Objectives: Evaluate current accepted risk factors for well-differentiated thyroid carcinoma, and develop a predictive model to determine one's risk of malignancy given a thyroid nodule.
Study design: Retrospective analysis of 600 patients.
Subjects and methods: Patients with benign thyroid nodular disease and with well-differentiated thyroid cancer were randomly selected. Patient, clinical, and investigational data were compared by means of univariate and multivariate regression analyses.
Results: Age, regional lymphadenopathy, ipsilateral vocal cord palsy, solid and/or calcified nodules, and an aspiration biopsy being malignant or suspicious predicted for cancer (P < 0.05). Regional lymphadenopathy and vocal cord palsy are perfect predictors of malignancy. Multivariate analysis indicated age, solid and/or calcified nodules, and all fine-needle aspiration biopsy results to be significant in assessing risk (P < 0.05).
Conclusion: Taking individual risk factors in isolation is not always reliable. Using a predictive model, one can anticipate a patient's risk of malignancy when the diagnosis is unclear.