A novel combined model integrating collagen properties, radiomics and clinical data to predict gastric cancer prognosis.
[INTRODUCTION] Gastric cancer (GC) is one of the most common malignant tumors in the world, and there is still no good method to predict the prognosis of GC patients.
- p-value p = 0.013
- p-value p = 0.005
- 95% CI 1.02-1.18
- HR 1.1
APA
Xie Y, Zhong G, et al. (2026). A novel combined model integrating collagen properties, radiomics and clinical data to predict gastric cancer prognosis.. Frontiers in oncology, 16, 1801350. https://doi.org/10.3389/fonc.2026.1801350
MLA
Xie Y, et al.. "A novel combined model integrating collagen properties, radiomics and clinical data to predict gastric cancer prognosis.." Frontiers in oncology, vol. 16, 2026, pp. 1801350.
PMID
42038391
Abstract
[INTRODUCTION] Gastric cancer (GC) is one of the most common malignant tumors in the world, and there is still no good method to predict the prognosis of GC patients. Collagens are the major component of the extracellular matrix (ECM), and abnormalities in collagen are closely associated with the development of GC.
[METHODS] This study investigated collagen signatures in GC using bioinformatics analysis and picrosirius red staining. A total of 196 GC patients were included, with 138 assigned to the training group and 58 to the validation group. Multivariate analysis was performed to evaluate prognostic factors, and a predictive model was constructed by integrating collagen properties, radiomics, and clinical data.
[RESULTS] The study revealed that collagen signatures were associated with GC initiation and progression. Multivariate analysis confirmed that the picrosirius red risk score (HR = 1.1, 95% CI: 1.02-1.18, p = 0.013) and Radscore (HR = 2.34, 95% CI: 1.3-4.21, p = 0.005) were independent prognostic factors for overall survival. The combined model demonstrated high accuracy, with a C-index of 0.9 in the training group and 0.84 in the validation group, along with good goodness of fit and net benefit.
[DISCUSSION] These findings suggest that a prognostic model integrating multiple types of data, including collagen characteristics, radiomics, and clinical factors, can better predict the overall survival of GC patients.
[METHODS] This study investigated collagen signatures in GC using bioinformatics analysis and picrosirius red staining. A total of 196 GC patients were included, with 138 assigned to the training group and 58 to the validation group. Multivariate analysis was performed to evaluate prognostic factors, and a predictive model was constructed by integrating collagen properties, radiomics, and clinical data.
[RESULTS] The study revealed that collagen signatures were associated with GC initiation and progression. Multivariate analysis confirmed that the picrosirius red risk score (HR = 1.1, 95% CI: 1.02-1.18, p = 0.013) and Radscore (HR = 2.34, 95% CI: 1.3-4.21, p = 0.005) were independent prognostic factors for overall survival. The combined model demonstrated high accuracy, with a C-index of 0.9 in the training group and 0.84 in the validation group, along with good goodness of fit and net benefit.
[DISCUSSION] These findings suggest that a prognostic model integrating multiple types of data, including collagen characteristics, radiomics, and clinical factors, can better predict the overall survival of GC patients.
같은 제1저자의 인용 많은 논문 (5)
- Adverse events in facial plastic surgery: Data-driven insights into systems, standards, and self-assessment.
- Tumor immune-vascular crosstalk: synergy and translation of immune checkpoint inhibitors and anti-angiogenic agents in melanoma.
- Claudin18.2 positive gastric cancer: biology, tumor microenvironment, and therapeutic strategies.
- Characteristics and influencing factors of sleep disturbance in breast cancer patients: a cross-sectional study.
- Potential role of leptin in colorectal cancer liver metastasis involving lipid metabolic reprogramming and immunosuppressive macrophage polarization.