The Society of Radiologists in Ultrasound (SRU) provided sonographic criteria for the management of thyroid nodules in an article published in 2005 [
1]. The SRU is an American society, and the criteria were developed by a panel of experts from several medical disciplines, including radiology, endocrinology, cytopathology, and surgery. These guidelines attempted to define recommendations for nodules that should and should not undergo ultrasound-guided fine-needle aspiration (FNA). The SRU recommendations assert that FNA is appropriate for nodules that are 10 mm or larger and have microcalcifications, nodules that are 15 mm or larger and are solid or have coarse calcifications, nodules that are 20 mm or larger and are mixed solid and cystic, and nodules with substantial growth since the prior ultrasound study. Despite the guidelines, the number of thyroid nodules undergoing ultrasound-guided FNA has continued to increase because more thyroid nodules undergo biopsy as a result of fear of missing a malignancy [
2].
Many thyroid nodules undergoing sonographic workup are small incidental thyroid nodules detected on imaging. Although early diagnosis may be favorable in some malignancies, this may not be the case for thyroid cancer because the majority of the malignancies detected are small papillary carcinomas [
3]. Many experts believe that these small thyroid cancers represent “pseudodisease” and that most patients die with thyroid cancer rather than of thyroid cancer [
1,
3]. Thus, an ideal approach to selecting which incidental thyroid nodules should undergo workup would not be to diagnose all cancers but, rather, to “diagnose cancers that have reached clinical significance, while avoiding unnecessary tests and surgery in patients with benign nodules” [
1]. This statement is emphasized by the SRU recommendations.
Although the specific intent of the SRU recommendations was to reduce the unnecessary workup of thyroid nodules, there were concerns that the recommendations could in fact lead to an increase in the number of thyroid FNAs and subsequent thyroid surgeries [
1]. To date, the effect of the guidelines on thyroid FNA volume and malignancy rate and the acceptance of the guidelines in practice have not been evaluated. The aim of this study was to determine the proportion and type of thyroid nodules undergoing ultrasound-guided FNA at our institution that do not meet the SRU recommendations, which we refer to as “SRU-negative.” Our hypothesis was that many thyroid nodules undergoing FNA do not meet SRU recommendations and that SRU-negative malignancies are uncommon and are less aggressive.
Materials and Methods
Subjects and Pathologic Categorization
This study was approved by our institutional review board and was compliant with the HIPAA. We retrospectively reviewed the records of 400 consecutive ultrasound-guided thyroid FNA encounters through the department of radiology during a 12-month period from July 2010 through June 2011. An encounter was defined as presentation to the department of radiology on a given date for FNA of one or more thyroid nodules.
The criteria for selecting which nodule or nodules to biopsy were determined by the referring endocrinologists and endocrine surgeons, not the radiologist. In our institution the most common practice for diagnostic ultrasound is to report thyroid nodules 10 mm or larger in the impression section of the ultrasound report without adding a statement to recommend biopsy of the nodule unless there are highly suspicious features such as microcalcifications or hypoechogenicity.
The medical records of all included patients were reviewed for the age and sex of the patient, pathologic diagnosis, and management. The pathologic diagnosis was defined by a surgical pathologic result if available (
n = 87 patients) and an FNA cytopathologic result (
n = 263 patients) if the patient did not undergo surgery. At our institution, FNA cytopathology results are characterized by the Bethesda System for Reporting Thyroid Cytopathology [
4]. The Bethesda class categories include the following: class I, non-diagnostic or unsatisfactory; II, benign; III, atypia of undetermined significance or a follicular lesion of undetermined significance; IV, a follicular neoplasm or suspicious for a follicular neoplasm; V, suspicious for malignancy; and VI, malignant.
Nodules with FNA cytopathologic results revealing Bethesda class I, III, IV, or V are not regarded as definitive results, and the Bethesda System [
4] recommends repeat biopsy for classes I and III and diagnostic surgery for classes IV, V, and VI. A class II result is considered benign, and patients with a class II result do not undergo further biopsy or surgery. In this study, patients with a Bethesda class I, III, IV, or V result who did not undergo repeat FNA or surgery were excluded from the study group. All malignancies (Bethesda class VI) were further followed for the stage of disease, treatment response, and overall survival.
Application of SRU Recommendations
Diagnostic ultrasound (iU22 unit, Philips Healthcare) images of the thyroid nodules were obtained before the biopsy using a 12-MHz transducer. Thyroid nodules were measured on the ultrasound unit by the technologist or radiologist at the time of imaging and were documented in the examination report. These sizes were used, and the nodule was not measured retrospectively. A board-certified radiologist (7 years of experience) reviewed the ultrasound images on a PACS workstation for findings according to the SRU recommendations. SRU recommendations were met if the biopsied nodule had any of the following characteristics: size of 10 mm or larger with microcalcifications, size of 15 mm or larger with solid composition or coarse calcifications, size of 20 mm or larger with mixed solid-cystic composition, or substantial growth since the prior ultrasound. Because the SRU consensus statement does not explicitly define the requirements for “substantial growth,” a nodule was considered to show substantial growth if interval growth was the reason for the FNA and the nodule possessed no other criteria meeting SRU recommendations. Past ultrasound studies preceding the last diagnostic thyroid ultrasound were not reviewed specifically for substantial change.
Biopsy encounters were categorized on the basis of the sonographic findings as meeting the SRU recommendations for biopsy, which we refer to as “SRU-positive,” or not. We measured potential reduction of workup by the number of patient encounters rather than individual nodules because the greatest cost saving is in preventing any FNA rather than preventing FNA of a second or third thyroid nodule. In patients who had more than one nodule biopsied during an encounter (n = 120), the encounter was categorized as SRU-positive if any nodule met the SRU recommendations or as SRU-negative if none of the nodules met the recommendations. SRU-negative encounters comprise the subgroup of patients for whom workup with FNA could have been avoided if the SRU recommendations had been applied at the time of biopsy.
Statistical Analysis
Nodules biopsied in the SRU-positive and SRU-negative encounters were compared for size, malignancy rate, and characteristics of the malignancies (size, histologic diagnosis, and stage). The proportion of nodules not meeting the SRU recommendations was calculated as follows:
(SRU-negative nodules / total nodules) × 100.
The malignancy rates in the SRU-positive and SRU-negative groups were compared with the Fisher exact test. The sensitivity and specificity of the SRU recommendations for thyroid malignancy were also calculated.
Discussion
The SRU recommendations were developed by an expert group of radiologists to assist physicians in deciding which nodules seen on ultrasound should undergo FNA. One of the concerns of the SRU was the unknown effect of the guidelines on the workup rate of thyroid nodules and the possibility that they could increase, rather than decrease, FNA and subsequent surgery. This study shows that using the SRU criteria for ordering ultrasound-guided FNA at our institution could have eliminated one in four biopsy encounters compared with current practice at our institution. The malignancies diagnosed in nodules that did not meet SRU criteria had a less aggressive histologic diagnosis than the SRU-positive malignancies.
The SRU recommendations have been studied by Ahn et al. [
5], who applied the criteria to nodules undergoing ultrasound-guided biopsy in their institution in South Korea. In their study, nodules smaller than 10 mm were also biopsied, which is not standard practice for nodules in low-risk patients in the United States without other suspicious sonographic findings [
1,
6]. In our study, only 13 nodules were smaller than 10 mm. In their study [
5], Ahn et al. performed an analysis of a subset of nodules 10 mm or larger and found that 40% of nodules did not meet SRU recommendations (396 SRU-negative nodules / 996 total nodules). This proportion of SRU-negative encounters is much greater than in our study (24%) and likely reflects the differences in biopsy selection between our institutions. The performance of the SRU recommendations reported as sensitivity and specificity for malignancy also differed between our studies: Ahn et al. found sensitivity and specificity to be 72% and 42%, respectively, whereas we found a higher sensitivity of 83% but a lower specificity of 25%. Sensitivity and specificity are dependent on the prevalence of disease; the prevalence of malignancy in our study group was 8%, compared with 16% in the study by Ahn and colleagues. The malignancy rate in our study is more consistent with those of other studies, and a recent meta-analysis of more than 25,000 FNA biopsies reported 5.4% of nodules were classified as malignant (Bethesda class VI) and 8.1% were classified as suspicious for malignancy (Bethesda class V) or malignant (Bethesda class VI) [
7].
Multiple guidelines exist for the evaluation and workup of thyroid nodules based on sonographic criteria. In addition to the SRU guidelines, other major guidelines are from the National Comprehensive Cancer Network (NCCN), American Association of Clinical Endocrinologists (AACE), and American Thyroid Association (ATA) [
6,
8,
9]. These guidelines may be the guidelines that some of our endocrinologists and surgeons use in recommending biopsy. The NCCN guidelines are most similar to the SRU guidelines in that the size cutoff for solid nodules without suspicious features is 15 mm. Both the AACE and ATA recommend biopsy for solid nodules larger than 10 mm. In addition, the AACE recommends biopsy of any size nodule with one or more suspicious sonographic findings [
8]. Overall, the larger size cutoff of the SRU and NCCN recommendations would result in higher specificity (workup of fewer benign cases) than the AACE and ATA guidelines for our cohort of patients. As seen from our results, the specificity for the SRU recommendations is already very low (25%).
The cost of higher specificity for malignancy is a potentially higher false-negative rate and more missed malignancies. The false-negative rate in our study was 17% for the SRU recommendations, but all the SRU-negative malignancies were smaller than 15 mm and were localized to the thyroid. Cancers of this size, histologic type, and stage have excellent survival of greater than 99%, and several experts believe that many of these patients will die with—rather than of—thyroid cancer [
6,
10]. A contentious question for future research is whether treatment of small localized papillary cancers improves survival [
11]. Postmortem studies show that a certain number of small thyroid cancers will remain indolent in a patient's life, with up to 36% of patients having undiagnosed thyroid cancer at death [
12]. If the small SRU-negative cancers were undiagnosed, untreated, and progressed to a larger size, the survival could still be excellent. Ito et al. [
13] followed 340 papillary microcarcinomas that were selected to not receive treatment: After 10 years, there were new nodal metastases in 3% of cases and no cancer deaths [
13]. For these reasons, the SRU recommendations specifically aim to “diagnose cancers that have reached clinical significance, while avoiding unnecessary tests and surgery in patients with benign nodules” [
1]. Our study shows that SRU recommendations achieve this goal by missing only localized papillary carcinomas while reducing workup by one quarter.
There are several limitations to this study. First, this study is a retrospective study at one academic center conducted over a limited period of 12 months. Our results may not be generalizable to different practice types and referral patterns. Second, FNAs performed in the department of radiology do not include all thyroid FNAs at our institution. FNAs of thyroid nodules are also performed using ultrasound guidance by surgeons, endocrinologists, and pathologists, and these FNAs were not included. The nodules biopsied by clinicians and pathologists might have a higher proportion of SRU-negative nodules due to more liberal biopsy practices because of convenience of biopsy in a clinic, or the opposite could occur because clinic biopsies may be performed on larger nodules that are less technically challenging to biopsy. Third, ultrasound studies were reviewed by a single reader, and additional readers could yield slightly different results in the performance of SRU recommendations because of interreader variability. We also did not review all past ultrasound studies for the criterion of “substantial growth” because what constitutes substantial growth has not been specifically defined by the SRU. Finally, we applied the criteria based on the imaging findings and clinical information available. There could be other clinical information that was not available in the medical records that put the patients at higher risk of malignancy to warrant a biopsy.
In conclusion, the SRU recommendations achieve the goal of “diagnosing cancers that have reached clinical significance, while avoiding unnecessary tests and surgery in patients with benign nodules” [
1]. Using SRU recommendations can reduce workup of one in four thyroid biopsy encounters compared with current practice without strict guidelines. Potentially missed malignancies in SRU-negative nodules are less aggressive by histologic type, size, and stage compared with SRU-positive malignancies.