Prediction of Claudin18.2 expression in gastric adenocarcinoma by habitat radiomics based on spectral CT
Abstract
To explore the predictive value of habitat radiomics derived from Dual-layer spectral detector CT (DLSCT) for the preoperative evaluation of Claudin18.2 (CLDN18.2) expression in gastric adenocarcinoma. A retrospective cohort of 111 eligible patients with gastric adenocarcinoma who underwent preoperative dual-phase DLSCT examinations was randomly divided into training and test sets at a 7:3 ratio. On the basis of the DLSCT images, the habitat subregions were delineated by applying K-means clustering to the voxel intensity values. Habitat features were extracted and filtered from these subregions to construct a habitat radiomics model. A clinical model, a conventional radiomics model, and a combined habitat radiomics and clinical data nomogram model were subsequently constructed. Model performance was assessed using the receiver operating characteristic (ROC) curve area under the curve (AUC), calibration curves, and decision curves. The AUC values for the nomogram model in the training and test sets were 0.802 and 0.757, respectively, which were superior to those of the clinical model (AUC: 0.643 and 0.604, respectively), conventional radiomics model (AUC: 0.713 and 0.716, respectively) and habitat radiomics model (AUC: 0.768 and 0.726, respectively). The calibration curves demonstrated good consistency between the predicted results and actual outcomes. The results of the decision curve analysis(DCA) indicated that the model has high clinical applicability. The habitat radiomics model based on DLSCT outperforms the traditional radiomics model. Furthermore, integrating habitat radiomics with clinical indicators yields promising non-invasive predictive performance for preoperative CLDN18.2 stratification in gastric adenocarcinoma; large-scale prospective multicenter validation is required prior to clinical application.
Related
- Nodule subtype dependency of CT-based radiomics for preoperative prediction of spread through air spaces in stage I lung adenocarcinoma: a multi-centre study
- MRI-derived habitat radiomics integrated with transformer deep learning for preoperative prediction of high CD9 expression: a novel prognostic marker in adult-type diffuse gliomas
- Radiomics modeling to predict the tumor immune-microenvironment of mucinous adenocarcinoma of gastric-type cervical cancer
- CT-based multi-station lymph-node radiomics for preoperative prediction of lymph node metastasis in colorectal cancer: a proof-of-concept comparison with primary-tumor radiomics
- Habitat-based ultrasound radiomics for preoperative prediction of axillary lymph node metastasis in breast cancer: a two-center study
BibTeX
@article{chen2026prediction,
title = {{Prediction of Claudin18.2 expression in gastric adenocarcinoma by habitat radiomics based on spectral CT}},
author = {Jinghua Chen and Xinyu Pan and Huachun Ma and Chen Tang and Lanlan Lu and Xu Yan and Jian Huang and Xiaoyu Tang and Tingting Qian and Kefu Liu},
journal = {Scientific Reports},
year = {2026},
month = oct,
doi = {10.1038/s41598-026-73281-5},
url = {https://www.nature.com/articles/s41598-026-73281-5},
note = {CC BY-NC-ND 4.0},
}
RIS
TY - JOUR
TI - Prediction of Claudin18.2 expression in gastric adenocarcinoma by habitat radiomics based on spectral CT
AU - Jinghua Chen
AU - Xinyu Pan
AU - Huachun Ma
AU - Chen Tang
AU - Lanlan Lu
AU - Xu Yan
AU - Jian Huang
AU - Xiaoyu Tang
AU - Tingting Qian
AU - Kefu Liu
PY - 2026
DA - 2026/10/02
JO - Scientific Reports
DO - 10.1038/s41598-026-73281-5
UR - https://www.nature.com/articles/s41598-026-73281-5
AB - To explore the predictive value of habitat radiomics derived from Dual-layer spectral detector CT (DLSCT) for the preoperative evaluation of Claudin18.2 (CLDN18.2) expression in gastric adenocarcinoma. A retrospective cohort of 111 eligible patients with gastric adenocarcinoma who underwent preoperative dual-phase DLSCT examinations was randomly divided into training and test sets at a 7:3 ratio. On the basis of the DLSCT images, the habitat subregions were delineated by applying K-means clustering to the voxel intensity values. Habitat features were extracted and filtered from these subregions to construct a habitat radiomics model. A clinical model, a conventional radiomics model, and a combined habitat radiomics and clinical data nomogram model were subsequently constructed. Model performance was assessed using the receiver operating characteristic (ROC) curve area under the curve (AUC), calibration curves, and decision curves. The AUC values for the nomogram model in the training and test sets were 0.802 and 0.757, respectively, which were superior to those of the clinical model (AUC: 0.643 and 0.604, respectively), conventional radiomics model (AUC: 0.713 and 0.716, respectively) and habitat radiomics model (AUC: 0.768 and 0.726, respectively). The calibration curves demonstrated good consistency between the predicted results and actual outcomes. The results of the decision curve analysis(DCA) indicated that the model has high clinical applicability. The habitat radiomics model based on DLSCT outperforms the traditional radiomics model. Furthermore, integrating habitat radiomics with clinical indicators yields promising non-invasive predictive performance for preoperative CLDN18.2 stratification in gastric adenocarcinoma; large-scale prospective multicenter validation is required prior to clinical application.
SN - 2045-2322
ER -