JoVE - Journal of Visualized Experiments
JoVE Visualize
Contact Us

Deep learning nomogram for predicting neoadjuvant chemotherapy response in locally advanced gastric cancer patients

  • 0Department of Radiology, The Fourth Hospital of Hebei Medical University, Shijiazhuang, People's Republic of China.
Abdominal Radiology (new York) +

|

May 26, 2024

|

May 26, 2024

View abstract on PubMed

Summary

This summary is machine-generated.

A deep learning nomogram accurately predicts neoadjuvant chemotherapy (NAC) response in locally advanced gastric cancer (LAGC) patients using CT scans. This tool aids personalized treatment decisions for gastric cancer.

Area Of Science

  • Radiology
  • Oncology
  • Artificial Intelligence

Background

  • Locally advanced gastric cancer (LAGC) requires effective neoadjuvant chemotherapy (NAC) to improve treatment outcomes.
  • Predicting NAC response is crucial for tailoring treatment strategies and improving patient prognosis.

Purpose Of The Study

  • To develop and validate a deep learning radiomics nomogram for predicting NAC response in LAGC patients.
  • To integrate multi-phase contrast-enhanced computed tomography (CECT) imaging features with clinical data for enhanced prediction accuracy.

Main Methods

  • A multi-center retrospective study included 322 LAGC patients.
  • Handcrafted radiomics and EfficientNet V2 deep learning models were applied to CECT images.
  • A nomogram was constructed using handcrafted features, deep learning features, and clinical data.

Main Results

  • The nomogram demonstrated excellent discriminative ability with an Area Under the ROC Curve (AUC) of 0.848 for the training set.
  • The model showed superior performance compared to clinical models and handcrafted radiomics alone.
  • The nomogram exhibited good calibration and clinical utility via decision curve analysis.

Conclusions

  • A validated deep learning radiomics nomogram can accurately predict NAC response in LAGC patients.
  • The nomogram, utilizing CECT images and clinical data, offers personalized treatment insights.
  • This approach supports tailored therapeutic strategies for LAGC patients undergoing surgical resection.

Keywords: