A risk model for predicting partial late radiation-induced injuries in T3-4 nasopharyngeal carcinoma based on dosimetric parameters and clinical factors
Abstract Objective This study aimed to construct a predictive risk model for late radiation injury in patients with T3-4 Nasopharyngeal Carcinoma (NPC) by integrating radiation dosimetric parameters and clinical features. The primary objective was to assess individual patient risk levels, enabling t...
Saved in:
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Springer
2025-08-01
|
| Series: | Discover Oncology |
| Subjects: | |
| Online Access: | https://doi.org/10.1007/s12672-025-03401-6 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Abstract Objective This study aimed to construct a predictive risk model for late radiation injury in patients with T3-4 Nasopharyngeal Carcinoma (NPC) by integrating radiation dosimetric parameters and clinical features. The primary objective was to assess individual patient risk levels, enabling timely adjustments to treatment strategies and facilitating early interventions to mitigate the occurrence of severe radiation-related complications. Methods Patients diagnosed with T3-4 NPC from January 2012 to December 2020 were retrospectively collected. The dosimetric characteristics of enrolled patients were obtained from DVH and the corresponding clinical characteristics were collected. LASSO regression and Random forest were used to screen for significant correlation factors affecting the occurrence of late radiation injury. ROC-AUC Curves will be used to evaluate the prediction ability of the models, and Calibration Curves will be used to evaluate the consistency of the model's prediction ability with the real situation. Decision Curves were used to evaluate the clinical application value of the two models. Results A total of 361 patients with T3-4 NPC were enrolled, and the patients were randomly divided into training group (N = 253) and test group (N = 108) by 7:3. Seven significant correlation factors were identified by Lasso regression. The AUC values of the model-lasso in the training group and the test group were 0.9126 (95%CI 0.889–0.924) and 0.8380 (95%CI 0.806–0.858), respectively. The AUC values of the model-rsf in the training group and the verification group were 0.917 and 0869, respectively. Conclusion In this study, a predictive risk model for late-radiation events of T3-4 NPC was successfully constructed. This model has a good prediction effect both in training cohort and test cohort. This model is expected to predict the risk of late radiation-induced injuries in advance and adjust the radiotherapy plan in time to reduce the occurrence of advanced radiation injury. |
|---|---|
| ISSN: | 2730-6011 |