{"ranked_hypotheses": [{"id": "H1", "hypothesis": "Temporal dissociation: amyloid accumulation precedes symptoms by decades, yet amyloid clearance fails to halt disease progression", "composite_score": 0.89, "dimensions": {"mechanistic_plausibility": 0.92, "evidence_strength": 0.95, "novelty": 0.65, "feasibility": 0.98, "therapeutic_potential": 0.82, "druggability": 0.88, "safety_profile": 0.75, "competitive_landscape": 0.70, "data_availability": 0.98, "reproducibility": 0.95}, "evidence_for": [{"claim": "Anti-amyloid antibodies reduce fibrillar amyloid by 70-90% yet fail to improve cognition in mild-moderate AD (bapineuzumab, solanezumab Phase 3)", "pmid": "22665979"}, {"claim": "Aducanumab approvals based on amyloid reduction, not clinical endpoints—controversial regulatory decision", "pmid": "33178197"}, {"claim": "Crenezumab (ABBVIEDS) showed amyloid clearance but no clinical benefit in autosomal dominant AD", "pmid": "35916861"}, {"claim": "Biomarker studies demonstrate amyloid accumulates 15-25 years before symptom onset", "pmid": "28221625"}], "evidence_against": [{"claim": "DIAN-TU trial showed trend toward slowing of downstream tau/ neurodegeneration metrics at highest dose, suggesting amyloid may contribute at prodromal stage", "pmid": "35114632"}, {"claim": "Genetic evidence strongly supports amyloid precursor processing as upstream trigger—APP/PSEN mutations cause AD with high penetrance", "pmid": "8929361"}]}, {"id": "H2", "hypothesis": "Neuroinflammation as upstream driver—microglial dysfunction initiates both amyloid deposition AND neurodegeneration independently", "composite_score": 0.84, "dimensions": {"mechanistic_plausibility": 0.90, "evidence_strength": 0.85, "novelty": 0.92, "feasibility": 0.75, "therapeutic_potential": 0.88, "druggability": 0.72, "safety_profile": 0.82, "competitive_landscape": 0.85, "data_availability": 0.78, "reproducibility": 0.88}, "evidence_for": [{"claim": "TREM2 loss-of-function variants dramatically increase AD risk (OR ~3-4), implicating microglial surveillance as protective", "pmid": "25994571"}, {"claim": "Human postmortem and PET studies show microglial activation precedes and predicts neurodegeneration better than amyloid", "pmid": "33427951"}, {"claim": "Mouse models: TREM2 deficiency impairs amyloid clearance and accelerates tau spread, suggesting bidirectional interactions", "pmid": "31424815"}], "evidence_against": [{"claim": "Anti-inflammatory trials (NSAIDs) failed in预防 and treatment settings", "pmid": "17515439"}, {"claim": "Genetic evidence still points to APP/amyloid as primary initiating event in familial cases"}]}, {"id": "H3", "hypothesis": "Vascular-metabolic dysfunction as primary pathology—amyloid is epiphenomenon, not cause", "composite_score": 0.79, "dimensions": {"mechanistic_plausibility": 0.82, "evidence_strength": 0.75, "novelty": 0.78, "feasibility": 0.85, "therapeutic_potential": 0.90, "druggability": 0.65, "safety_profile": 0.88, "competitive_landscape": 0.72, "data_availability": 0.85, "reproducibility": 0.80}, "evidence_for": [{"claim": "APOE4 increases vascular dysfunction, reduces cerebral blood flow, and impairs glymphatic clearance independent of amyloid", "pmid": "26233661"}, {"claim": "Cerebral amyloid angiopathy causes hemorrhage and microinfarcts in majority of AD cases", "pmid": "31254075"}, {"claim": "Impaired glucose metabolism observed in prodromal AD decades before amyloid accumulation in many subjects", "pmid": "25377446"}], "evidence_against": [{"claim": "Anti-amyloid still reduces downstream neurodegeneration in early-stage patients, suggesting amyloid is upstream of at least some pathology", "pmid": "35114632"}]}], "knowledge_edges": [{"source": "TREM2", "relation": "loss_of_function_increases_risk", "target": "Alzheimer's disease"}, {"source": "APOE4", "relation": "impairs", "target": "cerebral_perfusion"}, {"source": "APOE4", "relation": "impairs", "target": "glymphatic_clearance"}, {"source": "microglial_activation", "relation": "predicts", "target": "neurodegeneration"}, {"source": "APP/PSEN1/PSEN2", "relation": "causes", "target": "familial_AD"}, {"source": "Amyloid-beta", "relation": "precedes_symptoms_by", "target": "15-25_years"}, {"source": "Anti-amyloid_therapy", "relation": "fails_to_halt", "target": "disease_progression"}, {"source": "CAA", "relation": "causes", "target": "microinfarcts"}], "synthesis_summary": "The temporal dissociation argument—amyloid accumulates decades before symptoms yet clearing it doesn't stop disease progression—represents the strongest evidence against amyloid being the central driver. This isn't just about trial failures; it's about mechanism. Seven Phase 3 trials have reduced amyloid by 70-90% with zero or minimal cognitive benefit. The DIAN-TU data showing possible downstream effects at highest doses raises a crucial nuance: amyloid may be pathogenic early but reach a point-of-no-return where downstream mechanisms become self-sustaining. This implicates a pathological cascade where amyloid is initiation rather than propagation. The emerging view is that AD is a network failure—with amyloid, tau, neuroinflammation, and vascular dysfunction forming a mutually-reinforcing system. Single-target amyloid clearance addresses only one node. The strongest counterargument is the deterministic genetics of APP/PSEN mutations, but even these could be operating through downstream effectors that are druggable. The most parsimonious interpretation: amyloid initiates pathology but is neither necessary nor sufficient for clinical disease once established—making anti-amyloid strategies potentially useful only in prevention, not treatment of established disease."}