Evaluation of Ultrasonographic, Cytological, Histopathological, and Clinical Findings in Childhood Thyroid Nodules in Terms of Malignancy: A Single-Center Cross-Sectional Analysis and a Six-Parameter Diagnostic Model Proposal
Discuss this preprint
Start a discussion What are Sciety discussions?Listed in
This article is not in any list yet, why not save it to one of your lists.Abstract
Objective: To assess ultrasonographic, cytological, histopathological and clinical features of thyroid nodules in children and to create a six parameter diagnostic model that has the ability to predict the malignancy with high precision. Design: A cross-sectional study with a retrospective design and based in one tertiary centre. Methods: Medical records of 250 children with thyroid nodules (September 2024- August 2025): 250 children with thyroid nodules screened and educated by the supervisor on correct nodule self-impression and the significance of routine medical examinations were reviewed. The variables analyzed were demographic, ultrasonographic, thyroid result tests, autoantibody status, FNAB outcome, and the histopathologic condition. Following the removal of non diagnostic nodules that had not been confirmed by surgery, 243 nodules were considered. The independent predictors of malignancy were identified using multivariate logistic regression and a six-parameter scoring model was developed. ROC and diagnostic accuracy indices were used to evaluate model performance. Results: Out of 243 assessable nodules, 18 (7.4%) were malignant and 225 (92.6) were benign. Male sex, solid echotexture, non-regular margins, longer-than-wide shape, calcification, larger size (10 or more), higher vascularity, and family history were significantly related to malignancy. Multivariate analysis revealed six independent predictors male sex (OR 152), solitary nodule (OR 59), calcification (OR 41.2), nodule size 1 cm (OR 61.9) and increased vascularity (OR 220.1) and family history of thyroid disease (OR 23.5). A scoring system giving one point to each parameter had shown very good discriminatory property (AUC 0.980; 95% CI: 0.9610.999; p<0.001). A score of ≥4 provided sensitivity of 94.1, specificity of 95.5, positive likelihood ratio of 20.89 and negative likelihood ratio of 0.062. The negative predictive value of the model was found to be 99.5 that depicts a high level of utility in the elimination of malignancy. Conclusions: Childhood thyroid nodules exhibit unique risk pattern of malignancy unlike that of adults. Male sex, isolated structure, calcification, greater size, greater vascularity and family history is a very strong diagnostic image. The proposed six-parameter model offers close to perfect differentiation of benign and malignant nodules and can be used to inform decisions on biopsy, minimise unnecessary operations and facilitate the use of risk based follow-up plans in the management of thyroid nodules in children. It should be externally validated in bigger and multicenter cohorts.