Skill Checkup: Advancing Memory Loss in a 70-Year-Old Woman

A 70-year-old woman in the US presents to the clinic accompanied by her daughter, who describes 18 months of worsening forgetfulness, including repeating questions, misplacing items, and missing recent conversations, along with new difficulty managing her finances. The patient still pays her own bills but now needs written reminders and considerably more time than in the past.

She remains independent in bathing, dressing, cooking, and driving. Her daughter denies low mood, anhedonia, or gait change. On examination, the patient is alert and fully oriented, with a Mini-Mental State Examination score of 26/30 and a Clinical Dementia Rating global score of 0.5.

Formal neuropsychological testing shows delayed-recall performance below expectations for her age and education, with attention, language, and visuospatial domains intact. Medication review, complete blood cell count, comprehensive metabolic panel, thyroid-stimulating hormone, and vitamin B 12 level are unremarkable.

Structural brain MRI shows mild medial temporal lobe atrophy without significant white-matter disease or infarcts. Amnestic mild cognitive impairment describes a clinical syndrome of objective, memory-predominant cognitive decline. This is corroborated by an informant history and impaired delayed-recall performance on formal testing, along with preserved independence in daily activities despite complex instrumental tasks (eg, paying bills) taking more time and support than in the past.

The underlying etiology remains undetermined and requires biomarker evaluation before it can be attributed specifically to an underlying cause such as Alzheimer's disease. Normal age-related memory change is excluded by the informant-noted decline and abnormal standardized delayed-recall testing, and depression is unlikely given no reported low mood or anhedonia, although it can present atypically and coexist with neurodegeneration in older adults.

Vascular cognitive impairment is unlikely given the insidious, amnestic-predominant course and absence of infarcts or significant white-matter disease on MRI , although mixed vascular-neurodegenerative pathology cannot be fully excluded. After evaluation by a dementia specialist, uncertainty about the underlying cause of the patient's amnestic syndrome persists, and she and her daughter wish to pursue further testing to clarify the etiology.

She has no known family history of early-onset dementia. Amyloid PET is appropriate for patients with mild cognitive impairment or dementia syndromes consistent with Alzheimer's disease pathology (amnestic presentation, onset at age ≥ 65 years) when etiology remains uncertain after specialist assessment and a positive or negative result would directly inform diagnosis and management, both of which apply here.

Emerging blood-based biomarkers can help triage which patients need this confirmatory testing, but amyloid PET or cerebrospinal fluid confirmation is required before treatment eligibility can be established. Repeating neuropsychological testing could document the trajectory of the patient's decline but would not establish amyloid status or amyloid-targeted therapy candidacy — the immediate question her syndrome raises.

FDG-PET can identify hypometabolic patterns characteristic of frontotemporal dementia or dementia with Lewy bodies but does not directly demonstrate amyloid pathology and is more useful for atypical or diagnostically ambiguous presentations than her typical late-onset amnestic picture.

APP/PSEN1/PSEN2 genetic testing is reserved for suspected autosomal-dominant disease, which typically presents before age 65 years with a multigenerational family history, not this patient's late-onset, apparently sporadic case. Amyloid PET, performed and interpreted by a qualified nuclear medicine physician, is visually read as positive, with tracer uptake extending into neocortical regions.

This is most pronounced in the precuneus, posterior cingulate, and frontal cortices, without a clear gray-white matter distinction. This biomarker result, combined with her amnestic clinical syndrome, establishes a diagnosis of mild cognitive impairment owing to early-stage Alzheimer's disease.

The patient and her daughter ask about disease-modifying treatment options. APOE epsilon 4 genotyping is recommended before initiating anti-amyloid antibody therapy because ARIA incidence differs substantially by genotype. In the CLARITY AD trial of lecanemab, ARIA occurred in 5.4% of APOE epsilon 4 noncarriers, 10.9% of heterozygotes, and 32.6% of homozygotes.

Donanemab trials show a comparable genotype-dependent gradient, with ARIA-edema/effusions (ARIA-E) risk roughly doubled in heterozygotes and more than fourfold higher in homozygotes relative to noncarriers. Current FDA labeling for both agents states that this testing should be performed to inform the risk discussion, although a patient may still be treated without it; in that case, epsilon 4 homozygosity and its associated higher ARIA risk cannot be identified.

MRI microhemorrhage count reflects a fixed eligibility threshold , not genotype-based risk. Plasma phosphorylated-tau217 is validated for amyloid detection and Alzheimer's disease diagnosis , but not ARIA prediction. Serum neurofilament light chain is a nonspecific injury marker uninvolved in ARIA or eligibility determination.

APOE genotyping shows that the patient is heterozygous for the epsilon 4 allele. After a discussion of the risks and benefits of donanemab and lecanemab, the patient and her daughter elect to proceed with lecanemab. The family expresses a preference for in-person monitoring during the highest-risk early treatment period, so she and her daughter elect to begin lecanemab via intravenous infusion.

Current FDA prescribing information for lecanemab (revised July 2026) recommends a baseline brain MRI before treatment, followed by MRI after 1, 2, 3, and 6 months of treatment. A 6-month interval and annual imaging are both inadequate and miss the highest-risk window. Serious ARIA-E clusters within the first 14 weeks of treatment, prompting the FDA's 2025 addition of a pre-third-infusion scan.

Symptom-triggered imaging alone is also inadequate because ARIA is usually asymptomatic even when radiographically apparent. The patient tolerates the first four infusions without incident. MRI obtained per protocol before the fifth infusion shows a new focus of fluid-attenuated inversion recovery (FLAIR) hyperintensity measuring 6 cm in greatest dimension, involving the right parietal cortex and subcortical white matter, without gyral swelling, sulcal effacement, or additional sites of involvement; there is no new microhemorrhage or superficial siderosis.

She reports no headache, confusion, visual change, or gait difficulty, and her neurologic examination is unchanged from baseline. Current FDA labeling classifies ARIA-E as moderate when FLAIR hyperintensity measures 5-10 cm at a single site, whereas continued dosing is permitted only for asymptomatic, mild ARIA-E confined to one site under 5 cm.

For an asymptomatic patient with moderate ARIA-E, labeling specifies suspending dosing until MRI demonstrates radiographic resolution, with a follow-up MRI considered at 2-4 months and resumption guided by clinical judgment, not automatic permanent discontinuation, which is reserved for more severe or recurrent events.

Dose reduction has no basis in the prescribing information, which structures ARIA-E management as a binary continue-or-suspend decision based on radiographic severity and symptom grade rather than partial dose adjustment. Dosing is suspended, and a follow-up MRI 10 weeks later shows resolution of the FLAIR hyperintensity, with the patient remaining asymptomatic throughout.

Lecanemab is resumed at the same dose per clinical judgment, and she continues the every-2-week infusion schedule with continued MRI surveillance per protocol. At reassessment 6 months later, treatment continuation is reviewed using serial cognitive, functional, and safety assessments.

Editor's Note: Skill Checkups are wholly fictional or fictionalized clinical scenarios intended to provide evidence-based educational takeaways. Additionally, this article was created using several editorial tools, including generative AI models, as part of the process. Human review and editing of this content were performed prior to publication.