Cytology versus calcitonin assay in fine-needle aspiration biopsy wash-out fluid (FNAB-CT) in diagnosis of medullary thyroid microcarcinoma

Widespread use of sensitive ultrasound examination led to an increasing detection of medullary thyroid microcarcinoma (micro-MTC). This prospective study evaluated the diagnostic accuracy of Fine-needle Aspiration Biopsy Cytology (FNAB-C) and calcitonin assay in Fine-needle Aspiration Biopsy wash-out fluid (FNAB-CT) in thyroid nodules less than 1 cm with elevated serum calcitonin(sCT). 87 thyroid nodules from 60 patients with elevated sCT (>10 pg/ml) were included and 51 were thyroid nodules less than 1cm. FNAB-CT and FNAB-C were performed to distinguish medullary thyroid carcinoma (MTC) lesions before surgery, histopathologic diagnoses served as main reference standards. FNAB-CT had a greater performance over FNAB-C for preoperative diagnosis of MTC (diagnostic accuracy: 98.85 vs 61.90%, sensitivity: 98.55 vs 55.07%, specificity: 100 vs 97.44%), especially for micro-MTC: FNAB-C established a sensitivity and diagnostic accuracy of 48.78 and 58% respectively, while FNAB-CT reached 97.56% sensitivity and 98.04% diagnostic accuracy. FNAB-CT demonstrated high diagnostic accuracy in diagnosing micro-MTC. Patients with microscopic thyroid nodules and elevated sCT level should perform FNAB-CT to exclude the diagnosis of MTC lesions.


Introduction
Medullary thyroid carcinoma (MTC) is a neuroendocrine malignancy originated from the calcitonin (CT)-secreting parafollicular C cells, which accounts for~2% of all malignant thyroid neoplasms and 13.4% of the total deaths attributable to thyroid cancer [1]. According to the greatest dimension in histopathology, MTC can be classified into microcarcinoma (micro-MTC, ≤10 mm) and macrocarcinoma (macro-MTC, >10 mm). The routine usage of sensitive ultrasound enabled the early detection of thyroid microcarcinoma and some retrospective analyses suggested an increasing trend in the proportion of micro-MTC in recent decades [2][3][4][5][6]. Comparing with its larger counterpart, micro-MTC has a similar aggressive behavior (including extra-thyroid extension, regional and distant metastases) but a relatively atypical diagnostic feature [2,3,[7][8][9][10].
These authors contributed equally: Ziyuan Liu, Wenzhong Zhou * Xi Chen cc_viai@yahoo.com.cn * Lei Ye lei_yelei@163.com Serum calcitonin(sCT) level generally rises early in MTC patients and parallels with tumor progression. Setting a recognized cutoff as 10 pg/ml, it has a sensitivity of nearly 100% in detecting MTC [11][12][13]. Nevertheless, routine sCT measurement in nodular thyroid disease remains controversial, in addition to cost-effectiveness concerns, evidence from population-based studies indicated a high false-positive rate in patients with "indeterminate" elevated sCT levels (10-100 pg/ ml) [12,14,15]. As tumor size significantly correlates with preoperative sCT levels, patients with micro-MTC usually have basal sCT level in this "indeterminate" range (10-100pg/ ml), making it difficult to distinguish early stage MTC from other physiological and pathological conditions [8,16,17]. Stimulation tests had been proposed to confirm MTC suspicious cases, especially for patients with "indeterminate" sCT [11,12,18]. Whereas, due to the improved sensitivity of sCT assay, one recent study found no superiority of the stimulation tests in early diagnosis of MTC [19]. In addition, calcitonin can be ectopic produced by neuroendocrine tumors [13,20]. Thereby, the ability of sCT to command an optimal preoperative evaluation for micro-MTC, is limited.
Ultrasound guided fine-needle aspiration biopsy cytology (FNAB-C) is highly recommended for differential diagnosis of benign thyroid nodules and malignant thyroid neoplasms. However, for MTC, its diagnostic performance may be less convincing, a meta-analysis of fifteen relevant studies indicated FNAB-C can only identify about one-half of all 641 MTC lesions [21]. In a more recent retrospective study, based on original interpretation alone, the sensitivity of FNAB-C was 68.3% among 145 MTC cases in Asia-Pacific region [22]. The high false-negative rate can be attributed to the morphologic heterogeneity of MTC, and some researchers found micro-MTC has more "mimic" cytological features such as less oncocytic change and more colloid presentation [9,[22][23][24]. Moreover, sampling error of FNAB-C occurs frequently in subcentimeter thyroid nodules [25,26]. Thereby, the ability of FNAB-C to command an optimal preoperative evaluation for MTC lesions, particular in micro-MTC, is also limited.
Calcitonin assay in fine-needle aspiration biopsy wash-out fluid(FNAB-CT) was firstly described as a diagnostic tool by Boi et al. [27] Thus far, previous studies had indicated FNAB-CT has an ideal diagnostic performance (nearly 100% sensitivity and specificity), some researchers suggested it as an ancillary tool to confirm inconclusive cytopathologic interpretation [22,[27][28][29][30][31][32][33][34][35][36].However, according to the revised American Thyroid Association (ATA) guidelines for management of MTC, FNAB and the follow up calcitonin measurement in the FNAB washout fluid were recommended only for thyroid nodules that are 1cm or greater in size [15]. The diagnostic performance of FNAB-CT has not been evaluated independently on micro thyroid MTC nodules.
To estimate the diagnostic value of FNAB-CT (versus FNAB-C) in microscopic thyroid nodules among MTC suspicious thyroid nodules, we conducted a prospective study following the latest STARD statement (STAndards for Reporting Diagnostic accuracy studies) [37]. In consideration of the high sensitivity of basal sCT levels in detecting MTC, MTC suspicious thyroid nodules in the present study were defined as thyroid nodules with elevated basal sCT(>10 pg/ml).

Study design and participants
A prospective study was conducted from September 2017 to December 2020 in accordance with the STARD statement. The selected patients met at least two criterias: (1) ultrasounddetected thyroid nodules; (2) persistent elevated serum calcitonin (sCT > 10 pg/ml for more than 3 months during follow up); The flow of participants through the study was described in Fig. 1. A total of 78 participants underwent FNAB were consecutively included, 108 thyroid nodules including 57 nodules less than 1 cm (defined as microscopic thyroid nodules) had fine-needle aspiration biopsy performed and wash-out fluid collected for calcitonin assay; 105 thyroid nodules including 56 micro nodules had cytology analyzed. Finally, 60 participants with 87 thyroid nodules met endpoints, including 51 microscopic thyroid nodules from 39 participants. In addition, peripheral blood samples were taken from 53 participants for RET mutations screening of exons 8, 10, 11, 13, 14, 15 and 16 by standard PCR-based Sanger sequencing.

Hormonal assay
All tests were performed in a College of American Pathologists (no. 7217913) accredited laboratory. Serum calcitonin (sCT) and FNAB-CT measurement was performed using a chemiluminescent immunoassay (Mindray Medical International, Shenzhen, China) with a sensitivity of 1.0 pg/ml. The interand intra-assay CVs for CT were 10 and 5%, respectively. Normal range for sCT was 0-10 pg/ml. As no association was found between the size of MTC lesions and the FNAB-CT levels, according to the cutoff proposed by Boi et al. [27], normal range for FNAB-CT was set as 0-36 pg/ml, for both microscopic and macroscopic thyroid nodules.

Ultrasonography and FNAB-C
Ultrasound examination was performed using the MyLab ™ Platform ultrasound system (Esoate SpA, Genoa, Italy) equipped with 4-13 MHz linear transducers.
US guide fine-needle aspiration biopsy was performed by certified interventional radiologists using a freehand technique with fine needle (22-gauge generally, 25-gauge for deep-seated or high vascularity lesions) attached to 5-mL plastic syringes. Each target lesion was aspirated two or three times. The obtained material was quickly smeared on glass slides and fixed immediately in 95% ethanol and submitted for haematoxylin-eosin staining.
Cytologic analysis was performed and categorized according to the six-tiered (Category I-VI) Bethesda System for Reporting Thyroid Cytopathology (TBSRTC) [39]. For the purpose of the study, all samples were checked for presentation of morphologic features of MTC. Cytologic diagnose as MTC or MTC suspicious were defined as FNAB-C positive.

SOP for FNAB-CT sample collection and transport
FNAB-CT sample was collected and transported according to standard operating procedures (SOP) as follows: 1. 1ml saline solution was added in a clean EP tube (2.5 ml) and precooled on ice. 2. After smear preparation of specimens for cytologic analysis, the same needle and syringe were quickly washed in precooled saline solution back and forth for at least 5 times. 3. EP tube with collected wash-out fluid were transported to laboratory on ice and centrifuged at 6000g for 10 min at 4°C, supernatant was transferred to new EP tubes for FNAB-CT measurement.

Pathological analysis
For thyroid nodules submitted to total thyroidectomy, histologic diagnosis including immunohistochemical stains for calcitonin and chromogranin A was performed to differentiate between MTC and non-MTC lesions. All of the histopathologic diagnoses were confirmed independently by two experienced pathologists.

Statistical analysis
Statistical analyses were performed using SPSS Version 21 and a P < 0.05 was considered statistically significant. True positive (TP) and true negative (TN) were defined as the correct prediction of the presentation MTC. The diagnostic performance, including sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) and overall accuracy were evaluated for FNAB-CT and FNAB-C. Receiver operating characteristic (ROC) analysis was performed to determine the area under the curve (AUC) value of FNAB-CT.

Clinical and demographic characteristics of the patients and lesions
As shown in Fig. 1, 78 patients with persistent elevated serum calcitonin and ultrasound-detected thyroid nodules were recruited, a total of 108 and 105 thyroid nodules received FNAB-CT and FNAB-C test, respectively. All 57 microscopic thyroid nodules had FNAB-CT performed and 56 had FNAB-C performed. At the end of the ). Among 53 patients tested, germline RET mutation was found in 20 patients (16 with microscopic thyroid nodules) at codons 634 (n = 15), 618 (n = 2), 611 (n = 1) and 533 (n = 2) (Table S1).

Comparison of FNAB-CT with FNAB-C in diagnosing MTC
FNAB-CT were performed on all 87 thyroid nodules. With the cutoff value as 36pg/ml, the sensitivity and specificity were 98.55 and 100%. Nineteen thyroid nodules showed negative FNAB-CT levels and 18 were confirmed as non-MTC lesions. False negative results occurred in a single case of MTC nodule with a size of 9.5*6.2*6.7mm; Sixtyeight thyroid nodules showed positive FNAB-CT level, all were histopathologically confirmed as MTC. The mean value of FNAB-CT in 69 MTC was significantly higher than in 9 non-MTC lesions (1993.40 ± 326.34 vs 3.45 ± 2.57pg/ml, p < 0.0001).

Comparison of FNAB-CT with FNA-C in diagnosing micro thyroid MTC nodule
Subsequently, we explored the performance of FNAB-CT in detecting MTC in microscopic thyroid nodules. As in Table 4 The limited diagnostic accuracy of FNAB-C should ascribe to the high incidence of false negative diagnosis and micro-MTC had a high false-negative rate of 41.18%. As in Table 5, among total 31 FNAB-C negative MTC, 21 was smaller than 1cm in size, including four sample error and 17 interpretation error: 12 as follicular neoplasm, three as multinodular goiter, one as papillary thyroid carcinoma (PTC) suspicious and one as Hashimoto thyroiditis. The other 10 false negative results occurred in macroscopic nodules(>1 cm) and included one sample error and nine interpretation error as follicular neoplasm. The majority (30/ 31) of these FNAB-C negative MTC were revealed by FNAB-CT before surgery.
To be noted, four patients with thyroid nodules (two had only one microscopic nodule) and elevated sCT (92.44, 149.71, 467.7, >2000 pg/ml) levels were excluded MTC diagnosis based upon FNAB-CT results. Subsequently, two of them found calcitonin-secreting pancreatic neuroendocrine tumors and the other two had normal sCT level after quitting herbal therapy. In addition, two patients with borderline elevated sCT (10.21, 16.43pg/ml) and microscopic   thyroid nodules, received total thyroidectomy and central lymph node dissection due to markedly high levels of FNAB-CT(>2000 pg/ml), histopathological findings confirmed MTC in both patients and one with lymph node metastases.

Discussion
The present study prospectively demonstrated FNAB-CT, with a high diagnostic accuracy, is a reliable tool in detecting MTC before surgery, especially for microscopic thyroid nodules with elevated serum calcitonin(sCT). Sensitive ultrasound examination facilitated the early detection of thyroid microcarcinoma including micro-MTC. However, probably to avoid the overtreatment of papillary thyroid microcarcinoma(micro-PTC), the current guidelines only recommended thyroid nodules ≥1cm to be evaluated by FNAB-C with or without ancillary tests including FNAB-CT [15,40]. For micro-MTC lesions, data are lacking. Some reports of cases showed the superiority of FNAB-CT over FNAB-C in detecting micro-MTC [28,29,32]. However, only a limited number of microscopic thyroid nodules were recruited and evaluated promiscuously with macroscopic nodules in some diagnostic studies [31,33,34]. In the present study, we estimated the two diagnostic tools in micro-MTC independently: FNAB-CT showed better diagnostic performance: sensitivity as 97.56%, specificity as 100%, PPV as 100%, NPV as 90.91% and overall accuracy as 98.04%; whereas FNAB-C established sensitivity and NPV only as 48.78 and 30%, respectively. These data indicated that, for thyroid nodules less than 1cm with elevated sCT, FNAB-CT should be performed to exclude micro-MTC.
As neuroendocrine tumor, MTC had typical cytologic appearance as eccentrically placed nuclei chromatin (salt and pepper) and usually have a mixed cellular population: mainly spindle-shaped, plasmacytoid or epithelioid cells, and can also present multiple cellular variants, including bizarre giant cells, oncocytic cells, clear cells, and cells with a small cell carcinoma-like appearance. The architectural patterns of MTC are variable as well, ranging from classic discohesive single-cell pattern to cohesive cell fragments. Thus the differential diagnosis for MTC is complicated and should be confirmed by immunolocalization of calcitonin [23]. When cytologic specimen lacking plasmacytoid morphology and presenting a microfollicular arrangement or round cells with finely granular cytoplasm, MTC can easily mimic follicular lesions. Also, the mixture of normal thyroid follicular cells during sampling contributed to the frequent misinterpretation of MTC as follicular neoplasm [9,23,24]. Consistent with previous studies, follicular neoplasm in present study accounted for interpretation error of the majority (21/26) of MTC lesions. We also found micro-MTC tend to have more non-diagnostic results and more variable cytological misdiagnoses. This is reasonable because a subcentimeter size is more challenging in aspiration and resulted in inadequate specimen and intermediate cytological feature [7,9,[23][24][25][26].
Notably, though calcitonin was regarded as the biomarker of MTC and an elevated basal sCT(>10 pg/ml) had been proven approximates 100% sensitivity in MTC screening, the   PPV was reported only as 10~40%. Many false positive results occurred in patients with systemic diseases (such as chronic renal failure), pharmacological treatments (such as proton pump inhibitors) or simply, smokers. They usually had mildly elevated sCT(10-100pg/ml), defined as "indeterminate" sCT by some researchers [13,17].Therefore, efforts have been made to enhance the cutoff value to improve the diagnostic accuracy. Costante G, et al described a 100% PPV for MTC with a cutoff of 100pg/ml, but the sensitivity dropped to 60% [12]. Colombo C, et al suggested a gender specific cutoff as 18.7pg/ml for female and 68pg/ml for male, obtaining 100% PPV but led to underdiagnosis of two micro-MTC [18]. Kwon et al. set a cutoff value for macro-MTC as 65pg/ml to improve diagnostic accuracy but was not able to exclude the presence of micro-MTC [17]. We evaluated the diagnostic performance of FNAB-CT in 29 patients with "indeterminate" sCT (10-100pg/ml) and found FNAB-CT exhibited a 100% diagnostic accuracy (versus 61.76% of FNAB-C, Table S2). Indeed, 36.54% (19/52) of the patients with MTC in present study had sCT values ranged from 10 to 100pg/ml, the mean size of their MTC nodules was significantly smaller than MTC nodules with sCT above 100pg/ml(7.59 ± 4.24mm vs 14.32 ± 9.08mm; p < 0.001). Furthermore, even significantly elevated sCT are not necessarily associated with MTC. Ectopic calcitonin-secreting neuroendocrine tumors (CTsecreting NETs) usually evoke "determinate" sCT elevations (>100 pg/ml) [20].In a retrospective study focused on macro-MTC, when set the cutoff value of sCT as 65pg/ml, the only two false positive diagnoses were from CT-secreting NETs [17]. In the present study, most of the patients (34/37) with sCT levels above 100pg/ml had at least one histopathologically confirmed MTC lesions (mainly macro and multiple MTC). Two participants with extremely high levels of sCT (467.7pg/ml and >2000 pg/ml) were proven ectopic calcitonin-secreting pNETs, Their thyroid nodules were classified as TIR3 or TIR4a on ultrasound examination, and were excluded MTC diagnoses for negative FNAB-CT results. Therefore, we suggest thyroid nodules with elevated sCT should be further evaluated by FNAB-CT to exclude the diagnosis of MTC, and CT-secreting NETs should be considered in differential diagnoses when FNAB-CT suggests significant non-MTC derived hypercalcitonemia.
The clinical significance of present study lied mainly in micro-MTC. As reported, the incidence of micro-MTC is growing [2][3][4][5][6]. Unlike micro-PTC, a significant portion of micro-MTC presents an aggressive clinical feature. In a retrospective study based on SEER database, 37% (65/ 176) and 5% (16/310) patients with micro-MTC showed regional metastases and distant metastases at diagnosis, significantly decreasing their 10-year survival rates [2]. Similar results have been reported by other studies [3,8]. Two recent studies found MTC patients diagnosed in recent time period had greater proportion of micro-MTC and better clinical outcome [4,6]. However, size of the tumor alone is not an independent prognostic factor of MTC [10]. Micro-MTC should be managed as macro-MTC. Due to the low specificity of FNAB-C and sCT, we suggested FNAB-CT when a thyroid nodule less than 1cm was suspected of MTC.
In contrast to previous studies [27-31, 33, 34], the characteristic of the present study should be highlighted: (1) first evaluation of FNA-C and FNAB-CT in micro-MTC independently; (2) the prospective design and accordance to STARD statement; (3) detailed description of clinical and demographic characteristics in included patients and lesions. Conversely, we totally included 87 thyroid nodules and 51 microscopic thyroid nodules, the sample size is a limitation.
In conclusion, we confirmed FNAB-CT as a reliable tool to detecting micro-MTC. When cytology analysis had approximately 50% false-negative rate, FNAB-CT kept a nearly 100% diagnostic accuracy. We suggest thyroid nodule less than 1cm with suspected sonographic characteristic should be tested serum calcitonin; for those with elevated sCT, FNAB-CT should be analyzed to exclude the diagnosis of MTC, so that the subsequent appropriate surgical procedures can be determined.

Data availability
All data generated or analyzed during this study are included in this article.

Compliance with ethical standards
Conflict of interest The authors declare no competing interests.
Ethics approval This study was approved by the board of medical ethics of Ruijin Hospital, Shanghai Jiaotong University.
Consent to participate Written informed consents were obtained from all participates included in the study.