The Accuracy of Manual Fine Needle Aspiration Cytology in the Evaluation of Thyroid Cancer

Abstract

Background The most common endocrine malignancy is thyroid carcinoma. Thyroid carcinoma constitutes for 0.5% of all cancer deaths.

In New Zealand, thyroid carcinoma is less than 1% of all cancer registrations. In New Zealand, between 1971 and 2008, the incidence increased from 0.8 to 2.2 per 100,000 in males and 2.5 to 5.6 per 100,000 in females. In this country, the five-year survival rate is 91%.

Fine needle aspiration cytology (FNAC) is an established technique for the investigation of thyroid nodules. The technique was pioneered in Sweden and has subsequently been adopted throughout the world. Manual FNAC is an inexpensive outpatient procedure which can provide a rapid and 99% accurate diagnosis. While FNAC is the principal screening and triage procedure in the evaluation of palpable thyroid nodules, there is controversy as to whether this should be performed under ultrasound guidance or as a purely manual technique.

Aim The objective of this study is to determine the accuracy of manual FNAC for the diagnosis of histologically proven thyroid malignancy in a community pathology practice over a 15-year period.

Study design This was a retrospective study. All cases of histologically confirmed thyroid abnormalities for the Otago and Southland regions were identified for the period of 15 years from 1995-2009. All the FNAC data available for these patients were reviewed blind to histological outcomes. Data were retrieved from the New Zealand cancer registry, the databases of the Southern Community Laboratory, Dunedin Hospital and Invercargill Hospital. The sensitivity, specificity, positive predictive value, negative predictive value and accuracy of FNAC were determined for the original diagnostic reports and the study diagnosis. Results A total of 242 cases were retrieved and reviewed in this study. FNAC results were diagnostic in 95% of cases, indeterminate in 2.5% and unsatisfactory in 2.5%. In this study, the false negative and false positive rates were (n=15) 6.6% and (n=5) 2.2% respectively.

Of the 141 histologically confirmed neoplastic cases, 122 were correctly identified as a neoplasm on FNAC and recommended for surgery. The 122 cases were papillary carcinoma (n=46), medullary carcinoma (n=3), anaplastic carcinoma (n=1), lymphomas (n=2) and suspicious for medullary carcinoma (n=1). Sixty-five of 76 Follicular neoplasm/ Hurthle cell neoplasm and four of six indeterminate cases reported on FNAC underwent surgery and histology confirmed neoplasm. Nineteen histologically confirmed neoplastic cases were diagnosed as benign (n=15) and unsatisfactory (n=4) on FNAC.

Of the 87 confirmed benign cases on histology, 80 were correctly diagnosed on FNAC. The remaining seven benign cases on FNAC were diagnosed as unsatisfactory (n=2), indeterminate (n=1), and follicular neoplasm (n=4).

Four of five cases of anaplastic carcinoma had no biopsy and were diagnosed solely on FNAC.

The 10 cases of micropapillary carcinoma were diagnosed as follicular neoplasm (n=5), Hurthle cell neoplasm (n=2), indeterminate for follicular lesion (n=1) and benign colloid nodules (n=2).

The statistical analysis was performed excluding micropapillary carcinoma (n=10) and unsatisfactory cases (n=6) considered indeterminate for follicular neoplasm, follicular neoplasm and suspicious as true positive. The overall sensitivity for thyroid neoplasm was 89.4% (95% CI=0.86, 0.92), specificity was 94.1% (95% CI= 0.88, 0.98), positive predictive value (PPV) was 96.2% (95% CI=0.92, 0.99), negative predictive value (NPV) was 84.2% (95% CI=0.79, 0.87) and accuracy was 91.0% (95% CI=0.87, 0.94). The sensitivity for diagnosing malignancy on FNAC was 97.6% and positive predictive value was 100%, if indeterminate for follicular lesions, follicular neoplasm and suspicious for malignancy considered as true positive. FNAC was highly sensitive in diagnosing papillary carcinoma of thyroid with 98.2%. A significant statistical correlation between FNAC and histopathological diagnosis was observed, p=0.000.

Significant difference was not observed between original and review cases in terms of sensitivity, specificity, PPV, NPV and accuracy. In this study, a review of 242 cases led to 13 (5.4%) discrepant to the original results on FNAC. This study showed 26 (10.7%) cases of disagreement between the first and second screener in the diagnosis of FNAC. Forty-six (19.3%) cases of histology diagnosis of the first screener differed from the second screener.

Conclusion The current study on manual FNAC was useful in differentiating thyroid lesions that required surgery. FNAC is cost-effective, minimally invasive and highly accurate for investigating patients with thyroid cancer.