Lateral ventricle morphology analysis via mean latitude axis

Area of Science:

• Neuroimaging
• Medical Image Analysis
• Developmental Neuroscience

Background:

• Statistical shape analysis is crucial for neuroimaging but often relies on surface alignment.
• Medial descriptions offer an alignment-independent shape metric using thickness information.

Purpose of the Study:

• To propose and validate a joint framework for computing local medial thickness using a mean latitude axis from the Spherical Harmonic (SPHARM-PDM) shape framework.
• To apply this novel method to analyze the morphology of lateral ventricles in neonates with mild ventriculomegaly (MVM).

Main Methods:

• Developed a joint framework integrating SPHARM-PDM with medial representations to compute local medial thickness.
• Applied the method to 3D MRI scans of 30 MVM neonates and 60 age- and sex-matched healthy controls.
• Utilized local MANOVA for statistical comparison between MVM and control groups, contrasting with traditional SPHARM-PDM analysis.

Main Results:

• Both surface and mean latitude axis analyses successfully differentiated MVM and healthy lateral ventricle morphology.
• The mean latitude axis analysis identified significant differences along the entire lateral ventricle shape.
• MVM neonates exhibited enlarged lateral ventricles in the tail and head regions compared to controls.

Conclusions:

• The proposed joint framework provides significant insights into local thickness analysis, surpassing traditional SPHARM-PDM methods.
• This study is the first to precisely quantify 3D lateral ventricle morphology in MVM neonates using advanced shape analysis.
• Medial representations offer a robust, alignment-independent approach for neuroimaging shape analysis in pediatric populations.