MRI Findings in Normal Pressure Hydrocephalus
MRI without IV contrast is the preferred imaging modality for NPH, demonstrating characteristic findings of ventriculomegaly with Evans index >0.3, callosal angle <90°, enlarged temporal horns, tight high-convexity sulci, widened Sylvian fissures, and aqueductal flow void. 1, 2
Essential Structural Features
Ventricular Enlargement Pattern:
- Ventriculomegaly not entirely attributable to cerebral atrophy or congenital enlargement, with Evans index (maximal width of frontal horns/maximal width of inner skull) >0.3 1, 2
- Rounded frontal horns with marked enlargement of temporal horns and third ventricle 1, 3
- No macroscopic obstruction of CSF flow 1, 2
Disproportionately Enlarged Subarachnoid-Space Hydrocephalus (DESH) Pattern:
- Tight high-convexity and medial subarachnoid spaces (effaced sulci along high convexities) 2, 4, 5
- Enlarged Sylvian fissures 2, 6, 4
- This combination of ventriculomegaly with tight high-convexity sulci and enlarged Sylvian fissures is highly specific for NPH and differentiates it from simple atrophy 7, 4
Corpus Callosum Changes
- Callosal angle <90° (narrowed posterior callosal angle) 1, 2, 6
- Corpus callosum thinning and elevation 6
- The callosal angle should be measured using standardized AC-PC line protocol and evaluated alongside other NPH features 2
White Matter and CSF Flow Abnormalities
Periventricular Changes:
- Periventricular white matter hyperintensities on T2-weighted imaging (evidence of altered brain water content) 1, 2, 6
- MRI has higher sensitivity for detecting these periventricular changes compared to CT 2
CSF Flow Dynamics:
- Aqueductal or fourth ventricle flow void on MRI 1, 2, 3
- Increased CSF flow void through the cerebral aqueduct correlates with good response to shunt surgery 1
- This flow void cannot be visualized on CT, making MRI superior for prognostic assessment 2
Cortical Atrophy Assessment
- Absence of or only mild cortical atrophy despite significant ventriculomegaly 1, 6
- This disproportionate relationship between ventricular size and cortical atrophy is a key distinguishing feature from neurodegenerative diseases 7, 5
Advanced MRI Techniques for Shunt Responsiveness
Cine MRI with Phase-Contrast Imaging:
- Hyperdynamic aqueductal CSF flow on cine MRI helps identify shunt-responsive NPH patients 1, 2
- Elevated aqueductal CSF stroke volume measured by phase-contrast MRI demonstrates high positive predictive value (up to 100% in some studies) for shunt responsiveness 3, 8
- Increased ventricular flow rate on cine MRI is supportive but not necessary for probable NPH designation 1
Supportive but Not Required Features
- MRI performed before symptom onset showing smaller ventricular size 1
- Widening of temporal horns not fully explained by hippocampal atrophy 6
- Enlarged basal cisterns 6
Critical Diagnostic Considerations
Common Pitfalls to Avoid:
- Do not diagnose NPH based on ventriculomegaly alone, as this is nonspecific and common in aging 7, 5
- Recognize that approximately 30% of NPH cases have concomitant Alzheimer pathology, which may complicate diagnosis 7
- MR spectroscopy cannot reliably distinguish NPH from other dementias and does not correlate with shunt outcomes 1, 9
When MRI is Contraindicated:
- CT head without IV contrast is an acceptable alternative that can identify ventriculomegaly, narrowed callosal angle, effaced sulci, and widened Sylvian fissures 2, 9
- However, CT has lower sensitivity for periventricular white matter changes and cannot detect aqueductal flow void 2, 9
- CT may miss small obstructing lesions 2
Integration with Clinical Assessment
The MRI findings must be interpreted in conjunction with the clinical triad of gait disturbance (hypokinetic gait with feet appearing "glued to the floor"), cognitive impairment, and urinary incontinence 3, 6. When MRI demonstrates the complete constellation of NPH features—particularly DESH pattern with aqueductal flow void—and clinical symptoms align, proceed to CSF diversion testing (large-volume lumbar puncture or external lumbar drainage) to confirm shunt responsiveness 2, 3.