temporal, parietal, and frontal hypometabolism on FDG PET, the clinical trial data ⁵ , ¹³ suggest that there is an increased likelihood of a blunted response to AAMA therapy.

5. WEIGHING THE BURDEN OF COMORBID NEURODEGENERATIVE AND CEREBROVASCULAR DISEASES

With the introduction of AAMA therapy, the possibility of multi‐etiologic contributions to a cognitive syndrome has ceased to be merely a hypothetical matter. In the elderly, multi‐etiology cognitive impairment is common ¹⁸ , ¹⁹ even in a person with elevated brain Aβ. Because non‐AD neuropathological entities may contribute additively to cognitive disability, ²⁰ candidates for AAMA therapy with evidence for non‐AD processes would be less likely to achieve benefits from an AAMA. The CPG outlines a general framework for investigating the etiology of cognitive impairment, ¹ , ² but interpreting multi‐etiology considerations and judging the primacy of AD pathology is a novel diagnostic challenge. Interpreting the possible role of non‐AD etiologies in a person with elevated brain Aβ being considered for AAMA therapy is an under‐appreciated consideration in diagnosis.

There are two questions to be addressed for estimating the likelihood of clinically impactful non‐AD etiologies in a person with elevated brain Aβ. First, how convincing is the evidence for AD as a primary cause: are the elevations in brain Aβ and the extent of neurodegeneration as measured by FDG PET or tau PET ¹⁷ in relation to the degree of cognitive impairment compelling? Second, how strong a case is there that clinical features as well as biomarkers suggest an alternative etiology? ²¹ Astute clinical judgment is at a premium here for weighing the likelihood of AD versus other etiologies in order to estimate the primacy of AD as the cause of cognitive impairment.

Frontotemporal degenerations with amnestic, language, or behavioral presentations may sometimes be mistakenly attributed to AD pathology, but typical frontotemporal degenerations rarely co‐occur with AD pathology. ²² Lewy body disease and limbic predominant TAR DNA‐binding protein 43 (TDP‐43) proteinopathy (LATE) ²³ , ²⁴ are the two degenerative conditions that most commonly co‐occur with AD pathology. With both disorders, their contribution may be suspected by positive evidence for Lewy Body disease or by inference from the pattern of multiple biomarkers in the case of LATE.

Clinical evidence for Lewy body disease (α‐synucleinopathy) ²⁴ might include signs and symptoms of parkinsonism, prominent well‐formed visual hallucinations, or rapid eye movement sleep behavior disorder. Knowledge of and sensitivity to these features is strongly related to clinical experience with the disorder and with being able to conduct an unhurried history and examination. Emerging FDG PET imaging features, ²⁵ skin biopsies, ²⁶ and CSF biomarkers ²⁷ may provide additional insights about the presence of α‐synuclein pathology when clinical features of DLB are equivocal.

The clinical diagnosis of LATE is a provisional one. ²³ LATE's common clinical appearance, that of an amnestic cognitive disorder with an indolent course, is almost indistinguishable from that caused by AD. On the other hand, the degree of hippocampal atrophy due to LATE often exceeds that seen in AD, and LATE also has a distinctive pattern of FDG hypometabolism that differentiates it from AD. ²⁸ Because of the non‐specificity of the clinical presentation of LATE and the lack of a fluid biomarker for TDP43, suspicion for LATE must rely on a pattern of not‐elevated or marginally‐elevated brain Aβ and lack of an AD topographic pattern of neurodegeneration (by FDG PET) or tauopathy (by tau PET). ²⁹

Atherosclerotic/arteriolosclerotic cerebrovascular disease (A/A‐CVD) is common in the elderly with cognitive impairment antemortem ³⁰ and post‐mortem. ³¹ In persons being considered for AAMA therapy, the extent of A/A‐CVD has both diagnostic and safety ramifications. The relatively uncommon features of typical stroke‐like onset of cognitive impairment, multiple infarcts, or territorial infarcts are contraindications to AAMA therapy on both diagnostic and safety grounds. The more common occurrence of the extensive burden of white matter hyperintensities, multiple (i.e., >2) lacunar infarcts, or single regional infarcts are not necessarily exclusionary but should raise questions about the vascular contribution to cognitive impairment in a person being considered for AAMA therapy.

As a safety matter apart from etiological considerations, diagnosis of A/A‐CVD is germane to AAMA therapy because extensive athero‐ or arteriolo‐sclerotic cerebrovascular disease may pose a hemorrhagic risk for the use of AAMA therapy. Similarly, evidence for cerebral amyloid angiopathy in the form of more than four cerebral microbleeds or superficial siderosis with gradient recalled echo MRI sequences is an exclusion criteria for AAMA therapy. ⁴

The presence of non‐AD disease processes does not necessarily negate the therapeutic value of AAMA therapy. As with the other comorbid conditions that occur in persons with cognitive impairment presumed due to AD, the evidence for other pathologies needs to be taken into consideration as part of a thorough assessment of appropriateness for treatment. Unfortunately, it may only be with future cumulative experience that the clinical relevance of non‐AD processes on treatment outcomes will be established.

6. ON A FINAL CONSENSUS RECOMMENDATION TO TREAT OR NOT TO TREAT

The diagnostic process described here is meant to identify patients most likely to achieve a benefit from AAMA therapy. When a patient meets all entry criteria, therapeutic activities can be initiated. However, the consequence of diagnostic exclusivity necessarily leads to a recommendation to forego AAMA therapy in many patients who would be eager to receive it. Some patients might be very unhappy and will undoubtedly plead to be treated despite safety concerns or the compelling threats to treatment effectiveness. On the one hand, it should be straightforward to justify a no‐treatment recommendation for a patient who lacks elevated brain Aβ or who has moderately severe dementia. In contrast, a recommendation to forgo AAMA therapy because of excess co‐morbidity may seem logical to a physician but deeply disappointing to patients and families who had embraced the benefits of AAMA therapy.

The CPG emphasis on shared decision‐making ¹ , ² recommends itself for an AAMA treatment‐diagnostic discussion. Instigation of AAMA therapy requires that the patient, family, and clinician have an understanding of how the therapy will alter the trajectory of the illness and what the consequences of side effects could be. Inherent in the expectations of benefits is the degree to which the patient's diagnosis enables the achievement of treatment goals. Given the small therapeutic advantage that the AAMAs offer, limiting treatment to persons with the fewest additional comorbid conditions and non‐AD neurological contributors to cognitive impairment would seem to offer the greatest chance to achieve those goals. The role of the experienced clinician is to describe, elucidate, and clarify the diagnostic inclusion features, alternative etiologic considerations, comorbidities, and general health. But the clinician is not there merely to recite a checklist. The experienced clinician must actively but empathetically assist in bridging the gap between expectations of benefit and the reality of the patient's overall suitability for AAMA therapy. Hopefully, in most situations, physicians, patients, and families will reach a treatment consensus.

CONFLICT OF INTEREST STATEMENT

Dr. David S. Knopman serves on a Data Safety Monitoring Board for the Dominantly Inherited Alzheimer Network Treatment Unit and a study of nicorandil for the treatment of hippocampal sclerosis of aging sponsored by the University of Kentucky. He was an investigator in Alzheimer clinical trials sponsored by Biogen, Lilly Pharmaceuticals, and the University of Southern California, and is currently an investigator in a trial in frontotemporal degeneration with Alector. He has served as a consultant for Roche, AriBio, Linus Health, Biovie, and Alzeca Biosciences but receives no personal compensation. He receives funding from the NIH. Author disclosures are available in the supporting information.

Supporting information

Supporting Information

Knopman DS. Precision diagnosis of cognitive impairment due to Alzheimer's disease for therapeutic interventions. Alzheimer's Dement. 2025;21:e14043. 10.1002/alz.14043