MRI-Based Radiomics-Clinical Model for Predicting Overall Survival in Lung Cancer Patients: A Single-Center Study
Technology in Cancer Research & Treatment · 2026-08 · 25
DOI: 10.1177/15330338261508400 ↗Kurz gefasst
Fachartikel in der Zeitschrift Technology in Cancer Research & Treatment, online erschienen am 14.11.2026. Verfasst von Jian Li und 9 weiteren Autor:innen. Beteiligte Einrichtung: Department of Radiotherapy, Harbin Medical University Cancer Hospital, Harbin, Heilongjiang, China. Hauptthemen laut MEDLINE: Radiomics, Lung Neoplasms, Magnetic Resonance Imaging, Carcinoma, Non-Small-Cell Lung. Die Arbeit nennt 32 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Themen: Onkologische Pflege
Abstract (Originalsprache)
Introduction
Despite advances in lung cancer management, overall survival (OS) remains limited due to tumor heterogeneity. This study aimed to evaluate the potential of pretreatment MRI-based radiomics, combined with clinical parameters, for prognostic assessment in non-small cell lung cancer (NSCLC).
Methods
This retrospective, single-center study included 70 consecutive patients with NSCLC who underwent pretreatment MRI and were randomly divided into a training cohort (n = 49) and an internal validation (n = 21) cohort. Radiomic features were extracted and filtered for reproducibility (ICC ≥ 0.75), followed by feature selection using univariate Cox regression and 10-fold cross-validated least absolute shrinkage and selection operator Cox (LASSO-Cox) regression within the training cohort. The training-derived Radiom model was evaluated in both cohorts. In exploratory full-cohort analyses, a clinical-only model (Clinical), the Radiom model (Radiomics), and a combined clinical-radiomics model (ClinRadiom) were compared using the concordance index (C-index), time-dependent area under the curve (AUC), calibration, decision-curve analysis, and 1,000-resample bootstrap internal validation.
Results
Among the 70 patients, 47 deaths occurred and 23 observations were censored. The median follow-up was 2.98 years (95% CI, 2.44-4.87). Seven radiomic features were retained to construct the Radiom score. The Radiom model achieved C-indices of 0.722 and 0.671 in the training and internal validation cohorts, respectively, with 1-, 2-, and 3-year AUCs of 0.653, 0.848, and 0.774 in the internal validation cohort. In exploratory full-cohort comparisons, ClinRadiom showed the highest apparent C-index (0.743; 95% CI, 0.662-0.823) and an optimism-corrected C-index of 0.733, with 1-, 2-, and 3-year AUCs of 0.700, 0.921, and 0.889, respectively. Adding Radiom to the Clinical model significantly improved model fit (likelihood-ratio p < 0.001).
Conclusion
The Radiom and ClinRadiom models showed potential for OS risk stratification in NSCLC. MRI-based radiomics may provide complementary non-invasive prognostic information, although validation in larger multicenter cohorts is required.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Department of Radiotherapy, Harbin Medical University Cancer Hospital, Harbin, Heilongjiang, China.
- Department of Radiotherapy, Hainan Cancer Hospital, Haikou, Hainan, China.
- Department of Thoracic Surgery, Harbin Medical University Cancer Hospital, Harbin, Heilongjiang, China.
ORCID-Profile: Lei Yao
Zitieren
APA
Li, J., Hu, Y., Wang, C., Liu, H., Zhang, Z., Hu, S., Ma, J., Liu, S., Xu, J., & Yao, L. (2026). MRI-Based Radiomics-Clinical Model for Predicting Overall Survival in Lung Cancer Patients: A Single-Center Study. Technology in Cancer Research & Treatment, 25. https://doi.org/10.1177/15330338261508400
Vancouver
Li J, Hu Y, Wang C, Liu H, Zhang Z, Hu S, et al. MRI-Based Radiomics-Clinical Model for Predicting Overall Survival in Lung Cancer Patients: A Single-Center Study. Technology in Cancer Research & Treatment. 2026;25. doi:10.1177/15330338261508400
BibTeX
@article{Li2026,
author = {Li, Jian and Hu, Yuefeng and Wang, Chengcheng and Liu, Hongkai and Zhang, Zijie and Hu, Songliu and Ma, Jianli and Liu, Shilong and Xu, Jianyu and Yao, Lei},
title = {MRI-Based Radiomics-Clinical Model for Predicting Overall Survival in Lung Cancer Patients: A Single-Center Study},
journal = {Technology in Cancer Research & Treatment},
year = {2026},
volume = {25},
doi = {10.1177/15330338261508400}
}
EUVIMED