Quality Control of Ultrasound Images During Early Pregnancy Via AI

Official title

Deep Learning-based Quality Control of Ultrasound Images During Early Pregnancy

Overview

This research integrates artificial intelligence to enhance early pregnancy ultrasonography quality control, focusing on specific fetal sections. In collaboration with prominent medical institutions, the investigators have amassed extensive fetal ultrasound data. The investigators aim to develop a deep learning model that can accurately identify essential anatomical areas in ultrasound images and evaluate their quality. This tool is expected to significantly decrease misdiagnoses of conditions like Down Syndrome and neural system deformities by ensuring real-time image quality assessment.

Detailed description

This research is dedicated to integrating artificial intelligence technology to optimize the quality control process of early pregnancy ultrasonography. The ultrasound images involved primarily focus on the median sagittal section, NT section, and choroid plexus of the fetus during early pregnancy. In this regard, the investigators have collaborated with renowned medical institutions such as Beijing Obstetrics and Gynecology Hospital, Peking University Third Hospital, Changsha Hospital for Maternal and Child Health Care, and Second Xiangya Hospital of Central South University to retrospectively and prospectively collect a vast amount of early pregnancy fetal ultrasound image data. Based on this, the investigators plan to establish a model rooted in deep learning. This model will be capable of precisely identifying key anatomical regions in standard ultrasound scan images. Furthermore, by recognizing these anatomical structures, the model will determine whether the ultrasound image meets the standard scanning quality. This model is anticipated to serve as a powerful auxiliary tool in obstetric ultrasonography, enabling real-time assessment of ultrasound image quality, thereby significantly reducing the rates of missed and misdiagnosed fetal diseases such as Down Syndrome and neural system malformations.

Interventions

• Other Image quality control
  The investigators identify the region of interest in the relevant section to give a conclusion on whether the image is standard or not, guiding clinicians to standardize the operation, and reducing the rate of misdiagnosis and underdiagnosis.

Primary outcome measures

• PR curve of image quality control module [Time frame: one month]

Secondary outcome measures (1)

• The accuracy of intelligent analysis system in image quality control module [Time frame: one month]

Eligibility criteria

Inclusion criteria

• Women in early pregnancy who have detailed personal information and ultrasound images.
• The ultrasound images should clearly show the fetus's median sagittal, NT, and choroid plexus views.

Exclusion criteria

• Ultrasound images from women in mid to late pregnancy.
• Ultrasound images that are unclear or blurry, making evaluation difficult.
• Women who did not provide complete personal and medical information during the ultrasound scan.

Criteria are shown verbatim from the registry (in English). Final eligibility is always assessed by the study center.

Healthy volunteers: Yes

Study design

Observational model

Cohort

Study locations

China · 4 centers

• Beijing Obstetrics and Gynecology Hospital affiliated to Capital Medical University — Beijing
• Peking University Third Hospital — Beijing
• Changsha Hospital for Maternal and Child Health Care — Changsha
• Second Xiangya Hospital of Central South University — Changsha

Identifiers

NCT: NCT06002412 · CASMI005

Primary sources (government registries)

• ClinicalTrials.gov (US NIH)

View this study on ClinicalTrials.gov ↗