JoVE - Journal of Visualized Experiments
JoVE Visualize
Contact Us

Spectral CT radiomics features of the tumor and perigastric adipose tissue can predict lymph node metastasis in gastric cancer

  • 0Department of Radiology, Affiliated Kunshan Hospital of Jiangsu University, Kunshan, China.
Abdominal Radiology (new York) +

|

January 25, 2025

|

January 25, 2025

View abstract on PubMed

Summary

This summary is machine-generated.

This study developed a nomogram using dual-layer spectral CT radiomics to predict lymph node metastasis in gastric cancer. The combined model shows high accuracy, potentially improving treatment strategies and patient outcomes.

Area Of Science

  • Radiology
  • Oncology
  • Medical Imaging

Background

  • Gastric cancer (GC) lymph node metastasis (LNM) significantly impacts prognosis.
  • Accurate preoperative prediction of LNM is crucial for guiding treatment strategies.

Purpose Of The Study

  • To develop and validate a nomogram for LNM prediction in GC using radiomics features from dual-layer spectral computed tomography (DLCT).
  • To integrate radiomics features of the tumor and adjacent perigastric adipose tissue for enhanced prediction accuracy.

Main Methods

  • Retrospective analysis of 175 GC patients, divided into training (n=125) and validation (n=50) cohorts.
  • Extraction of radiomics features from tumor and perigastric fat using DLCT spectral images.
  • Construction of radiomics models, clinical-DLCT model, and a combined clinical-DLCT-radiomics nomogram, validated with Bootstrap and ROC analysis.

Main Results

  • Optimal radiomics models achieved high Area Under the Curve (AUC) values (up to 0.923 in training, 0.821 in validation).
  • The clinical-DLCT model showed moderate predictive performance (AUC 0.728 in training, 0.657 in validation).
  • The combined nomogram demonstrated superior predictive efficacy with AUCs of 0.935 (training) and 0.876 (validation).

Conclusions

  • The developed nomogram, integrating clinical factors (Nct, ECV<sub>ID</sub>) and DLCT radiomics features, shows significant potential for predicting LNM in GC.
  • This radiomics-based nomogram can aid in personalized treatment planning and potentially improve clinical outcomes for GC patients.

Keywords: