Predictors of Malignancy in Thyroid Nodules

Background: Better and early prediction of malignant thyroid nodule, and sure differentiation from benign one, is crucial need to decrease the rate of non-indicated surgeries. The aim of the work: To find risk factors and predictors of malignancy in patients with thyroid nodule[s]. Patients and Methods: Fifty patients with thyroid nodule[s] were included. All patients were evaluated by history taking, clinical examination and laboratory investigations. Imaging studies included thyroid ultrasound [US]. All were also submitted to fine needle aspiration cytology [FNAC] before treatment by total or subtotal thyroidectomy. Excised tissues were sent to histopathological analysis. Results: The incidence of malignancy was 22.0%. both benign and malignant groups were comparable as patient characteristics, complaints except the significant increase in cosmetic disfigurement among patients with benign nodules [92.3% vs. 45.5%]. Patients with malignant nodules had a significant increase of ill-defined margins, intranodular vascularity and enlarged lymph nodes [72.7%, 63.6% and 72.7% vs 5.1%, 12.8% and 25.6% respectively]. Ill-defined margins, enlarged lymph node and high grade in FNAC were the predictors of thyroid malignancy. Conclusion: Predictors of malignant thyroid nodules are illdefined edges, enlarged lymph node and high grade in FNAC. This helps clinicians to spare more benign cases from surgical interference.


INTRODUCTION
Thyroid nodules are found in 4% to 8% by palpation and in 10% to 41% by ultrasound study [1] . Thyroid cancer is more common than all other endocrine cancers with increasing incidence worldwide [2] .
The most prevalent types of thyroid malignancies are differentiated types, papillary and follicular malignancies [3] .
Clinically there are simple characteristics that are known to be risky regarding malignancy including; male sex, age below 30 years and above 60 years, family history of previous thyroid malignancy, previous exposure to neck irradiation and a hard nodule that may cause compression symptoms [4] .
Being a non-invasive and inexpensive tool, ultrasound became the most frequently used imaging test in evaluating the thyroid gland [5] .
In addition to clinical assessment ultrasound helps to identify risky nodules with risky criteria like microcalcifications, hypoechogenicity, ill-defined edged nodules and solid pattern [6] .
Fine needle aspiration biopsy [FNAB] is considered as the gold standard for the diagnosis of thyroid malignancy, but unfortunately about 25% of its biopsies are non-diagnostic [7] .
More recently, Thyroid stimulating hormone [TSH] is considered as a predictor of malignancy in thyroid nodules in many studies. There is a relationship between TSH and malignancy as the TSH promotes thyroid cell proliferation through cyclic AMP [8] .
Irrespective of all these developments, there is still a debate about the ideal approaches and actual predictors of malignant thyroid nodules. Thus, the current study had been designed.

AIM OF THE WORK
The aim of this study is to find risk factors and predictors of malignancy in patients with thyroid nodule[s].

PATIENTS AND METHODS
This was a prospective study carried out at the Department of Surgery, Damietta Faculty of Medicine, Al-Azhar University, Egypt. It included 50 patients with thyroid nodule[s] from November 2019 to June 2020. The study protocol was approved by the Local Ethics and Research Committee. In addition, an informed consent was signed by every patient after full clarification of study protocol. Patients' privacy and rights were assured.
The exclusion criteria were: patients with previous thyroid surgery for thyroid cancer, patients with previous cervical nodal biopsy, patients whose FNAC was established as malignant.  [9] .
Finally, computed tomography was carried out to detect retrosternal extension of the thyroid gland for those patients in whom US neck couldn't detect the inferior border on the gland, evaluate the local extension in more advanced stages of thyroid cancer, and it is appropriate for a suspicious mass with bulky cervical lymph nodes. In addition, all patients submitted to fine needle aspiration cytology [FNAC] before treatment by total, subtotal or hemi-thyroidectomy. The cytological specimens were examined and classified according to the Bethesda classification [11] .
Intraoperatively, patients were assessed for risk factors for thyroid cancer [such as worrisome metastatic lymph nodes, local invasion suggesting extrathyroidal extension [ETE], criteria of local structures invasion and metastatic lymph nodes confirmed on intraoperative frozen section analysis.

RESULTS
The incidence of malignancy in the current study was 22.0% [11 patients]. Patient age ranged between 17 and 70 years, the average age [SD] was 40.91±11.67 years. Both benign and malignant subgroups were comparable as regard to the patient age group, patient gender, complaints except cosmetic disfigurement which is significantly higher among benign when compared to malignant subgroups [92.3% vs 45.5%], a history of radiation exposure, positive family history and smoking habit [ Table 1].
TSH ranged between 0.01 to 18.0mIU/liter and there was no significant difference between patients with malignant or benign thyroid nodules [ Table 2].
In the current work, patients with benign when compared to malignant thyroid nodules revealed nonsignificant differences regarding to the number of nodules, consistency, echogenicity, calcification or size. On the other side, patients with malignant nodules had a significant increase in ill-defined margin, intranodular vascularity and enlarged lymph nodes [72.7%, 63.6% and 72.7% vs 5.1%, 12.8% and 25.6% respectively] [ Table 3].
The results of the FNAC revealed that, patients with malignant thyroid nodules had significantly higher grade of Bethesda grade when compared to patients with benign nodules [no any patients in malignant group had grades I or II, while 9.1%, 18.2%, 18.2% and 54.5% had grades III, IV, V and VI; while in benign group, the majority were grades I and II [5.1% and 74.4% respectively], 15.4% were grade III, 2.6% were grade IV and only 2.6% were grade VI [ Table 4].
In the current work, ill-defined margins, enlarged lymph node and high grade in FNAC were the predictors of thyroid malignancy [ Table 5].

DISCUSSION
Thyroid nodules are a common clinical problem, with prevalence rate of 5.0% of females and 1.0% of males living in iodine-sufficient countries. Prediction of malignant nodules is a critical issue and the development of ultrasound adds to the field a noninvasive modality. However, its value and predictive parameters are questioned [12] . The current work is an effort to identify predicting variables [clinical, intraoperative or radiological], fine needle aspiration cytology in the identification of malignant thyroid nodules.
Results revealed that, patients with benign nodules complained more significantly than those with malignant nodules of cosmetic disfigurement [92.3% vs 45.5%]. However, these percentages recognized disfigurement as the most common complaint. Otherwise, patients with malignant nodules had a significant increase in ill-defined margin on ultrasound examination, higher intranodular vascularity, more enlarged lymph nodes and higher grades on Bethesda classification. However, running multiple logistic regression revealed that, ill-defined margins, enlarged lymph nodes and results of FNAC [high grade] were the predictors of malignant nodules.
For the final diagnosis of malignancy, the histopathological examination was used as the goldstandard in the current work, although previous studies considered FNAC as the gold standard for the diagnosis of most thyroid nodules [13,14] . However, FNAC had a failure rate of extending between 30% and 70.0% [15] .
This high failure rate could be attributed to coarse [macro] or micro-calcification especially with thick-walls of thyroid nodules makes the penetration of this wall very difficult and even penetration achieved the enough sample for cytology is hard to be obtained [13][14] .
The incidence of malignant nodules in the current work is 22.0% which is higher than the previously reported value by Alexander et al. [16] [incidence rate 5-15%].
The incidence rate of thyroid nodules is increasing due to the introduction of new methods of diagnosis [i.e, wide use of ultrasound]. However, minority of nodules require surgery [17] .
Cozzolino et al. [18] reported a high incidence of malignant nodules reaching 38.5%. This heterogeneity in the incidence of malignancy was attributed to different selection criteria.
In the current work, malignant nodules were more common among males than females and in older than younger patients. However, the difference was statistically non-significant. This is comparable to results of Al-Hakami et al. [19] who reported that, thyroid malignancy was more prevalent in males than females, and in patients who were older than or equal to 45 years [but with significant difference].
Similar results to the current study were reported by Witczak et al. [20] who reported non-significant difference between benign and malignant nodules regarding patient's age or gender. However, Paul et al. [10] reported a significant increase of malignancy among females, which is contradictory to the current study and could be attributed to different inclusion criteria and sample size. In addition, geographic differences could explain this contradiction.
In the current work, both benign and malignant groups had no difference regarding family history of thyroid cancer. However, previous studies had reported that, positive family history is one of the known clinical risk factors for the development of malignant nodules [12,21] . However, and similar to the current study, Cozzolino et al. [18] found that, positive family history is not associated with the risk of thyroid malignancy.
No significant difference could be detected regarding the previous exposure to radiation. Pellegriti et al. [22] reported that exposure to radiation is the most likely contributing factor to cancer thyroid but also other environmental carcinogens can contribute.
The single nodule was increased among malignant when compared to benign nodules [63.6% vs 38.5% respectively]. However, the difference was statistically non-significant. Frates et al. [6] reported that a solitary nodule is associated with higher risk of malignancy than multiple nodules [p <0.01].
In addition, Li JZ et al. [23] reported that the incidence of thyroid cancer in higher among patients a single nodule. However, Gandolfi et al. [24] reported that multiple nodules should no longer be considered an indicator of benign disease and should be assessed as a solitary nodule.
On the other side, El-Gammal et al. [25] reported that multiple nodules were significantly associated with malignancy than solitary nodules.
In the current study, small nodule size less than 2cm was more common among malignant nodules. However, the difference is not statistically significant. This is comparable to Albuja-Cruz et al. [26] who reported non-significant association between nodule size and risk of malignancy.
Moreover, Megwalu et al. [27] reported similar results denying the association between tumor size and risk of malignancy.
In the current work, the only predictors of malignant nodules were ill defined margins, enlarged lymph node and high grade in FNAC.
These results agree with Jinih et al. [28] who reported that, their results suggested that, nodule size is not accurately predict the risk of thyroid malignancy irrespective of the results of FNAC.
Results of the current work also in agreement with Cozzolino et al. [18] who reported that, although nodular vascularity was more frequent in malignant nodules at univariate analysis, it was not a factor to predict thyroid malignancy at multivariate analysis.
However, there study contradict the current one in the defined predictors of malignancy as they only found a small tumor size and microcalcification are the predictors of malignant nodules. The fact that, they restricted their study to specific populations with specific criteria of thyroid nodules [grey zone; intermediate cytopathology thyroid nodules] could explain this contradiction.
The value of vascularization in the evaluation of thyroid malignant nodules has been much debated [29] .
Brunese et al. [30] argued that intralesional vascularity raises the suspicion of malignancy, while Frates et al. [6] demonstrate that intralesional vascularity not has a role. Jaheen and Sakr [31] reported that microcalcifications are a significant risk factor of malignancy. This could be recognized in the current study too.
Paul et al. [10] reported that ill-defined edges suggest malignant infiltration of adjacent thyroid parenchyma with no pseudo capsule formation. Jaheen and Sakr [31] also reported that ill-defined edges are a significant predictor of malignancy with a significant predictive value [p=0.029]. Another study by Witczak et al. [20] reported similar results. These results are confirmed in the current study.
In the current work, echogenicity could not differentiate between benign and malignant nodules and could not predict malignant transformation.
Kapali et al. [32] reported that the sensitivity of hypoechogenicity in predicting malignancy was 84%.
On the other hand, another trial found that hypoechoic pattern is not significant in predicting malignancy by Jaheen and Sakr [31] as in the current study.
Solid consistency could not be found to predict malignant thyroid. This is agreeing with Macias et al. [33] who reported that the solid nature of the thyroid nodule is not a significant predictor of malignancy.
In agreement with the current work, Kapali et al. [32] reported that, enlarged lymph nodes are predictors of malignancy in thyroid nodules, and similar results were reported by Jaheen and Sakr [31] , while Paul et al. [10] reported that, enlarged lymph nodes are of low value in predicting malignancy in thyroid nodules.
It is clearly evident that there was a gradual increase in incidence of malignancy from with increased Bethesda grade, with the highest incidence in grade six, making higher grades significant in prediction of malignancy in thyroid nodules.

Conclusion:
There are significant predictors of malignancy in thyroid nodules such as ill-defined edges, enlarged lymph node and high grade in FNAC. This helps clinicians to spare more benign cases from surgical interference. However, the small number of included patients represent a limitation of the current work, and future larger scale trials are required.

Financial and Non-financial Relationships and Activities of Interest
None