Background:
The risk of malignancy associated with thyroid nodules is ~5-15%. The Bethesda classification1 stratifies the risk based on fine needle aspiration (FNA) cytology and is used to guide management. However, false negatives remain a concern and is estimated between 1.3-11.5%. This study examined the accuracy of thyroid FNA by comparing the results with final histopathology, and evaluating the sensitivity, specificity and predictive values of FNA for the diagnosis of thyroid malignancy.
Methods:
Medical records of 449 patients who underwent FNA for thyroid nodules whom 187 were operated and have final pathological diagnosis were retrospectively reviewed. FNAs were classified according to the Bethesda classification. We calculated the malignancy risk for each category by follow up histopathology in all 187 cases that underwent subsequent surgeries at our institution.
Results:
Of the 550 FNAs performed, 187 cases proceeded to surgery (thyroidectomies or hemithyroidectomies). Malignancy rates at our institution were 21.05% for the non-diagnostic group; 10.0% for benign group, 44.44% for follicular lesion of undetermined significance (FLUS) group, 43.75% for the suspicious for follicular neoplasm group, 71.43% for the suspicious for malignancy group and 94.74% for the malignant group.
Sensitivity was 83.33%, specificity 71.29%, PPV 57.97%, NPV 90.0%, and diagnostic accuracy was 75.17%.
Conclusions:
Thyroid FNA has high sensitivity and specificity, but false negative and false positive results cause concern. It is difficult to calculate the true frequency of false negatives because only a small percentage of patients with benign FNA undergo surgery. Our findings do not match the published data. Our malignancy rate is higher for the benign group (10%) compared to published literature of 0-3% with a benign FNA result. This suggest that when making treatment recommendations and counselling patients, we should use data from our own institution in addition to published values.