Fluid and Imaging Biomarkers in Lecanemab Therapy for Early Alzheimer’s Disease: Current Evidence, ARIA Risk Prediction, and Treatment Monitoring

Background/Objectives: Lecanemab, a humanized anti-amyloid-β protofibril antibody, achieves substantial amyloid clearance and modest slowing of clinical decline in early Alzheimer’s disease (AD). Amyloid-related imaging abnormalities (ARIA) represent the most clinically important safety concern, yet no fluid biomarker has been prospectively validated for ARIA prediction in lecanemab-treated patients. This review critically synthesizes current evidence on fluid and imaging biomarkers for ARIA risk stratification and treatment-response monitoring, with emphasis on distinguishing clinically actionable markers from those that remain investigational, and on the interpretive challenge of differentiating pseudo-atrophy from true neurodegeneration. Methods: A narrative literature review was conducted across PubMed, Google Scholar, Scopus, Web of Science, Embase, and the Cochrane Library, covering English-language publications from January 2020 to April 2026, supplemented by landmark earlier studies, regulatory documents, and appropriate-use recommendations. Evidence was synthesized narratively and classified into four tiers: Established, Emerging, Exploratory, and Hypothesis-generating. Results: APOE ε4 genotype and baseline MRI markers of cerebral amyloid angiopathy are the only established, clinically actionable ARIA risk stratifiers. In CLARITY-AD, ARIA-E and ARIA-H occurred across all APOE genotype groups. Plasma p-tau217 demonstrates the strongest evidence for downstream biological response monitoring, while neurofilament light chain (NfL) dynamics help distinguish pseudo-atrophy from ongoing neurodegeneration. GFAP is exploratory, while sTREM2, endothelial markers, and endogenous anti-Aβ antibodies remain hypothesis-generating, with no prospective validation in lecanemab-treated cohorts. Assay-platform variability substantially limits cross-study comparability for all quantitative fluid biomarkers. Conclusions: Current ARIA monitoring remains anchored in APOE genotyping, baseline MRI assessment, and serial MRI surveillance. Plasma p-tau217 and NfL provide complementary biological context for treatment-response interpretation but cannot replace established imaging protocols. Future studies should prioritize prospective multicenter validation of multimodal ARIA prediction models integrating fluid biomarkers, APOE genotype, baseline cerebrovascular imaging, and standardized assay platforms.