Multimodal machine learning for menopause status prediction using LLM-extracted ultrasound features - Takeaways - MDSpire

Multimodal machine learning for menopause status prediction using LLM-extracted ultrasound features

  • By

  • Weiwei Yin

  • Zhengyuan Shen

  • Chun Feng

  • Xia Zhang

  • Sihao Shen

  • Yiyue Jiang

  • Zhenbo Cheng

  • Lihui Wang

  • Ling Liu

  • July 2, 2026

  • 0 min

Share

  • 1

    The study involved 713 women for training and 284 for validation to predict menopausal status using ultrasound features and anthropometric data.

  • 2

    Large language models (LLM) effectively extracted structured morphological features from unstructured ultrasound report text.

  • 3

    The highest AUC of 0.984 was achieved by combining anthropometric and hormone features in the validation set.

  • 4

    The multimodal prediction model maintained high accuracy even when certain feature types were missing.

  • 5

    This method offers a supplementary evaluation for menopausal status, particularly when hormone data is unavailable.

Original Source(s)

Multimodal machine learning for menopause status prediction using LLM-extracted ultrasound features

  • by Weiwei Yin, Zhengyuan Shen, Chun Feng, Xia Zhang, Sihao Shen, Yiyue Jiang, Zhenbo Cheng, Lihui Wang, Ling Liu

  • July 2, 2026

Related Content

The rise of perimenopause misinformation

Top 10 States With Rising Demand for Physicians

These 10 states report physician shortages, projected physician supply gaps, or physician-to-population ratios below national benchmarks.

Adiposity and domain-specific menopausal symptom severity in midlife women: a cross-sectional clinical study in eastern china