Genetic signatures of responsiveness in idiopathic normal pressure Hydrocephalus: Insights from whole-exome and LASSO-based analysis

Background: Idiopathic normal pressure hydrocephalus (iNPH) is a potentially reversible neurological disorder characterised by the triad of gait disturbance, cognitive decline, and urinary incontinence, together with ventriculomegaly despite normal cerebrospinal fluid (CSF) pressure (1–3). Although traditionally considered non-genetic and multifactorial, emerging data suggest that rare pathogenic variants and cilia-related mechanisms may contribute to its pathogenesis and modulate responsiveness to CSF diversion (8–12). Methods: We performed whole-exome sequencing (WES) in 33 consecutive patients with clinically and radiologically confirmed iNPH who underwent a standardised 120-h lumbar drainage (LD) protocol. Responders were defined as those showing ≥20 % improvement in gait speed or ≥3-point increase in Montreal Cognitive Assessment (MoCA) score compared with baseline. Variants were filtered according to ACMG/AMP guidelines (19) and population allele frequency (minor allele frequency <1 %). Least Absolute Shrinkage and Selection Operator (LASSO) logistic regression (20) with ten-fold cross-validation was applied to identify gene-level predictors of LD responsiveness. Results: After filtering, 110 genes with at least one pathogenic or likely pathogenic variant were retained for modelling. LASSO regression at the λ.1se threshold identified a five-gene panel—FANCD2, ATR, ORAI1, MUC1, and RP1L1—involved in DNA damage response, calcium signalling, epithelial barrier integrity, and ciliary architecture (21–28). The internally derived model achieved an accuracy of 81.8 %, sensitivity 68.8 %, specificity 94.1 %, and positive predictive value 91.7 % for prediction of LD response. Conclusions: Rare pathogenic variants in genes linked to genome stability, immune and calcium signalling, and ciliary structure may influence LD responsiveness in iNPH. The five-gene model represents a biologically plausible, hypothesis-generating tool for preoperative risk stratification. Validation in larger, multicentre cohorts, integration with shunt outcomes, and functional studies are required before any clinical implementation.