Related Experiment Video
Updated: Jul 10, 2026
Gene Regulation and Targeted Therapy in Gastric Cancer Peritoneal Metastasis: Radiological Findings from Dual Energy CT and PET/CT
Published on: January 22, 2018
18F-FDG PET/CT radiomics combined with double-expressor status enhances prognostication and high-risk stratification
Qi Jiang1, Silu Cui2, Panpan Luan2
1Department of PET/CT Center, Jiangsu Cancer Hospital and Jiangsu Institute of Cancer Research and The Affiliated Cancer Hospital of Nanjing Medical University, Nanjing, Jiangsu Province, China.
Background:
In the R-CHOP era, the IPI score has reduced efficacy in identifying high-risk diffuse large B-cell lymphoma (DLBCL) patients. Precise prognostic biomarkers and scoring systems are urgently needed to achieve refined stratification of this high-risk population. This study aimed to clarify the combined prognostic significance of 18F-fluorodeoxyglucose positron emission tomography/computed tomography (18F-FDG PET/CT) radiomics and double-expressor status in DLBCL patients (DEL) receiving first-line R-CHOP therapy, and to evaluate their incremental value over the the International Prognostic Index (IPI) score in identifying high-risk DLBCL patients and predicting their prognosis.
Methods:
This study included 210 real-world patients with newly diagnosed DLBCL treated with R-CHOP. Radiomic parameters including maximum standardized uptake value (SUVmax), total metabolic tumor volume (TMTV), total lesion glycolysis (TLG), and maximum lesion diameter (Dmax) were extracted from 18F-FDG PET/CT, with lesion segmentation performed at a 41% SUV threshold. With PFS as the primary endpoint and OS as the secondary endpoint, optimal cutoffs were identified by the max-stat method. Survival analysis and Cox regression were applied to construct a prognostic risk score model. Model validation was performed using five-fold cross-validation and 1000 bootstrap resamplings. Predictive performance was evaluated by AIC, c-index and AUC, with AUC differences compared via multiple statistical tests.
Results:
Multivariate analysis identified MTV, Dmax and DEL status as independent predictors of progression-free survival (PFS) and overall survival (OS), leading to the development of the MDE prognostic model. Five-fold cross-validation confirmed significant PFS/OS risk stratification across all training and validation cohorts (all P < 0.05). MDE was a significant independent PFS predictor (P < 0.0001), outperforming the IPI (P = 0.205). The MDE model showed superior predictive performance over IPI: for PFS, mean AIC (536.564 vs. 575.546) and C-index (0.778 vs. 0.677); for OS, mean AIC (424.227 vs. 445.86) and C-index (0.775 vs. 0.724). Its AUC (0.797) was significantly higher than IPI (0.700, all P < 0.05) per DeLong's test, bootstrap method and Venkatraman's permutation test. Compared with IPI high-risk patients (5-year PFS: 34.6%, OS: 32.6%), MDE identified a higher-risk subgroup (5-year PFS: 22.4%, OS: 29.7%) and improved discriminatory capacity across IPI subgroups (all P < 0.05).
Conclusions:
The MDE scoring system integrates baseline 18F-FDG-PET/CT radiomic features and DEL status. It identifies higher-risk DLBCL patients more effectively than the IPI, exhibits superior prognostic predictive value for treatment-naive DLBCL patients, and demonstrates excellent reproducibility.
Trial Registration:
The research protocol of this study was approved by the Ethics Committee of Jiangsu Cancer Hospital (Approval KY-2025-095, approval date: 16 July 2025).
Clinical Trial Number:
Not applicable.
