Related Experiment Video
Updated: Sep 13, 2025
Predictive Immune Modeling of Solid Tumors
Published on: February 25, 2020
PET/CT-Based prognostic model enhances early survival prediction in angioimmunoblastic t-cell lymphoma
Xiaoxia Xu1, Xiangxiang Ding1, Xiao Jin1
1Key laboratory of Carcinogenesis and Translational Research (Ministry of Education), Beijing Key Laboratory of Research, Investigation and Evaluation of Radiopharmaceuticals, NMPA Key Laboratory for Research and Evaluation of Radiopharmaceuticals (National Medical Products Administration), Department of Nuclear Medicine, Peking University Cancer Hospital & Institute, Beijing, China.
Background:
To develop and validate a new prognostic model using baseline PET parameters and clinical indicators for predicting early overall survival (OS) in Angioimmunoblastic T-cell lymphoma (AITL) patients.
Methods:
We conducted a retrospective cohort study from December 2009 to December 2023 (n=124) at a single center. The model's predictors included baseline clinical characteristics, pathological indicators, laboratory metrics, and PET/CT parameters. Independent prognostic factors were identified using Cox regression and presented as nomograms. The C-index assessed predictive accuracy, while calibration plots and decision curve analysis evaluated prediction accuracy and discrimination ability. The model's accuracy was compared with existing prognostic systems using C-index, NRI, ROC, and Kaplan-Meier survival curves.
Results:
SUVmax, β2MG, platelet, and albumin were identified as independent risk factors. The C-index for OS was 0.78 (95% CI: 0.70-0.85); for 1000 bootstrap samples, it was 0.76 (95% CI: 0.61-0.93). Calibration curves showed excellent agreement between predictions and actual observations. The AUC for 6-month and 1-year OS were 0.91(95% CI: 0.82-1.00) and 0.85 (95% CI: 0.77-0.94), respectively. The model outperformed PIAI, IPI, and PIT in predictive capacity.
Conclusion:
The new prediction model reliably estimates outcomes for AITL patients, demonstrating high discrimination and calibration.
