A new automatic algorithm to extract craniofacial measurements from fetal three-dimensional volumes

Area of Science:

• Medical Imaging
• Prenatal Diagnostics
• Computational Anatomy

Background:

• Three-dimensional (3D) ultrasound offers multiplanar views for fetal craniofacial evaluation.
• Current 3D ultrasound measurements are time-consuming and subjective, requiring expert input for area and view selection.

Purpose of the Study:

• To develop and evaluate an automated image analysis system for precise measurement of fetal craniofacial structures using 3D ultrasound.
• To assess the system's performance in the second trimester of pregnancy with a novel 3D volume analysis algorithm.

Main Methods:

• A universal facial surface template model was created from a fetal phantom.
• The system fitted this template to 3D fetal facial volumes from 11 fetuses using model deformation.
• Automatic measurements included biparietal diameter (BPD), occipitofrontal diameter (OFD), interorbital diameter (IOD), bilateral orbital diameter (BOD), and vertex-nasion distance (VN).

Main Results:

• High correlation coefficients were achieved: 0.99 for BPD and OFD, 0.80 for BOD, 0.83 for IOD, and 0.99 for VN.
• No significant differences were found between the automated system's measurements and manual expert measurements.
• The system demonstrated precise measurement capabilities for fetal craniofacial structures.

Conclusions:

• The proposed automated 3D ultrasound system precisely measures fetal craniofacial structures, showing potential for clinical application.
• The system's accuracy and efficiency could enhance prenatal evaluation and may be applicable to other clinical fields.