Self-supervised 3D deep learning on preoperative contrast-enhanced computed tomography for predicting high pathologic nodal burden in esophageal squamous cell carcinoma: temporal and external multicohort validation - Takeaways - MDSpire

Self-supervised 3D deep learning on preoperative contrast-enhanced computed tomography for predicting high pathologic nodal burden in esophageal squamous cell carcinoma: temporal and external multicohort validation

  • By

  • Qian Li

  • Yongxin Li

  • Jing Bai

  • Jinze Zhang

  • Zhenkai Nie

  • Yanxin Ren

  • Zhantao Li

  • Jin Guo

  • Meng Li

  • Tingting Zhang

  • Xinbo Liu

  • June 11, 2026

  • 0 min

Share

  • 1

    The study involved 1,060 patients with esophageal squamous cell carcinoma undergoing preoperative contrast-enhanced CT and curative-intent surgery.

  • 2

    A self-supervised 3D deep learning model was developed to predict high pathologic nodal burden, achieving high performance across multiple cohorts.

  • 3

    The model demonstrated area under the receiver operating characteristic curve values of 0.881 in the internal test cohort and 0.860 in both temporal and external cohorts.

  • 4

    Sensitivity and specificity for the self-supervised model were 0.581 and 0.845, outperforming a guideline-inspired comparator in the external cohort.

  • 5

    The study highlights the potential of self-supervised learning to improve preoperative nodal staging accuracy in esophageal squamous cell carcinoma.

Original Source(s)

Self-supervised 3D deep learning on preoperative contrast-enhanced computed tomography for predicting high pathologic nodal burden in esophageal squamous cell carcinoma: temporal and external multicohort validation

  • by Qian Li, Yongxin Li, Jing Bai, Jinze Zhang, Zhenkai Nie, Yanxin Ren, Zhantao Li, Jin Guo, Meng Li, Tingting Zhang, Xinbo Liu

  • June 11, 2026

Related Content

Global trends and academic landscapes of AI applications in basal cell carcinoma research: a bibliometric analysis

Case Report: Marjolin ulcer characterized by squamous cell carcinoma arising from chronic wounds: clinical features, diagnostic and therapeutic challenges, and insights—a report of four cases

Molecular and microenvironmental drivers of malignant transformation from oral potentially malignant disorders to oral squamous cell carcinoma