Ultrasound-Guided Fine Needle Aspiration of Deep Thyroid Nodule: Is There a Correlation between the Nodule's Depth and Nondiagnostic Results?

Objective To evaluate whether thyroid nodule depth correlates with nondiagnostic results in ultrasound-guided fine needle aspiration cytopathology. Background Many factors correlate with nondiagnostic ultrasound-guided fine needle aspiration cytology (FNAC) results, including older age, macrocalcification, small-sized nodules, aspirin medication, and cystic portion in more than 50% of the thyroid nodules. However, there are few studies which have examined whether there is a relationship between the depth of nodules and the percentage of nondiagnostic results in cytology (Bethesda category I). We conducted this study in order to investigate if such a correlation exists. Materials and Methods FNAC was performed on 283 thyroid nodules between January 2019 and December 2020. Cytological analyses of the nodules were reviewed and sorted as nondiagnostic and diagnostic according to the Bethesda score. Patient files and ultra sound (US) scans were reviewed for clinical information (such as age, sex, and ethnic group) and sonographic features of nodules (such as depth, size, cystic portion, type of calcification, and echogenicity) and were compared between the nondiagnostic and diagnostic nodule results. The depth of a nodule was calculated as the shortest distance from the skin to the most superficial border of the nodule in the axial plane, using our medical center's computer program, which allows reviewing all saved shots of the US scan. Results Age, sex, and ethnicity were not significantly different between the nondiagnostic group and the diagnostic group (p > 0.05). Nodule diameter, cystic portion, calcification, and echogenicity were also not associated with the frequency of nondiagnostic results. The depth of nodules ≥9 mm was correlated with nondiagnostic US-guided FNA cytological results (OR = 2.55, p=0.018). Conclusions Deep thyroid nodules correlated with nondiagnostic US-guided FNA cytological results. Further studies are needed for optimizing the approach to deep thyroid nodules in order to improve the efficacy of FNA in deep thyroid nodules.


Introduction
yroid nodules are common, being diagnosed in 34% (27% in men, 41.7% in women) of the adult population [1][2][3]. e majority of nodules are benign [4]. Ultrasonography (US) is the most important diagnostic tool in the workup of thyroid nodules. e clinical importance of thyroid nodule diagnosis rests on the need to exclude thyroid cancer, which occurs in 5%-15% of cases in correspondence to age, sex, previous exposure to ionizing radiation, family history, and other factors [4][5][6][7][8].
e Bethesda system for reporting thyroid cytopathology (TBSRTC) standardized the reporting of thyroid cytopathology results [15]. It includes six diagnostic categories, which are linked to certain ranges of malignancy risk and clinical management guidelines [16]. In TBSRTC, inadequate samples are reported as Category I.
is category applies to samples that are nondiagnostic or unsatisfactory due to obscuring blood, overly thick smears, air drying of alcohol-fixed smears, or an inadequate number of follicular cells that got on the aspirates [17].
A high nondiagnostic rate is correlated with older age, macrocalcification, small-sized nodules, cystic portion more than 50% of the thyroid nodules, hypoechogenicity, heterogeneous echogenicity, aspirin medication, lesser experience of the performing physician, lower procedural volume, and aspiration without rapid on-site evaluation (ROSE) [14,18,19].
High nondiagnostic FNAC result rates continue to hinder the full potential of FNAC as a diagnostic tool. e percentages for nondiagnostic US-guided FNAC in the literature range widely from 2 to 29% [16,18,20,21]. Such percentages of nondiagnostic results increase the length of time to definitive diagnosis, patient stress, costs, and lead to unnecessary thyroidectomy. Moreover, studies have shown that nondiagnostic samples can harbor malignancy in 7% to 28% of cases [7,18,20].
Based on our experience in examining deep nodules, needle navigation and nodule aspiration may be more challenging. Surprisingly, we found only two studies investigating the relationship between the depth of thyroid nodules and the rate of nondiagnostic FNA cytology results [19,22].
In this study, we aim to evaluate whether thyroid nodule depth as measured by US can predict nondiagnostic cytological results in thyroid FNAC.

Study Population.
A retrospective chart and image review was conducted on 215 patients with 283 thyroid nodules who underwent US FNAC between January 2019 to December 2020 in Ziv Medical Center, Israel. e local Ethics Committee approved the protocol. Clinical information, US, and FNA cytology records were reviewed for all patients.
We followed the National Comprehensive Cancer Network (NCCN) and the American yroid Association (ATA) guidelines for deciding which nodules were candidates for sampling. High risk features considered included undefined or lobulated margins, microcalcifications, tallerthan-wide shape, hypoechogenicity, extrathyroidal extension of thyroid nodule, and prior exposure to radiation [23,24].

Ultrasound Examination.
A single head and neck surgeon with more than 10-years' experience in the area of thyroid malignancy performed all the US-guided FNA procedures. e depth of each nodule was considered as the shortest distance from the skin to the most superficial margin of the nodule in the axial plane ( Figure 1). Every evaluated nodule was assessed for size, cystic component, calcification, location of nodule, and echogenicity. e size of the nodule was assessed by measuring two diameters: a horizontal diameter (dimension A) and an anterior-posterior diameter (dimension B) ( Figure 1). Nodules without the cystic component were considered as solid, otherwise the cystic portion was classified as less than 50% or more than 50%. We excluded all nodules that were purely cystic, because these nodules are invariably nondiagnostic. Calcification when it existed was classified as microcalcification or macrocalcification, when calcification foci size was less than 1 mm or bigger than 1 mm, respectively. Echogencity of a nodule was classified as hyperechoic, isoechoic, hypoechoic, and heteroechoic. Location of the nodule was classified as in the right lobe of the thyroid gland, left lobe, or into the isthmus.
2.3. Ultrasound-Guided FNA Procedure. All biopsies were performed using an A 23 Gauge needle, attached to a 3 ml syringe. e needle was 30 millimeter long, not measuring the hub. When it was attached to the syringe, their length together was 130 millimeters. e surgeon adjusted the ultrasound depth and focus as needed to get better visualization of the nodule and the needle tip.
Under US guidance, we collected three syringes (passes) from every nodule. e skin overlaying the nodule was locally anesthetized by lidocaine + prilocaine 2.5% + 2.5% cream (Rafarm S.A, Athenes, Greece) which was applied for 45 minutes for minimum, and the skin cleaned and sterilized with alcohol 70% solution (Floris, Israel). e needle was penetrated into the nodule by holding the syringe, and without the use of a holder. It was moved back and forth within the nodule. An equal vacuum was created on the syringe from the moment the tip of the needle was inserted into the nodule until the hub of the needle was filled with aspirated material and then it was extruded from it.

Cytological Preparation and Diagnosis.
Following every FNAC procedure, an experienced cytopathologist immediately expelled the collected materials on glass slides, smeared, and directly fixed the specimen in 95% ethyl alcohol. Staining was performed using the Papanicolaou method. ROSE was not performed. e samples were reported using the Bethesda System. An adequate sample was defined as having at least six clusters of thyroid follicular cells (10 cells) for each specimen. When fewer cells were obtained but abundant colloid was observed, the specimen was considered adequate and benign. Whenever an atypical or other diagnosis (such as thyroiditis) was detected, the specimen was considered as adequate.
e chief cytopathologist reviewed and confirmed the results of all slides.

Statistical Analysis.
Quantitative data were shown as mean ± standard deviation, whereas numbers and percentage were provided for the qualitative data. Pearson's chisquare test was used to examine the relationships between study groups for categorical parameters such as noudel depth, time of surveillance, age, and both A and B dimensions. Variables were tested for normality according to Kolmogorov-Smirnov. After that, we applied the ROC curve constructed, the chi-square test to measure the differences between the study groups (Bethesda 1 vs. Bethesda 2-6), and the Kruskal Wallis H test to measure the differences between all of the study groups. Multiple logistic regression assessed the correlations between the study groups and risk factors, providing ORs and 95% confidence interval (CI), with adjustment for significant and potential (age and gender) confounders.
A p-value of 5% or less was considered statistically significant. e statistical analyses were analyzed using the SPSS version 25 (SPSS Inc., Chicago, ILL, USA).

Results
A total of 283 nodules were biopsied from 215 patients. Of these, 40 aspirations (14%) were classified as nondiagnostic. e other 243 aspirations (86%) were diagnostic and were categorized as follows: 227 as benign, 6 as atypical cells or follicular lesions of undetermined significance, 3 as follicular neoplasm or suspicious for follicular neoplasm, 2 as suspicious for malignant, and 5 as malignant. e average age was 57.5 ± 14.8 (Mean ± SD), and 175 (82%) of the patients were females. Baseline demographic data of patients with nondiagnostic nodules were not different compared with the diagnostic group (Table 1).
Kolmogorov-Smirnov test of normality indicated that nodule depth, time of surveillance, age, and both A and B dimension variables did not follow a normally distributed shape. us, we applied Mann-Whitney nonparametric tests to measure the differences between study groups (Bethesda 1 vs. Bethesda 2-6). By evaluation of sonographic characteristics, the portion of nondiagnostic specimens increased with deeper location (12.1 ± 4.7 mm in the Bethesda 1 group vs 9.7 ± 6.8 mm in the Bethesda 2-6 group, p < 0.001)) and shorter diameters of nodules (Dimension A: 14.5 ± 5.3 mm in the Bethesda 1 group vs 17.2 ± 8.3 mm in the Bethesda 2-6 group, p < 0.05 and Dimension B: 13.5 ± 6.7 mm in the Bethesda 1 group vs 16.4 ± 7.8 mm in the Bethesda 2-6 group, p < 0.01).
ere was no significant difference in correlation to the side of the nodule, cystic portion, calcification, or type of echogenicity.
Using the ROC curve constructed did not yield a significant result according to the AUC (p � 0.690). erefore, we used the chi-square correlations in order to assess the correlation between the Bethesda groups (1 vs. 2-6) and the depths of the nodules. By using the chi-square test (Table 2), we found that a depth of 8 millimeters is a clear cut-off which can predict nondiagnostic FNA cytology (Table 2). We used the binary logistic regression analysis (Table 3) for determining if there are multifocal influences on nondiagnostic specimens. For this purpose, we chose variables (from Table 1) with pvalue less than 0.05. e nodule's depth and diameter were the only candidate characteristics. In the multivariate analysis, nodules with a depth of ≥9 mm significantly increased the odds of nondiagnostic FNA (OR, 2.5; p � 0.022; CI, 1.14-5.45), when compared with the depth <9 mm (Tables 2 and 3). Nodule diameters were not significantly different between the two study groups after cancellation of confounders. Kruskal-Wallis nonparametric test was performed to examine the difference in the nodule depth between the different Bethesda groups [1][2][3][4][5][6]. e only significant difference was found between groups 1 and 2 (p < 0.01). No differences were found between the Bethesda 1 and any of the Bethesda 3-6 groups.

Discussion
Due to its high sensitivity and specificity, U-guided FNA is the gold standard for thyroid nodule evaluation. e number of ultrasound-guided FNAs has risen.
Over the last two decades, the number of ultrasoundguided FNAs has increased dramatically due to increased acceptance of the additional benefit of ultrasound guidance [10,19]. Still, nondiagnostic FNAC result rates continue to hinder the full potential of FNAC as a diagnostic tool [18,25].
erefore, to avoid bias, all thyroid nodules included in our study underwent US-guided FNA by the same experienced operator and all slides were interpreted by the same experienced cytopathologist.
In our study, we found that the depth of thyroid nodules correlates independently with nondiagnostic FNA cytology results. A cut-off of 9 mm depth increased the odds of nondiagnostic FNA results. e clinical characteristics of the patients were not significantly different between the diagnostic and nondiagnostic group, and they were not shown to be correlated with a nondiagnostic result of the FNA.
ere is a paucity of research regarding thyroid nodule depth, especially its relationship with nondiagnostic results. Xia et al. also found a correlation between nodule's depth and accuracy of the exam; in their research, they found that nodule depth of 15 mm or more was an independent factor for nondiagnostic results [22]. Kavanagh et al. also found that the average depth of nondiagnostic thyroid nodules was 15 mm, and it was significantly deeper than the diagnostic group. [19] Our study and both of those studies considered  10.0 ± 6.6 12.1 ± 4.7 9.7 ± 6.8 <0.001 Time of surveillance, m (mean ± SD) 9.2 ± 4.7 10.4 ± 5.6 9.0 ± 4.5 0.090  the depth of a nodule as the shortest distance between the skin and the outermost margin of a nodule. Both studies found no relationship between nodule size (as opposed to depth) and nondiagnostic FNA cytology results. e tendency of FNAC of deep nodules to be nondiagnostic may result from a few technical entities: difficulty to identify properly the nodule in the thyroid gland by the US device, difficulty in navigating the needle's tip to the nodule through the soft tissue of the neck, and the difficulty to aspirate the nodule. More than that, aspirating a deeper nodule is followed by greater pain and discomfort to the patient, which makes the examination even more difficult due to the patient's movements. erefore, deeper nodules may take more time to aspirate. is issue is of outmost importance, because the more the needle stayed into a nodule, more blood could enter the aspirate and obscure the cytological material which hinders the cytological examination and may lead to nondiagnostic cytological results.
One study showed that hypoechoic and hyterogenous nodules are prone to be less diagnostic compared with isoechoic or hyperechoic nodules [26,27]. is association was theoretically related to fibrosis, hemorrhage, necrosis, and high cellular structures in these nodules. In another study, the highest diagnostic rates were in the hypoechoic nodules and lowest rate in heterogenic nodules [28]. We did not find correlation between echogenicity of thyroid nodule and the nondiagnostic rate of FNA cytology.
Other factors that might influence the nondiagnostic rate but were not examined in this study might be the needle width and the use of the French technique. ere is some evidence that on the one hand, wide core needles associate with more cellular aspirates, but on the other hand, they are associated with more obscuring red blood cells; therefore, they are the same diagnostic as narrow core needles [26,27,29]. ere is some evidence that nonaspiration fine needle cytology, or "the French technique" reduces the number of obscuring blood and has a better sampling accuracy [30][31][32]. In this technique, a free needle with an uncovered hub is inserted back and forth into the thyroid nodule without activating negative pressure. However, other studies have failed to show superiority of nonaspiration cytology over aspiration cytology [33][34][35]. In the case of deep thyroid nodules, using a mere needle by holding it into the hub could be manually challenging. erefore, an alternative option would be to connect the needle to a syringe in which the plunger has been removed. is complies with the rule that the needle's hub should be exposed to the outer atmosphere, and we are able to better control the needle by holding the syringe. is technique may improve the navigation of the needle's tip into the thyroid nodule. e percentage of nondiagnostic FNA cytology in our study was 14.1% (40/283). is is compatible with studies from other centers [16,19,22,25,36].As mentioned previously, the range of nondiagnostic FNA cytology is wide, and we believe that this variability relates to the difference in the nodule characteristics associated with nondiagnostic cytology between different study groups, physician experience, aspiration technique (short vs long axis), and variability in the use of ROSE. ROSE is a microscopic evaluation of the cytological material immediately after the aspiration.
ere is controversial evidence about the utility of using a ROSE for improving the diagnostic rate of US-guided FNA of thyroid nodules [37]. Some studies show a significant impact of ROSE on the nondiagnostic rate of FNA of thyroid nodules [38,39]. Even in a case where a relatively low initial rate of nondiagnostic results has been observed using ROSE could have significantly improved the accuracy of FNA cytology [18]. ROSE could be efficient on subcentimeter, mixed solid-cystic, macrocalcified, and hypervascular nodules [40].On the other hand, one meta-analysis concluded that the utility of ROSE depends robustly on the initial adequacy rates, the higher the initial rate of nondiagnostic FNA cytology, the bigger the impact of ROSE and vice versa [41]. Generally, ROSE requires extended duration of the procedure, longer patient discomfort, and higher costs [38]; however, it may have a significant implication on the subgroup of deep nodules.
Our study is based on large samples of patients, with a wide diversity of demographic and clinical characteristics that were included in the analysis.
ere were some limitations to our study. One limitation is that this was a retrospective study, so a selection bias could not be avoided. Another limitation is that ROSE was not applied on the study's population.

Conclusion
In our study, we showed that deeper thyroid nodules are associated with a higher rate of nondiagnostic FNA cytology. Different FNA methods may improve the diagnostic rate of deep thyroid nodules, reduce recurrent aspiration, and minimize patient discomfort. Further studies are needed for optimizing the approach to deep thyroid nodules in order to improve the efficacy of FNA in deep thyroid nodules. ese future studies could also focus on the length and gauge of the needles, as well as the duration of the FNA procedure.

Data Availability
e data used to support the findings of this study are included within the supplementary information file.