Background: This study explored the relationship between thyroid-associated antibodies, immune cells, and hypothyroidism to establish a predictive model for the incidence of hypothyroidism in patients with nasopharyngeal carcinoma (NPC) after radiotherapy.
Methods: A total of 170 patients with NPC treated at the Cancer Hospital of University of Chinese Academy of Sciences between January 2015 and August 2018 were included. The complete blood count, biochemical, coagulation function, immune cells, and thyroid-associated antibodies tested before radiotherapy were evaluated. A logistic regression model was performed to elucidate which hematological indexes were related to hypothyroidism development. A predictive model for the incidence of hypothyroidism was established. Internal verification of the multifactor model was performed using the tenfold cross-validation method.
Results: The univariate analysis showed that immune cells had no statistically significant differences among the patients with and without hypothyroidism. Sex, N-stage, antithyroid peroxidase antibody (TPO-Ab), antithyroglobulin antibody (TG-Ab), thyroglobulin (TG), and fibrinogen (Fb) were associated with hypothyroidism. Males and early N-stage were protective factors of thyroid function, whereas increases in TPO-Ab, TG-Ab, TG, and Fb counts were associated with an increased rate of hypothyroidism incidence. The multivariate analysis showed that TPO-Ab, TG-Ab, TG, and Fb were independent predictors of hypothyroidism. The comprehensive effect of the significant model, including TPO-Ab, TG-Ab, TG, and Fb counts, represented the optimal method of predicting the incidence of radiation-induced hypothyroidism (AUC = 0.796). Tenfold cross-validation methods were applied for internal validation. The AUCs of the training and testing sets were 0.792 and 0.798, respectively.
Conclusion: A model combining TPO-Ab, TG-Ab, TG, and Fb can be used to screen populations at a high risk of developing hypothyroidism after radiotherapy.
Download full-text PDF |
Source |
---|---|
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7243020 | PMC |
http://dx.doi.org/10.1155/2020/3860936 | DOI Listing |
Enter search terms and have AI summaries delivered each week - change queries or unsubscribe any time!