Can preoperative serum thyroglobulin and anti- thyroglobulin levels predict malignant potential of a thyroid nodule with atypia or follicular lesion of undetermined significance?

Background: The heterogeneity of atypia of undetermined significance (AUS) and follicular lesion of undetermined significance (FLUS) renders it difficult to classify the risk as benign, suspicious, or malignant. In the present study, we assessed whether patients’ preoperative demographic, radiological, and biochemical characteristics can predict the postoperative pathological diagnosis in patients with AUS/FLUS. Methods: Retrospective review of patients with AUS/FLUS who underwent surgery over a six-year period was conducted. The primary outcome was prediction of the malignant potential of AUS/FLUS nodules based on demographic data, thyroid imaging reporting and data system (TIRADS) score, and thyroglobulin (Tg) level. Results: Fifty-one cases were identified with a mean age of 40 ± 12 years and 36 (70.6%) were female. The final pathological diagnosis of malignancy was seen in 29 patients (56.9%). Patients with malignancy had a higher Tg level and TIRADS points (154 ± 167 vs. 463 ±733 μg/L, p = 0.0352 and 3.3 ± 1.4 vs. 5.9 ± 2.7, p = 0.0001, respectively). The final diagnosis did not differ by patient demographics (age and gender). Conclusion: AUS/FLUS nodules in patients with high Tg level and TIRADS score showed a higher risk of malignancy, which was observed in nearly 50% of the patients in this retrospective study. Therefore, preoperative Tg and high TIRADS score could be useful in malignancy-risk stratification in thyroid nodules with indeterminate cytology. Further prospective multicenter studies that may throw light on this common endocrine cancer among Saudis and specifically among women, are needed to confirm our findings. *Correspondence to: Amal Alhefdhi, Department of General Surgery, Breast and Endocrine Section, King Faisal Specialist Hospital & Research Center (KFSH&RC), Riyadh, Saudi Arabia, Tel: +966 11 4427755, Fax: +966 11 4427772; E-mail: alhefdhi@kfshrc.edu.sa


Introduction
Thyroid cancer is the most common endocrine cancer worldwide [1], and in Saudi Arabia [2,3], where it accounts for approximately 9% of all malignancies and 12% of female malignant cancers [4]. Thyroid cancer is considered the second most common cancer among Saudi women, and the fourth among Saudi men [4]. Thyroid nodules are the most common presentation of thyroid cancer [5]. However, thyroid nodules are highly prevalent, accounting for 20%-76% in the general population [6]. Since therapeutic management of malignant nodules is different from that of benign nodules, it is vital to distinguish the two [5,6].
Can preoperative serum thyroglobulin and antithyroglobulin levels predict malignant potential of a thyroid nodule with atypia or follicular lesion of undetermined significance?
Based on the ACR recommendations, FNAC is not required for TR1 and TR2 nodules, but it is mandatory for TR3, TR4, and TR5 nodules if they measure 2.5, 1.5, and 1.0 cm, respectively [10]. Although, both ACR and ATA do not recommend the FNAC for nodules that measure < 1.0 cm [6,8], FNAC is justified in nodules measuring 5 to 9 mm that fall under risk category [7,9].
Atypia of Undetermined Significance (AUS) or Follicular Lesion of Undetermined Significance (FLUS) is considered the third BSRTC category. The risk of malignancy in this particular group is 5%-15% [9][10][11], but because it is a heterogeneous group, categorizing AUS into the six malignancy-risk groups is difficult. To investigate AUS further, it is essential to identify a suitable biomarker, which remains a challenge. To predict malignancy in thyroid nodules, some have used molecular, cytological/morphologic and imaging, whereas others have focused on the levels of preoperative thyroglobulin (Tg), albeit with conflicting results [12]. Normal or well-differentiated malignant thyrocytes produce Tg, a glycoprotein that may be used as tumour marker for patients with differentiated thyroid cancer (DTC) [12].
Genetic testing of the thyroid fine-needle aspiration biopsy (FNAB) detects malignancy in patients with AUS/FLUS with high sensitivity and specificity [8,13,14]. Since all tertiary hospitals do not provide this genetic testing, the need for an alternative method is warranted. This study aimed to describe the correlation between preoperative demographic data, radiological characteristics required for identifying TIRADS-based malignancy-risk groups, biochemical characterization of Tg and anti-Tg antibody and final postoperative pathological diagnosis among patients with AUS/FLUS.

This study was approved by the Office of Research Affairs at King Faisal Specialist Hospital & Research Center (KFSH&RC), Riyadh, Saudi Arabia. Institutional review board approval number 2161182.
Between 2011 and 2016, adult patients with primary diagnosis of AUS/FLUS and who underwent surgery at the King Faisal Specialist Hospital & Research Center (KFSH&RC), Riyadh, Saudi Arabia, and who had documented preoperative Tg,TgAb and TSH levels, were included in this retrospective study, which comprised patients with a final diagnosis of both benign and malignant tumours. We excluded pediatrics cases (younger than 18 years old) and patients with missing records, and those who were operated at an external hospital. All cases had repeated FNAC with two consecutive results of AUS or FLUS on cytology. Patient data including age, gender, TSH, Tg level, Tg antibody (Ab) and US variables using the TIRADS score were retrieved from the electronic medical records. Missing variables were obtained from the patients' charts.
All statistical analyses were performed using SAS version 9.4 (SAS Institute Inc., Cary, NC, USA). Descriptive statistics for the continuous variables were reported as mean ± standard deviation and categorical variables were summarized as frequencies and percentages. The continuous variables were compared using Student's independent t-test, Pearson correlation test, and scatter diagram. The categorical variables were compared by Chi-square test. The level of statistical significance is set at p < 0.05.

Results
In this retrospective analysis, we screened 161 cases of AUS/FLUS. After applying the inclusion and exclusion criteria, 51 cases met our study criteria in which the patients had documented preoperative serum Tg, TgAb, and TSH levels. Patients' mean age was 40 ± 12 years. In general, the final pathological diagnosis of malignancy was noted in 29 patients (56.9%), while the remaining (43.1%) were benign. Females (n = 36) comprised 71% of the total. Though malignancy rate (73.3%) was higher among males, it was not statistically significant (p = 0.2142). Patient demographic characteristics did not differ by tumour status ( Table 1).
The mean serum Tg and TgAb levels were 330 ± 580 µg/L and 153 ± 583 U/ml retrospectively, while the mean TSH level was 1.84 ± 1.2 mU/L. The mean serum Tg was significantly higher among patients with malignancy compared with patients with benign tumours (463 ± 733 µg/L vs. 154 ± 167 µg/L; p = 0.035 respectively). The mean TgAb and TSH did not differ by tumour status (Table 1).
Of the 51 patients, 49 had available US images that were reexamined by a radiologist using the TIRADS score. Reports for two cases were unavailable. The TIRADS points were higher among patients with malignancy compared with those with benign tumours (5.9 ± 2.7 vs. 3.3 ± 1.4; p = 0.0001 respectively). The mean size of the largest thyroid nodule was 38 ± 24 mm, but the mean size of the nodule did not differ by tumour status ( Alhefdhi A (2020) Can preoperative serum thyroglobulin and anti-thyroglobulin levels predict malignant potential of a thyroid nodule with atypia or follicular lesion of undetermined significance? Petric et al. [24] reported that patient age, solitary tumour, and preoperative serum level of Tg were independent predictors of malignancy in follicular neoplasms. Strazisar et al. [25] reported that patient age and preoperative serum level of Tg were independent predictors of malignancy in Hürthle cell neoplasms. In a systematic review of 13 studies, with 9 that focused on thyroid nodules with indeterminate cytology, found a significance difference between benign and malignant nodules in association with mean or median serum preoperative Tg level [12]. Therefore, we concluded Tg to be an independent predictor of thyroid cancer, particularly with indeterminate cytology [12].
Though we could not detect a significant difference in the preoperative serum level of TgAb between benign and malignant groups, Karatzas et al. [26] found that in thyroid nodules with indeterminate cytology, there was a significant correlation between the TgAb and PTC (P < 0.001). Hosseini et al. [27] found that high TgAb level (≥ 30 IU/ml) may increase the risk of malignancy. On the contrary, Kim et al. [28] reported that a positive TgAb test is an independent predictor of thyroid nodule malignancy along with TSH results, regardless of autoimmune thyroiditis. In contrast, Yalcin et al.
[29] and Yazici et al. [30] found that high preoperative serum levels of serum Tg, anti-TPO, and TgAb, and high preoperative serum TSH levels were not predictive in thyroid cancer.
All patients with AUS/FLUS in our study with a thyroid malignancy had TIRADS 3 or more, and subsequently higher overall TIRADS points increase the risk of malignancy, which was implied from the risk of malignancy in TIRADS 3 (39%), 4 (66.7%), and 5 (90.9%), which is in agreement with the reported risks of cancer for the same category: 5%, 50%-80%, and >80%, respectively [17], with the caveat that these rates were based on the risk of malignancy using the TIRADS score in general. In this study, the TIRADS score was used only for AUS/FLUS group.
A previous report showed that US variables and TIRADS categories did not differ between benign and malignant nodules or between malignancies with one and two AUS/ FLUS nodules [18]. Park et al. [18] assessed the rate of malignancy and characteristics of thyroid nodules in 31 patients with AUS/FLUS and found a high (58.1%) rate of malignancy with no differences between the benign and malignant nodules in demographic or US characteristics.
Thus, AUS/FLUS nodules in patients with high Tg level and TIRADS score showed a higher risk of malignancy. This study is not without its limitations. The retrospective nature of the study and the small sample size limit the generalizability of our findings. Thus, we use it as a pilot study, and we are considering a prospective multicentre clinical study to confirm our findings.
In conclusion, elevated preoperative Tg and high TIRADS score could be useful in malignancy-risk stratification in thyroid nodules with indeterminate cytology. The limitations of this study can be overcome by pursuing prospective multicentre studies that may throw light on this common endocrine cancer among Saudis and specifically among women.
follows: 23 cases with TIRADS 3 of which nine (39%) had thyroid cancer (four follicular cancer and five papillary thyroid cancer (PTC); two with follicular variant); 12 cases with TIRADS 4 of which eight (66.7%) had thyroid cancer (two with follicular cancer and six with PTC; three with follicular variant); 11 cases with TIRADS 5 of which 10 (90.9%) had thyroid cancer (two with follicular cancer and eight with PTC; one with follicular variant). This is shown in table 2 and graphically depicted in figure 1.

Discussion and conclusion
Examination of the correlation between preoperative demographic data, radiological characteristics required for identifying TIRADSbased malignancy-risk groups, biochemical characterization of Tg and anti-Tg antibody and final postoperative pathological diagnosis among patients with AUS/FLUS showed that AUS/FLUS nodules in patients with high Tg level and TIRADS score showed a higher risk of malignancy, which was observed in nearly 50% of the patients in this retrospective study.
Age and gender were not significant variables influencing the differences between benign and malignant groups in this study, and are in line with previous studies [18,19]. However, the risk of malignancy of 56.9% was higher than most published studies, which ranged from 13.5% to 43% [11,20,21], but is in concordance with a recent study, which found a 58.1% rate of malignancy in AUS/FLUS nodules [18].
We found higher preoperative serum levels of Tg among patients with thyroid malignancy, suggesting that higher Tg level could increase the risk of malignancy. This finding was in concordance with Sands