A total of 93 patients met inclusion criteria with 72 females (77.4%) and 21 males (22.6%). Median age was 51.9 years with a range between 9 and 76. One patient had two suspicious lesions which were analyzed, resulting in 94 thyroid nodules available for statistical analysis.
shows the stratification of FNA results that were available for analysis. No patients had FNA reporting malignancy or
nondiagnostic/unsatisfactory per study design. There were 32 samples that were read as atypia or follicular lesion of undetermined significance and 26 samples read as follicular neoplasm or suspicious for follicular neoplasm. The rest were read as either benign or suspicious for malignancy (34 and 2 samples, resp.). Despite the 0–3% risk of malignancy for benign diagnostic category, there were additional indications for surgery including suspicious ultrasound findings, compressive symptoms, persistent growth, or even patient preference. There were four patients noted to have
nondiagnostic/insufficient FNA biopsies, one with a negative FS and positive final pathology (papillary carcinoma), two with FS suggesting positive although final pathology was negative, and one with both negative FS and final pathology. There were 2 patients who were diagnostic malignancy where their FS and final pathology were both positive (papillary carcinoma and metastatic melanoma).
Stratification of fine needle aspiration biopsies according to Bethesda system.
shows each patient and the correlation between the FNA Bethesda Criteria, FS, and final pathology. The sensitivity and specificity of FS were 76.9% and 67.9%, respectively. This is compared to the sensitivity and specificity for FNA, which were 53.8% and 74.1%, respectively. The PPV and NPV for FS were 27.8% and 94.8%, respectively. In comparison, FNA demonstrated a PPV of 25% and NPV of 90.9%. There were no changes in operative course as a consequence of a FS result.
Fine needle aspiration (FNA) Bethesda Criteria and frozen section (FS) compared with final pathology.
Our study did not demonstrate added information when intraoperative FS was used in patients with an FNA of benign, follicular neoplasm, suspicious for follicular neoplasm, or suspicious for malignancy reading. With this information, patients can be scheduled for a thyroid lobectomy with the knowledge that a diagnosis of malignancy will not be obtained until final pathologic analysis is completed. If malignancy is determined, a completion thyroidectomy can then be scheduled. The time and expense of intraoperative FS and scheduling operating room time for a total thyroidectomy, when it is not initially indicated, can therefore be spared.
The initiation of the Bethesda Criteria for Reporting Thyroid Cytopathology has greatly improved our ability to have accurate and meaningful conversations with our patients about their thyroid disease. In our case series, the sensitivity of FS was 76.9% compared to 53.8% for FNA, which is lower than previously published data [9
]. However, these publications were performed prior to the Bethesda Criteria and this may account for these differences. Even when evaluating only patients with a diagnosis of malignancy, Makay et al. observed FS to only have a 72% sensitivity [14
]. The PPV of both the FS and FNA biopsy were 27.8% and 25%, respectively. The challenges of FNA interpretation are well known and documented. Chang and colleagues [13
] showed that when there was discrepancy between the FNA and FS, the FS was shown to be more accurate (78.9% versus 21.1%).
Our results indicate that the addition of an FS does not allow for a more accurate and predictive result. There were no instances within our series where the FS altered the clinical course (i.e., conversion to a total thyroidectomy). There were no false positive results of malignancy for our FSs while there were two false negatives, one with a papillary microcarcinoma and the other a follicular carcinoma.
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; diagnosis may benefit from a more uniform language akin to that instituted for FNA biopsies. A standardized language may assist pathologists in their assessment as well as helping surgeons make decisions while in the operative theatre.