Nearly half of all global stillbirths and maternal and neonatal deaths result from complications during labor, birth, and the early postnatal period, occurring particularly in low-resource settings. These deaths are largely preventable through timely interventions such as cesarean sections, yet concerns persist about both underuse and overuse of such interventions. Effective strategies that enhance the quality of care during labor, birth, and immediately afterwards are crucial for reducing the incidence of stillbirths and maternal and neonatal deaths.

The WHO has long advocated for the monitoring of women in labor by skilled healthcare providers using a partograph—a clinical tool on paper that records observations to determine whether and when an intervention, such as labor augmentation or cesarean section, is necessary. Despite this, partographs have often been underutilized or not used at all in low- and middle-income countries due to various barriers.

Traditional methods, which include subjective digital vaginal examinations to ascertain fetal head position, rotation, and descent during delivery, have sometimes proven inaccurate. In this context, intrapartum transperineal ultrasound has emerged as an effective method for monitoring fetal head (FH) descent. A significant advancement of this technique is the measurement of the angle of progression (AOP), which provides an objective, accurate, and reproducible indicator. This technique offers a clear advantage over digital vaginal examinations by providing insights into the relationship between the pubic symphysis (PS) and FH (PSFH). According to guidelines from the International Society of Ultrasound in Obstetrics and Gynecology (ISUOG), the AOP is measured on a static 2D ultrasound image and is defined as the angle between the long axis of the pubic bone and a line from the lowest edge of the pubic symphysis that tangentially touches the deepest bony part of the fetal skull. Research has shown that an AOP of 120 degrees or greater is strongly associated with a high likelihood of spontaneous vaginal delivery, making it a valuable predictive indicator for the mode of delivery.

The first step in using this metric is performing PSFH segmentation (PSFHS) – the process of extracting visible PSFH contours from transperineal ultrasound images. However, PSFHS is challenging due to the need for precise identification and delineation of the PSFH boundaries. FH shapes and orientations can vary significantly during different labor stages, and surrounding structures such as amniotic fluid and placenta can overlap or obstruct parts of the head, causing segmentation ambiguity. Additionally, the size and position of PSFH can vary widely among individuals, complicating the development of a universally applicable segmentation model. The inherent characteristics of ultrasound images, such as poor resolution, noise, and artifacts, further challenge the PSFHS process, especially during dynamic changes in the relative positions of PS and FH in the second stage of labor.

To overcome these challenges and foster advancements in AI research, promoting data sharing and developing more comprehensive and representative datasets is essential. In support of this goal, we provided the dataset used for the PSFHS challenge of MICCAI 2023 (https://ps-fh-aop-2023.grand-challenge.org/). This dataset comprises two parts: one is the PSFHS dataset (https://doi.org/10.5281/zenodo.10969427) and the other is from the JNU-IFM dataset (https://doi.org/10.6084/m9.figshare.14371652). The images from the PSFHS dataset will also be utilized for the Intrapartum Ultrasound Grand Challenge (IUGC) 2024 of MICCAI 2024 (https://codalab.lisn.upsaclay.fr/competitions/18413).

Finally, we are delighted to share our work with the scientific community and domain experts in the prestigious journal, Scientific Data. We sincerely hope that this resource can provide valuable research groundwork and further insights for the community.

Reference

1. Fernandez-Turienzo C, Sandall J. Delivering high-quality childbirth care. Nat Med. 2024;30(2):348-349. doi:10.1038/s41591-024-02812-2
2. Zhou M, Wang C, Lu Y, et al. The segmentation effect of style transfer on fetal head ultrasound image: a study of multi-source data. Med Biol Eng Comput. 2023;61(5):1017-1031. doi:10.1007/s11517-022-02747-1
3. Ou, Z., Bai, J., Chen, Z., Lu, Y., Wang, H., Long, S., Chen, G., RTSeg-Net: A Lightweight Network for Real-time Segmentation of Fetal Head and Pubic Symphysis from Intrapartum Ultrasound Images. Computers in biology and medicine, 2024; 108501. doi:10.1016/j.compbiomed.2024.108501
4. Qiu, R., Zhou M, Bai, J., Lu, Y., Wang, H. PSFHSP-Net: An Efficient Lightweight Network for Identifying Pubic Symphysis-Fetal Head Standard Plane from Intrapartum Ultrasound Images. Med Biol Eng Comput. 2024 (In press)
5. Chen, Z. Ou, Y. Lu, J. Bai, Direction-guided and multi-scale feature screening for fetal head–pubic symphysis segmentation and angle of progression calculation, Expert Systems with Applications, 245 (2024) 123096. doi:10.1016/j.eswa.2023.123096
6. Lu Y, Zhou M, Zhi D, et al. The JNU-IFM dataset for segmenting pubic symphysis-fetal head [published correction appears in Data Brief. 2022 Apr 01;42:108128]. Data Brief. 2022;41:107904. Published 2022 Feb 2. doi:10.1016/j.dib.2022.107904
7. Lu Y, Zhi D, Zhou M, et al. Multitask Deep Neural Network for the Fully Automatic Measurement of the Angle of Progression. Comput Math Methods Med. 2022;2022:5192338. Published 2022 Sep 2. doi:10.1155/2022/5192338
8. Bai J, Sun Z, Yu S, et al. A framework for computing angle of progression from transperineal ultrasound images for evaluating fetal head descent using a novel double branch network. Front Physiol. 2022;13:940150. Published 2022 Dec 2. doi:10.3389/fphys.2022.940150
9. Chen G, Bai J, Ou Z, Lu Y, Wang H. PSFHS: Intrapartum ultrasound image dataset for AI-based segmentation of pubic symphysis and fetal head. Scientific Data. 2024. doi:10.1038/s41597-024-03266-4