Acta Neurochirurgica (2026) 168:155
This study validates the DESH radiological scoring system as a predictor of shunt surgery response in 127 patients with idiopathic normal pressure hydrocephalus (iNPH) at a Danish neurosurgical center, showing a significant positive association at a mean 7.4-month follow-up. The analysis adjusted for age, sex, and imaging modality, with blinded radiologic assessment and multivariate logistic regression demonstrating increased odds of a positive outcome per DESH point.
The paper also examines whether CT versus MRI confounds DESH grading, finding no evidence of modality-driven bias in this cohort but noting that formal equivalence requires paired within-patient imaging. Limitations include selection bias from preselected tested patients and lack of paired imaging, while strengths are external validation, cohort size, and blinded scoring.
Objective External validation of the DESH score as a prognostic marker for shunt response in idiopathic normal pressure hydrocephalus (iNPH), and assessment of whether CT vs MRI confounds DESH grading/outcome association.
Cohort Retrospective study of 127 shunt-operated iNPH patients (2013–2020) with preoperative CT and/or MRI and ≥2 months follow-up; mean follow-up 7.4 months.
Predictors assessed Radiologic markers (Evans’ index, callosal angle, Sylvian fissure dilation, tight high convexity, focal sulci) and their combined DESH score were compared against shunt response.
Primary finding Higher DESH score was significantly associated with objective shunt response after adjustment for age, sex, and modality (OR 1.35, p=0.036).
Modality impact No significant difference in DESH score assessments between CT and MRI, and including/excluding modality did not meaningfully change regression coefficients.
Outcome rate 81.9% of patients had a positive response to shunt surgery in this cohort.
Interpretation CT may be a practical alternative for DESH-based initial evaluation when MRI is impractical, but formal equivalence of CT vs MRI DESH scoring requires paired-imaging studies.
Key limitation Cohort was preselected based on prior diagnostic testing (e.g., infusion/tap testing), increasing pretest probability and limiting generalizability (potential selection bias).