Disproportionately elevated sulcal index (DESI): An automatically driven index representing disproportionate...

Idiopathic normal pressure hydrocephalus (iNPH) is frequently underdiagnosed due to non-specific symptoms and the risks of invasive testing. While disproportionately enlarged subarachnoid space hydrocephalus (DESH) is a hallmark imaging feature, manual assessment is subjective and qualitative. We developed the Disproportionately Elevated Sulcal Index (DESI), a fully automated deep learning-based volumetric biomarker, to objectively quantify these morphological changes.

We trained a U-Net model with an EfficientNet-B0 encoder on T1-weighted MRI scans from the Baltimore Longitudinal Study of Aging and Johns Hopkins Clinic (n = 1248) to segment Sylvian fissures and superior sulcal spaces. DESI was defined as the volumetric ratio of the Sylvian fissure to superior sulci within an AC-PC aligned wedge. The model was externally validated on the multi-site PENS trial dataset (n = 94), comparing NPH patients against healthy controls and participants with Alzheimer's disease and vascular dementia.

In external validation, DESI demonstrated high diagnostic accuracy. The index distinguished NPH patients with DESH features from non-DESH NPH cases with an Area Under the Curve (AUC) of 0.97. When differentiating NPH with DESH from a pooled group of healthy controls and neurodegenerative mimics, DESI achieved an AUC of 0.99 (sensitivity 98%, specificity 100%). In a broad comparison of all NPH cases versus all non-NPH groups, DESI maintained an AUC of 0.94.

DESI provides a robust, fully automated quantification of sulcal disproportion that effectively differentiates iNPH from normal aging and neurodegenerative mimics. This continuous, non-invasive metric offers a scalable tool for accurate iNPH screening and patient stratification in clinical settings.