Screening for Normal Pressure Hydrocephalus on Head CT Using Automated Callosal Angle Assessment

Background/Objectives: Normal pressure hydrocephalus (NPH) is a treatable cause of gait impairment and fall risk in older adults, yet it remains frequently underdiagnosed. This study aimed to validate an automated measurement of the callosal angle, a recognized imaging marker of NPH, adapted for use on routine head computed tomography (CT). Methods: We performed a retrospective analysis of 198 patients with probable NPH, confirmed by gait improvement following lumbar tap test, and 198 age- and sex-matched controls presenting with headache and negative head CT findings (mean age 74 ± 7 years; 60% male in both groups). Manual callosal angle measurements were independently obtained by trained residents and reviewed by neuroradiologists. Automated and manual measurements were compared using intraclass correlation, and diagnostic performance was assessed across threshold values. Results: Automated callosal angle measurements demonstrated strong agreement with manual measurements (ICC = 0.90). Using an automated callosal angle threshold of <90°, diagnostic accuracy was 84.1%, with sensitivity of 90.4% and specificity of 77.8%. Optimization to a 95° threshold yielded an accuracy of 85.9%, with both sensitivity and specificity of 85.9%. The area under the receiver operating characteristic curve was 0.915 (95% CI, 0.897-0.933). Conclusions: Automated callosal angle assessment on routine head CT provides reliable and scalable detection of NPH, supporting its use as a screening tool to facilitate earlier diagnosis and treatment of a potentially reversible cause of dementia.