Risk of malignancy and clinical outcomes of cyst fluid only nodules in the thyroid based on ultrasound and aspiration cytology

Abstract Background The number of extensive studies focusing on cyst fluid only (CFO) thyroid nodules is limited, and the risk of malignancy (ROM) in CFO nodules has not been well‐established. Thus, the purpose of this study was to investigate CFO nodules using cytology and ultrasound. In addition, we sought to define the ROM and determine the recommended clinical management of CFO nodules. Methods We retrospectively reviewed cytological preparations of 678 nodules that were originally identified as CFO nodules, including conventional specimens in 209 nodules, liquid based cytology (LBC) specimens in 221 nodules, and both conventional and LBC specimens in 248 nodules. Ultrasound reports with representative photographs were also reviewed. Results Of the 678 CFO nodules, 214 (31.6%) were reclassified into other categories, including non‐diagnostic/unsatisfactory (ND/UNS) except for CFO (n = 15), benign (n = 198), and malignant (n = 1). Conventional preparations (33.5%) were more frequently reclassified than LBC preparations (13.6%; P < .0001). Re‐aspiration for diagnosis was performed for only one calcified nodule. The rates of surgical resection and malignancy were 3.0% and 0.2%, respectively. Based on American Thyroid Association guidelines and the Kuma Hospital ultrasound classification, worrisome sonographic features were identified in 5.8% and 0% of CFO nodules, respectively. Conclusion We propose that CFO nodules should be classified as separate from ND/UNS nodules; they should be categorized as a subtype of benign nodules. However, it is essential that fine‐needle aspiration cytology be performed under ultrasound‐guided real‐time visualization of needle placement in the target nodule in all cases.


| INTRODUCTION
According to The 2017 Bethesda System for Reporting Thyroid Cytopathology (TBSRTC), 1 the diagnostic categories for thyroid nodules are classified as "non-diagnostic or unsatisfactory (ND/UNS)," "benign," "atypia of undetermined significance (AUS)," "follicular neoplasm (FN)," "suspicious for malignancy (SFM)," and "malignant," and the risk of malignancy (ROM) and recommended clinical management are specific to each category. ND/UNS applies to specimens that are unsatisfactory owing to the following: (a) fewer than six groups of well-preserved, well-stained follicular cell groups with 10 cells each; (b) poorly prepared, poorly stained, or obscured follicular cells; or (c) cyst fluid, with or without histiocytes, and fewer than six groups of 10 benign follicular cells. As the possibility of cystic papillary thyroid carcinoma cannot be excluded in the last scenario, namely cyst fluid only (CFO) nodules, the cases are classified as ND/UNS. 1,2 However, in Japanese Society of Thyroid Surgery (RSJSTS) 2015 reporting system, CFO nodules were not categorized under ND/UNS but as an independent diagnostic category. 3 The clinical value of CFO nodules depends largely on sonographic correlation. 1,4 When the nodule is a simple cyst without worrisome sonographic features, an endocrinologist might proceed as if it is a benign nodule. Alternatively, when the sonographic features suggestive of malignancy, the endocrinologist is not convinced that the sample is representative, and it might be clinically ND/UNS. 5

| MATERIALS AND METHODS
We reviewed a database of the cytology reports of 7295 patients that underwent thyroid fine-needle aspiration cytology (FNAC) at Kuma Hospital from January to December 2017. During this period, cytology specimens from 9767 thyroid nodules were available. Of these, 678 (6.9%) were reported as CFO nodules, based on the criteria of TBSRTC. 1 Most of the aspirations were performed using a 22-gauge needle, and an 18-gauge needle was used to drain cyst fluid. The procedure was done under ultrasound-guided real-time visualization of needle placement in the target nodule in all cases. The aspirated materials were prepared using conventional methods, liquid-based cytology (LBC), or both. The conventional smears were produced using a press and release method, 9 and were fixed with Cytorop (Alfresa, Japan), which is a cytological fixative. When excessive fluid was aspirated, the fluid components were removed by tilting the preparations, smears were done using the press and release method, and preparations were fixed 10 to 20 seconds after smearing. 9 For LBC, CytoRich RED collection fluid (Becton, Dickinson and Company, New Jersey) with haemolytic and proteolytic abilities was used for the CytoRich hand method. 10 Both preparations were stained with Papanicolaou stain.
Thereafter, the conventional specimens in 209 nodules, LBC specimens in 221 nodules, and both conventional and LBC specimens in 248 nodules were retrospectively reviewed. We also reviewed previous ultrasound findings saved in electronic medical records. Ultrasound was performed using the APLIO 80 SSA-770A (Toshiba Medical Systems Co., Ltd., Otawara, Japan) or the APLIO 500 TUS-A500 (Toshiba) with the PLT-805AT (Toshiba) or PLT-1005BT (Toshiba) probe. The ultrasound reports were interpreted based on both the Kuma Hospital ultrasound classification (KH-USC), previously described by Ito et al, 11,12 and the American Thyroid Association (ATA) guidelines. 4 The KH-USC classifications 5 and 4, 3, 2, and 1 roughly correspond to the high suspicion, intermediate suspicion, low suspicion, very low suspicion, and benign categories outlined in the ATA guidelines, respectively. The study was conducted in accordance with the guidelines of the ethics committee of Kuma Hospital; approval was granted by the ethic committee, and all subjects provided informed consent. The statistical significance of the data was determined using the Fisher's probability test, and P < .05 was considered statistically significant. Table 1 shows the revised classification of the 678 nodules originally  re-aspirated. The incidences of the re-aspiration in KH-USC categories 4, 3, 2, and 1 were 100%, 0%, 4.0%, and 0%, respectively.

| RESULTS
Re-aspiration was performed to facilitate diagnosis for only one calcified nodule in which malignancy was suspected by ultrasound. The second aspiration was performed 7 days after initial aspiration. In 11 KH-USC 3 nodules, re-aspiration was not done. In the remaining 17 nodules, cytological examination was combined with drainages of fluid or percutaneous ethanol injection therapy (PEIT). The duration between the initial and second aspirations was less than 2 weeks in 2 nodules, 2 weeks to 3 months in 5 nodules, and more than 3 months in 11 nodules. Cytological diagnoses of re-aspiration included CFO (13 nodules, 72.2%), benign (4 nodules, 22.2%), and AUS (1 nodule, 5.6%).    (7) Benign nodular goitre (  The ROM (0.2%) of the present study was strikingly lower than that (2.0%) of 10 years ago. The difference may be due to the development of ultrasound, improvement of aspiration skills and diagnostic ability, combined use of the LBC method, and a short follow-up period.
In countries otherthan Japan, CFO nodules seem to beclassified as ND/UNS because thepossibility of cystic papillary thyroid carcinoma cannot be excluded. When a nodule is entirely cystic and has no worrisome sonographic features, it is handled as a benign nodule. On the other hand, when the sonographic features are worrisome, it is considered clinically unsatisfactory. It is important to ascertain whether these classifications and handling strategies are appropriate.
If we think that cases diagnosed as CFO or benign should not contain malignancy, we will not be able to use the terms "CFO" or "benign" as diagnostic categories in the cytology. Thus, this concept is not practical. The presence of a few malignant cases should be allowed in cases diagnosed as CFO or benign. The problem is then defining the permissible ROM. This decision may involve consideration of both academic data and the social environment. In our study, the ROM of CFO nodules was 0.2%. The incidence was significantly lower than that of NDother nodules. The ROMs of nodules diagnosed as benign was 1.1%, and 13.7% for subsequently resected benign nodules. 19 The ROMs of CFO nodules are evidently lower than those of benign nodules. In Japan, the fact that CFO nodules reveal such lower ROMs indicates that these cases should not be handled as ND/UNS. 3 Based on ATA guidelines, worrisome sonographic features were identified in 5.8% of CFO nodules. Based on KH-USC, no CFO nodules with worrisome sonographic features were detected. As per the RSJSTS reporting system, we suggest that there is no reason to categorize CFO nodules with an extremely low ROM as ND/UNS. In addition, CFO and ND-other nodules differ in terms of their clinical management. 1,4 For benign nodules, the ATA guidelines recommend that follow-up should be determined by risk stratification based on ultrasound patterns. 4 We propose that CFO nodules should be classified separately from ND/UNS nodules; They should be classified as an independent category, similar to that in Japan, 3 or a subtype of benign nodules. However, it is essential that FNAC be performed under ultrasound-guided real-time visualization of needle placement in the target nodule in all cases. Our proposal is based on the data collected from a thyroid center with high patient volume in Japan. Therefore, further research analyzing data from other hospitals and other countries is warranted.