Fetal Lung Development Data for GDM and PE Assessment via Multimodal Fusion

1,183 human fetal lung ultrasound images were analyzed alongside rat model data including ultrasound, histopathology, and molecular profiles. The dataset was created by Yanran Du and last updated on 2026-05-18. It supports a multimodal deep learning framework for assessing fetal lung development in high-risk pregnancies.

Use Cases

• Training multimodal fusion models based on combined ultrasound imaging, histopathology, and molecular data.
• Assessing fetal lung development differences based on data from normal, GDM, and PE groups.
• Validating transfer learning approaches for clinical applications using animal model data.
• Conducting differential expression and functional enrichment analysis based on gene sequencing profiles.

Strengths

• Includes 1,183 human fetal lung ultrasound images for clinical validation.
• Multimodal integration of ultrasound, histopathology, immunohistochemistry, and gene sequencing data from rat models.
• Reported model performance metrics include accuracies up to 0.948 and AUCs up to 0.955 in animal studies.

Limitations

• Dataset size is 28.1 KB, indicating a very limited scope, likely containing summary results or a document rather than raw data.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment for model training.

Provenance

Source

figshare, author Yanran Du.

Collection Method

Data derived from rat models of GDM and PE and human fetal ultrasound images, analyzed via deep learning, histopathology, and molecular techniques.

Freshness

Last updated 2026-05-18 05:41:14; freshness should be verified.