Robust Alignment of the Human Embryo in 3D Ultrasound using PCA and an Ensemble of Heuristic, Atlas-based and Learning-based Classifiers Evaluated on the Rotterdam Periconceptional Cohort

TL;DR

The paper tackles automated, robust alignment of the human embryo in 3D ultrasound to enable consistent standard plane detection and biometric analysis in the first trimester. It introduces a PCA-based rigid alignment that yields four orientation candidates from embryo segmentations and uses three independent selection strategies—Pearson Heuristic, atlas-based cross-correlation, and a Random Forest—augmented by a Majority Vote to choose the standard orientation. On 2166 test images from 1043 pregnancies in the Rotterdam Periconceptional Cohort, the method achieved 99.0% correct PCA axis extraction and 98.5% overall orientation accuracy with Majority Vote, outperforming individual selectors. The approach is computationally efficient, training-light for most components, and provides a scalable pre-processing step with open-source code for clinical and research use in early pregnancy imaging.

Abstract

Standardized alignment of the embryo in three-dimensional (3D) ultrasound images aids prenatal growth monitoring by facilitating standard plane detection, improving visualization of landmarks and accentuating differences between different scans. In this work, we propose an automated method for standardizing this alignment. Given a segmentation mask of the embryo, Principal Component Analysis (PCA) is applied to the mask extracting the embryo's principal axes, from which four candidate orientations are derived. The candidate in standard orientation is selected using one of three strategies: a heuristic based on Pearson's correlation assessing shape, image matching to an atlas through normalized cross-correlation, and a Random Forest classifier. We tested our method on 2166 images longitudinally acquired 3D ultrasound scans from 1043 pregnancies from the Rotterdam Periconceptional Cohort, ranging from 7+0 to 12+6 weeks of gestational age. In 99.0% of images, PCA correctly extracted the principal axes of the embryo. The correct candidate was selected by the Pearson Heuristic, Atlas-based and Random Forest in 97.4%, 95.8%, and 98.4% of images, respectively. A Majority Vote of these selection methods resulted in an accuracy of 98.5%. The high accuracy of this pipeline enables consistent embryonic alignment in the first trimester, enabling scalable analysis in both clinical and research settings. The code is publicly available at: https://gitlab.com/radiology/prenatal-image-analysis/pca-3d-alignment.

Figures (3)

• Figure 1: Overview of our alignment method with an example 3D ultrasound image acquired in the 10th gestational week. Our method takes as input a 3D segmentation of the embryo and then calculates possible candidate images using Principal Component Analysis (PCA). Three selection methods assess all candidates and select the one in standard orientation. These are also combined in a Majority Vote.
• Figure 2: Example of a reduced 3D point cloud of a fetus from the 11th gestational week. The Pearson correlation coefficients, $r_{top}$ and $r_{bottom}$, are computed for both the top and bottom section, respectively. The selection criteria for the Pearson Heuristic method, $|r_{bottom}| > |r_{top}|$ and $r_{bottom} > 0$, hold true in the second point cloud.
• Figure 3: Columns (a) to (c) show correctly aligned test set images across gestational weeks, while columns (d) and (e) show failure cases to due non-neutral position and missing limbs in the segmentation, respectively. The mid-sagittal and mid-coronal planes as well as the 3D rendering are shown per column for each unique scan.