CT semantic features versus first-order histogram analysis for preoperative prediction of Lauren classification in gastric cancer
European Journal of Surgical Oncology · 2027-01 · 53 (1) · 112342
DOI: 10.1016/j.ejso.2026.112342 ↗Kurz gefasst
Fachartikel in der Zeitschrift European Journal of Surgical Oncology, 2027-01. Verfasst von Shuxiang Chen und 7 weiteren Autor:innen. Beteiligte Einrichtung: Department of Radiology, Shengli Clinical Medical College of Fujian Medical University, Fujian Provincial Hospital, Fuzhou University Affiliated Prov…. Die Arbeit nennt 29 Literaturangaben. Gefördert von: Fujian Provincial Department of Science and Technology.
Abstract (Originalsprache)
OBJECTIVES: Accurate preoperative prediction of gastric cancer Lauren subtypes is critical for individualized treatment decision-making. This study aimed to explore and compare CT-based semantic features, first-order histogram analysis, and their combined model for non-invasive discrimination of intestinal and diffuse Lauren subtypes of gastric cancer. METHODS: This retrospective study enrolled 140 patients with gastric cancer, who were randomly assigned to a training set (n = 98) and an internal validation set (n = 42). The collected variables included clinical data, conventional CT indicators (e.g., peritoneal metastasis, tumor size, and thickness), and first-order histogram features extracted from the primary tumor and peritoneum (e.g., mean, entropy, skewness). The endpoint was Lauren classification (intestinal vs. diffuse type). Three predictive models, including a CT semantic model, a first-order histogram radiomic model, and a combined model, were established using LASSO logistic regression. Model performance was comprehensively evaluated in terms of discriminative ability, calibration, and clinical net benefit. RESULTS: All three models achieved robust discriminative performance. The first-order radiomic model yielded the highest training-set AUC (0.904), followed by the combined model (0.879) and CT semantic model (0.844). Bonferroni-corrected pairwise comparisons showed no significant inter-model differences, demonstrating comparable predictive efficacy. Owing to its non-inferior performance and superior clinical accessibility, the CT semantic model was visualized as a nomogram for individualized preoperative Lauren subtype assessment. While calibration was satisfactory in the training cohort, mild calibration drift with a reduced slope and increased maximum calibration error was observed in internal validation, though discriminative power remained acceptable. DCA confirmed favourable clinical net benefit across a wide range of risk thresholds in both training and validation cohorts. CONCLUSIONS: Both the CT semantic features model and the first-order histogram model enable reliable preoperative Lauren subtype prediction. Featuring higher accessibility, the CT semantic features model needs no extra radiomic analysis. Though exploratory, these non-invasive tools may enable preoperative subtype identification and individualized treatment planning.
Schlagwörter
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Department of Radiology, Shengli Clinical Medical College of Fujian Medical University, Fujian Provincial Hospital, Fuzhou University Affiliated Provincial Hospital, Fuzhou, Fujian, 350001, China. Electronic address: 06…
- The School of Basic Medical Sciences, Fujian Medical University, Fuzhou, Fujian, 350122, China.
- Department of Breast Surgery, Shengli Clinical Medical College of Fujian Medical University, Fujian Provincial Hospital, Fuzhou University Affiliated Provincial Hospital, Fuzhou, Fujian, 350001, China.
- Department of Obstetrics and Gynecology, Shengli Clinical Medical College of Fujian Medical University, Fujian Provincial Hospital, Fuzhou University Affiliated Provincial Hospital, Fuzhou, Fujian, 350001, China.
- Department of Radiology, Shengli Clinical Medical College of Fujian Medical University, Fujian Provincial Hospital, Fuzhou University Affiliated Provincial Hospital, Fuzhou, Fujian, 350001, China.
- Department of Endoscopy, Shengli Clinical Medical College of Fujian Medical University, Fujian Provincial Hospital, Fuzhou University Affiliated Provincial Hospital, Fuzhou, Fujian, 350001, China.
Förderung
- Fujian Provincial Department of Science and Technology
Zitieren
APA
Chen, S., Cai, D., Chen, S., Chen, Y., Xiao, G., Zhang, H., Tong, Y., & Zhou, Y. (2027). CT semantic features versus first-order histogram analysis for preoperative prediction of Lauren classification in gastric cancer. European Journal of Surgical Oncology, 53(1), 112342. https://doi.org/10.1016/j.ejso.2026.112342
Vancouver
Chen S, Cai D, Chen S, Chen Y, Xiao G, Zhang H, et al. CT semantic features versus first-order histogram analysis for preoperative prediction of Lauren classification in gastric cancer. European Journal of Surgical Oncology. 2027;53(1):112342. doi:10.1016/j.ejso.2026.112342
BibTeX
@article{Chen2027,
author = {Chen, Shuxiang and Cai, Dingsheng and Chen, Shuo and Chen, Yifan and Xiao, Guiqing and Zhang, Huijuan and Tong, Yongxiu and Zhou, Yishen},
title = {CT semantic features versus first-order histogram analysis for preoperative prediction of Lauren classification in gastric cancer},
journal = {European Journal of Surgical Oncology},
year = {2027},
volume = {53},
number = {1},
pages = {112342},
doi = {10.1016/j.ejso.2026.112342}
}
EUVIMED