Thyroid nodules represent a significant problem in the current clinical and surgical practice [1]. Due to their high incidence and malignant potential, careful evaluation is required in order to adopt optimal management and spare unnecessary surgery. FS biopsy of a thyroid nodule allows an intraoperative pathological diagnosis at the time of operation which may guide surgical strategy. For malignant nodules, total thyroidectomy should be performed to avoid patient the risk of a second surgery which increase disease spread, morbidity and complication rates. However In case of benignity, FS could alter operative management from aggressive to conservative surgery [2]. The aim of this study was to review the place of FS in thyroid surgery and to assess its value in the management of patients with thyroid disease.

The pathologic reports issued for all thyroid surgery specimens, submitted for intraoperative pathology examination at the Department of Pathology of Habib Thameur Hospital during the period from January 2003 to December 2012 were reviewed retrospectively. During this period, the results of fine needle aspiration (FNA) were not always available to pathologists at the time of intraoperative consultation. FS examination was performed on all thyroid specimens (thyroidectomy, lobectomy and loboisthemctomy). Each of the specimens was examined by any one of seven staff pathologists who selected areas for microscopic FS examination after gross inspection of the specimens. In most cases one block of tissue was examined. The tissue blocks were frozen in a cryostat at -20°C and followed by hematoxylin-eosin staining. Frozen tissues were then fixed in 10% neutral and processed for paraffin section. Touch preparations were routinely performed. In cases where the lesion was clearly benign or clearly malignant the surgeon was thus informed and was also given any more specific diagnostic information which could be provided. Pathologic reports were analyzed in regard to FS results final histopathology result and tumor size. Frozen-section analysis results for the thyroid gland are coded as malignant for all thyroid cancers, lymphoma, or metastatic diseases. Goiter, hyperplasia, thyroiditis, and adenoma are considered benign. Reports of using ‘‘favor’’ or ‘‘possible’’ for any malignancy or suspicious for capsular invasion, follicular neoplasm are coded as deferred (defer reporting results until permanent sections are reported). Definite pathologic categorization is based on permanentsection diagnosis. Permanent pathologic diagnoses were coded as malignant for any report of papillary, follicular, Hürthle cell, medullary, or anaplastic thyroid carcinoma. Micro carcinomas were included with papillary cancers. Lymphoma and metastatic disease involving the thyroid gland were coded as malignant. The maximum diameter of the tumor or the largest of multiple tumors was recorded for all malignant lesions. All remaining diagnoses were recorded as benign, including goiter, adenoma, Hürthle cell adenoma, hyperplasia, and thyroiditis. All histological diagnoses were performed in accordance with WHO guidelines.

The results were considered as follows:

A true positive (TP) is a positive result for malignancy with subsequent final histopathologic verification. A true negative (TN) is a negative result for malignancy with subsequent final histopathologic confirmation. A false positive (FP) is a positive result for malignancy at FS and benign results for the final histopathologic diagnosis. A false negative (FN) is a reading consistent with benign disease and malignant on final histopathologic diagnosis.

Sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy for detection of malignancy are determined for frozen-section analysis according to the following equations: sensitivity=TP/(TP+FN); specificity=TN/(TN+FP); PPV=TP/(TP+ FP); NPV=TN/(TN+FN); and diagnostic accuracy=(TP+TN)/total cases.

Intraoperative FS examinations of the thyroid were performed in 1110 thyroid specimens. Of these, the diagnosis was deferred in 102 cases (9%), positive for malignancy in108cases (9.7%), and benign in 902 cases (81.3%). Final pathology results revealed 920 benign lesions (82.9%) of all and 190 carcinomas (17.1%). On final histologic diagnosis, from the 902 benign diagnoses of the FS, 847 (93.3%) were TN and 55 (4.95%) were FN. 50% of FN were represented by papillary micro-carcinomas (tumor size<1 cm) which were accidentally discovered. The 108 malignant results of the FS were represented by 102 (94.4%) TP and 6 (0.55%) FP. The 102 TP were represented by 87 papillary carcinomas (PC), 5 follicular carcinomas (FC), 5 anaplastic carcinomas (AC), 4 medullary carcinomas (MC) and one poorly differentiated carcinoma. The 100 deferred results revealed 67 (67%) benign lesions and 33 (33%) carcinomas with 28 PC, 4 FC with microinvasion and 1 poorly differentiated carcinoma. In our series, FS and final histopathological diagnosis agreed in 85.4% and disagreed in 5.5%. 9.1% of the cases were deferred (Table 1).

Table 1: Comparison of FS diagnoses with final histologic diagnoses.

The sensitivity/specificity pairs of FS analysis for allhistological subtypes were 64.7%/99.3%, dropping to 53.1%/92% with calculation including deferred responses. The specificity for FS analysis for carcinoma diagnoses were unmoved with the histological subtype, however its sensitivity for papillary carcinoma was 61.7%, 83.3% for follicular carcinoma and 100% for anaplastic carcinoma. The positive predictive value of intraoperative FS result was 94.4% and the negative predictive value one was 93.9% with an accuracy of 85.4% (p<0.001).

Generally, FS is considered to have enhanced specificity for diagnosing thyroid cancers [2]. Table 2 shows the different studies regarding the results obtained with the FS on the assessment of thyroid nodules. The sensitivity, specificity, PPV, NPV and accuracy of FS reported in the literature vary from 50% to 99%, 92% to 100%, 87% to 100%, 75% to 98.5%, and 69% to 98%, respectively. Our findings match those in the literature, with good accuracy.

Table 2: FS results in the assessment of thyroid nodules.

In our study, 4.95% of cases diagnosed by FS as benign lesions were FN. This result can be explained by a high rate of papillary microcarcinomas diagnosed only on final exam (50% of all FN). Malignant frozen-section diagnosis obtained 0.55% of FPs. Our findings indicate that FS is reliable in diagnosing benign and malignant tumors. Conversely, in 3.1% of those diagnosed as deferred frozen-section results, permanent-section examinations confirmed malignancy.

Usually, pathological diagnosis of papillary carcinoma in the thyroid is easy when compared follicular carcinoma. This is due to nuclear atypia which are specific of papillary carcinoma. However, our study showed opposite results, in fact we found that sensitivity of FS for papillary carcinoma is lower than that of follicular carcinoma. These particular and controverting results are due a high rate of follicular variety of papillary carcinoma in our series which diagnoses were challenging and habitually made after final histopathological examination. In the other hand, the major part of follicular carcinoma diagnosed were massively invasive making diagnosis easier. These are main reasons to explain lack of sensitivity of FS of papillary carcinoma compared to follicular carcinoma.

The true utility of this practice has long been debated. Proponents for routine use of intraoperative FS have demonstrated cost effectiveness and a reduced number of completion thyroidectomies. Others disagree on the basis that increased costs from additional operating time and pathologists needed to read the specimens are not justified by a significant benefit in patient outcomes [28]. However, FS is usually insufficient to determine true capsular or vascular invasion and deferral to a final pathologic diagnosis is often necessary in the setting of a follicular lesion [29]. Currently, preoperative fine needle aspiration (FNA) and intraoperative FS examination should be regarded as complementary procedures which, in some patients, may provide more diagnostic information than either procedure alone [8]. FS analysis often complements the fine needle aspiration cytology in cases with cytologic atypia or suspicious appearance, and is most useful at making a diagnosis of papillary thyroid carcinoma [28]. For FNA indeterminate specimens, a diagnostic thyroid lobectomy is often performed with intra-operative FS analysis in order to prevent a return dtrip to the operating room for completion thyroidectomy.

Our data supports the clinical usefulness of FS for the diagnosis and management of thyroid nodules. Future directions for the diagnosis and treatment of thyroid nodules should include addressing highly variable language used among pathologists in regard to thyroid intraoperative FS reporting. A standardized language may assist pathologists in their assessment as well as helping surgeons make decisions while in the operative theatre.