Background/Objectives: Alzheimer’s disease (AD), amyotrophic lateral sclerosis (ALS), and frontotemporal dementia (FTD) share molecular features yet differ clinically, suggesting underlying systems-level commonalities. We aimed to characterize shared and disease-specific multimorbidity architectures across AD, ALS, and FTD using an artificial intelligence–driven literature-based semantic network. Methods: We applied SemNet 2.0, constructed from over 35 million PubMed abstracts, to analyze disease and syndrome (DSYN) and pharmacological substance (PHSU) nodes. Nodes were ranked using HeteSim and mapped to a harmonized 13-category mechanistic ontology. We quantified pairwise disease intersections, ontology-level enrichment, rank similarity, and intersection–disease alignment, and constructed an integrated multimorbidity priority landscape integrating disease-specific and intersection-level hierarchies. Results: Across AD, ALS, and FTD, a convergent multimorbidity architecture centered on a shared metabolic and immune core was identified, accompanied by prominent neurobehavioral processes and intermediate systems including gastrointestinal, endocrine, hematological, hepatic, and sensory pathways. Disease-specific signatures shaped distinct vulnerability profiles within this shared structure, including cardiovascular enrichment in AD, neuromuscular and toxin-related pathways in ALS, and coupled neurobehavioral–metabolic features in FTD. PHSU patterns reinforced these findings, with centrally positioned compounds predominantly targeting inflammatory, metabolic, or neuromodulatory processes. Conclusions: These findings position AD, ALS, and FTD within a unified, AI-derived multimorbidity framework. This ontology-guided approach provides a computational, hypothesis-generating foundation for multimorbidity-aware biomarker discovery, risk stratification, and cross-disease therapeutic exploration in neurodegenerative disease.
Iyer et al. (Mon,) studied this question.