Development of a CT-based comprehensive model with deep learning for differentiating pathological types of pulmonary ground-glass nodules - Takeaways - MDSpire

Development of a CT-based comprehensive model with deep learning for differentiating pathological types of pulmonary ground-glass nodules

  • By

  • Jian Zhang

  • Boheng Liu

  • Ji Li

  • Yang Liu

  • Jipeng Jiang

  • May 26, 2026

  • 0 min

Share

  • 1

    The study developed an integrated model to differentiate benign from malignant pulmonary pure ground-glass nodules (pGGNs) using clinical, radiomics, and deep learning features.

  • 2

    A total of 1,067 patients with pGGNs were included, with clinical and imaging data collected for model training and validation.

  • 3

    The Support Vector Machine classifier achieved the highest performance for distinguishing pGGNs, with a validation AUC of 0.840.

  • 4

    The integrated model combining all features achieved a validation AUC of 0.871 for benign versus malignant classification.

  • 5

    For subtype classification, the integrated model reached a validation AUC of 0.853, indicating its potential for non-invasive decision support.

Original Source(s)

Development of a CT-based comprehensive model with deep learning for differentiating pathological types of pulmonary ground-glass nodules

  • by Jian Zhang, Boheng Liu, Ji Li, Yang Liu, Jipeng Jiang

  • May 26, 2026

Related Content

Novel approaches for separating graft-versus-leukemia effects from graft-versus-host disease

Icaritin plus TACE improves survival in advanced HCC with macrovascular invasion: a multicenter cohort study

Myelodysplastic syndromes complicated by atypical Sweet syndrome: a brief research report