Nathaniel R. Alegre,
Joseph Leonard Z. Òbusan
January 1-December 31, 2016
Background: Proper assessment and evaluation is important in thyroid cancer because of the low mortality rates in the early stages of this disease. Multiple classifications have been developed in an attempt to separate thyroid nodules by risk of malignancy.
Objective: The purpose of this study was to determine the validity of one of these classifications, TIRADS, in predicting benign and malignant thyroid nodules on ultrasound in adult Filipino patient.
Methods: This was a single center retrospective study in which ultrasound images of patients who underwent ultrasound guided thyroid biopsies were categorized according to their characteristics as defined by the TIRADS method and were compared to their cytopathologic results. The data was then assessed and presented as sensitivity, specificity, positive predictive value, negative predictive value, likelihood ratios and area under the ROC curve.
Results: A total of 64 patients were included in the study. Using TIRADS 2-3 for benign and TIRADS 4-5 for malignant, the sensitivity, specificity, positive predictive value, negative predictive value, positive likelihood ratio and negative likelihood ratio were measured at 88.24%, 48.94%, 38.46%, 92.0% 1.73 and 0.24, respectively. The positive predictive values for the individual categories were 0%, 10%, 33%, 38% and 100% for TIRADS category 2, 3, 4A, 4B and 4C, respectively. Empiric AUC was calculated at 0.754. Cohen's kappa for intraobserver variability is at 0.796 for TIRADS 2-5 and 0.869 for TIRADS 2-3 versus 4-5.
Conclusion: TIRADS categorization is a viable way of assessing thyroid nodules. Gradually increasing value with higher TIRADS categories are consistent with previous studies. Slightly higher sensitivity of TIRADS categorization compared to individual suspicious characteristics should increase the rate of detection of thyroid cancer while still avoiding some unwarranted diagnostic interventions.