A CT-based multi-scale fusion model with SHAP interpretation for preoperative differentiation between lung adenocarcinoma in situ/minimally invasive adenocarcinoma and invasive adenocarcinoma: a multicenter study - Takeaways - MDSpire

A CT-based multi-scale fusion model with SHAP interpretation for preoperative differentiation between lung adenocarcinoma in situ/minimally invasive adenocarcinoma and invasive adenocarcinoma: a multicenter study

  • By

  • Ning Dong

  • Shihang Sun

  • Zhongwei Li

  • Zhenjie Cong

  • Yi Lin

  • Jingxian Chen

  • Hu Zhang

  • Yuxin Liu

  • Shuxia Li

  • Jing Xu

  • Demin Kong

  • Qingfeng Yin

  • May 12, 2026

  • 0 min

Share

  • 1

    A multi-scale fusion model was developed to differentiate between lung adenocarcinoma in situ/minimally invasive adenocarcinoma and invasive adenocarcinoma.

  • 2

    The study analyzed 621 patients with ground-glass nodules from two centers, utilizing radiomics, habitat analysis, and deep learning.

  • 3

    The early fusion model outperformed others, achieving areas under the ROC curves of 0.988, 0.903, and 0.872 in training, validation, and testing cohorts.

  • 4

    The 2.5D ResNet50 deep learning architecture demonstrated the best performance among the evaluated models.

  • 5

    This model supports precise preoperative diagnosis and individualized treatment decisions for patients with ground-glass nodules.

Original Source(s)

Related Content