Abstract | Background: Methods: Data was collected from 174 patients with clinicopathologically confirmed TSESCC from October 2013 to June 2020. Procalcitonin (PCT), C-reactive protein (CRP), and interleukin-6 (IL-6) levels in serum were dynamically monitored during radiotherapy. Lasso analysis was used for feature screening before multivariate logistic regression analysis to reduce the multicollinearity of variables. A nomogram combined with biological factors and clinical signs for individualized risk assessment and precise prediction of RP was developed and assed the performance with respect to its calibration, discrimination. Results: Of the 174 patients, 30 patients developed RP (grade ≥2) while 144 patients did not. After variable screening by Lasso analysis and logistics multivariate regression analysis, the predictor variables that were finally retained in the nomogram prediction model included IL-6, CRP, and radiotherapy techniques. The model displayed good discrimination with an area under the curve (AUC) of 0.898 (95% CI: 0.849-0.947), with the sensitivity and specificity of 0.967 and of 0.736, respectively. This model also shows good calibration and clinical practical value. In addition, the study provided a web-based version of the dynamic nomogram to facilitate clinical application. Conclusions:
|
Authors | Ting Qiu, Yuxia Deng, Haochun Guo, Haijun Zhang |
Journal | Translational cancer research
(Transl Cancer Res)
Vol. 11
Issue 10
Pg. 3754-3766
(Oct 2022)
ISSN: 2219-6803 [Electronic] China |
PMID | 36388040
(Publication Type: Journal Article)
|
Copyright | 2022 Translational Cancer Research. All rights reserved. |