Large-Scale Validation of Aberration Correction Algorithms in Clinical Fetal Ultrasound Scans

Fetal ultrasound imaging is the primary modality for prenatal screening and diagnosis. Optimal image quality is essential to ensure accurate delineation and visualization of the inherently small fetal structures. One of the major challenges in obtaining high-quality ultrasound images is phase aberration, leading to ultrasound wavefront distortion, resulting in reduced image spatial resolution and contrast. Various aberration correction (AC) methods have been proposed; however, very few studies have evaluated AC using extensive clinical data. In this study, we collected 243 datasets from 82 patients, disregarding their body mass index (BMI), included all gestational ages and three different fetal anatomical views (brain, heart, and abdomen), to assess the performance and robustness of AC algorithms in obstetric ultrasound imaging. In addition, we compared two AC methods, namely, the multiple speed-of-sound beamforming (MSSB) and cross correlation (CC), using a quantitative image-quality metric, and asked three gynecologists to rank the uncorrected and corrected images. The results showed that AC using either method noticeably enhanced image quality. Quantitatively, image sharpness was increased in over 90% of the cases, with median increase values of 1.46% and 1.70% for MSSB and CC, respectively. From a clinical perspective, all clinicians generally preferred the AC images, and rated the image quality superior or similar after AC in over 75% of the time.