Radio Pathological Correlation of Thyroid Nodules Using Tirads Based Ultrasound Classification and Bethesda Classification for FNAC: a Prospective Study

Background: Thyroid gland is afflicted by various pathologies amongst which nodules are the cause of maximum concern because of their malignant potential. With the introduction of high resolution ultrasound and use of Thyroid imaging reporting and data system (TIRADS) classifi cation as a widely used universal grading system, there has been reduced inter-observer variability and increased inter-departmental communication. In this study, we studied the TIRADS ultrasound grading as a screening tool and compared it with the BETHESDA grading on FNAC. Material and methods: 200 patients with thyroid nodules were subjected to ultrasound and USG guided FNAC. Each was assigned a TIRADS and Bethesda grade. Findings were compared to assess the sensitivity, specifi city, PPV (positive predictive value) and NPV (negative predictive value) of ultrasound in differentiating benign from malignant nodules. RESULT: Out of 200 nodules examined, 116 nodules belonged to TIRADS 2 while 44, 13 and 27 belonged to TIRADS 3, 4 and 5 respectively. On FNAC, 162 patients belonged Bethesda 2 & 12, 7, 15 and 4 to Bethesda 3, 4, 5 and 6 respectively. The sensitivity, specifi city, PPV and NPV of ultrasound were found to be 92.3, 90.8, 60 and 98.75 % respectively. Conclusion: TIRADS is an effective risk stratifi cation system which should be routinely used in our clinical practice as it can predict the possibility of a particular nodule for being malignant to a great extent. Especially keeping in mind its high negative predictive value, FNAC can be deferred in TIRADS 2 patients which form a majority of cases reporting to pathology department for thyroid FNAC.

dized, category-based reporting system for thyroid FNACs 10 . It is as follows: Bethesda Class I-inadequate or unsatisfactory, Bethesda Class II -benign thyroid nodules, Bethesda Class III -follicular lesion of undetermined signifi cance, Bethesda Class IV -(suspicious for) follicular neoplasm, Bethesda Class V-suspicious for malignancy and Class VI-malignant.

MATERIAL AND METHODS
Th is study received approval from the institute's ethics committee (IEC) for human research. Informed consent was obtained from each participant, according to the Declaration of Helsinki, and the IEC approved pro forma.
Th e study was a prospective study and patients who had thyroid nodule in B-mode ultrasound were subjected to fi ne needle aspiration cytology (FNAC). Sample size was kept as 200 and study duration was 18 months (October 2018-March 2020).
Th e exclusion criteria were normal thyroid scan (TIRADS 1), proven case of thyroid malignancy (TI-RADS 6), and patients who had persistent Bethesda I (non-diagnostic) FNAC after two attempts.
Th e imaging was done using Esaote machine with high-frequency probe . During the procedure, patient lies in supine position with neck slightly extended and high resolution ultrasound examination of thyroid gland was done. Th e thyroid nodules, if present, were staged according to TIRADS. Th e neck was then assessed for any enlarged lymph nodes. FNAC of all nodules was done by the radiologist under direct USG guidance. Th e FNAC samples were read by an experienced pathologist and Bethesda staging was done. Th e data was subsequently transferred to a Microsoft Excel 2010 sheet and statistical analysis done using IBM SPSS for Windows version 22.0 (SPSS, Inc., Chicago, IL, USA).

RESULT
In our study, 80% of patients (160) were females and two third (66%) of patients were between third and sixth decade of life.

INTRODUCTION
Th yroid gland is affl icted by various pathologies ranging from diff use enlargement (goiter) to nodular lesions and thyroiditis.
Nodules are the cause of maximum concern among all thyroid pathologies because of their malignant potential. Th ere is a wide range of malignant potential among the clinically or radiologically detected thyroid nodules. Th e average prevalence of malignancy rates across the world in thyroid nodules, as shown by invasive procedure ranges from 4.0 to 6.5% 1,2 .
Th e American Th yroid Association defi nes thyroid nodules as "discrete lesions within the thyroid gland, radiologically distinct from surrounding thyroid parenchyma" 3 . Th ese nodules may be fi rst identifi ed clinically or incidentally on thyroid imaging (incidentalomas) 4 .
Earlier in the absence of any standardized system of USG reporting. there was diffi cult communication between radiologists and endocrinologists. Th en, the introduction of the Th yroid Imaging Reporting and Data System (TIRADS) by the American College of Radiologists 5 was an attempt to study and compare sonographic fi ndings of thyroid nodules to cytological fi ndings. Th is system was originally proposed by Horvath et al. 6 as a risk stratifi cation system and subsequently modifi ed by Jin Kwak et al. 7 into a more practical and reproducible format.
Th e fi ne-needle aspiration method for studying the thyroid was fi rst developed in Sweden in the Rudiunhelmet hospital of Stockholm in the 1950s 9 . Bethesda classifi cation system established a standar-FNAC as the gold standard) using the below formulas and were found to be as follows: -

DISCUSSION
Th e assessment of thyroid is a multipronged approach involving history with clinical examination, thyroid function tests, ultrasound thyroid and US-guided FNAC. However until recently, the USG evaluation of thyroid lesions was less standardized. Th en with the introduction of TIRADS and the endorsement of this classifi cation system by the American College of Radiologists 6 , we have a reproducible risk stratifi cation system similar to BIRADS grading system for breast lesions. Th is system attempts to correlate sonographic features to cytological classifi cation and gives an estimate of chances of malignancy in a particular nodule.
Th e sonological features included in our study are echogenicity, microcalcifi cations, taller than wider shape, presence of suspicious lymph node, irregular margins, and peripheral halo. It is to be emphasized here been tabulated in Table 1.
For all the calculations, TIRADS 4/ 5 and BETHES-DA 4/5/6 considered positive for malignancy and the rest as negative.
Out of 116 nodules belonging to TIRADS 2, none turned out to be Bethesda 4/5/6, thus risk of malignancy in TIRADS 2 was 0%. Out of the two indeterminate FNAC (Bethesda III) amongst these, one patient insisted on repeat FNAC where it was shown be benign hyperplastic nodule (Bethesda II) and other showed no increase in size on interval follow up on 3/6/9 months.
Th e risk of a malignant FNAC in TIRADS 3, 4 and 5 was 4.5, 38. 4 Table 4 compares the test characteristics (sensitivity, specifi city, PPV and NPV) of our screening test (TI-RADS) with those of prior studies by Singaporewalla et al. 18 and Periakaruppan et al. 15 and shows similar results.
One of the shortcomings of our study was that TI-RADS 4 was considered as a single class and not classifi ed into 4A, 4B and 4C which vary widely as far as malignant potential is considered.

CONCLUSION
TIRADS is an ultrasound-based eff ective risk stratifi cation system correlating cytological and radiological features. It can predict the chances of a particular thyroid nodule for being malignant to a great extent. Especially keeping in mind its high negative predictive value and the fact that TIRADS 2 nodules have 0% chances of malignancy, FNAC can be deferred in these TIRADS 2 patients. It will also help reduce the burden on pathology department as these patients form the bulk of nodules reporting for thyroid FNAC. that it is not the presence or the absence of a single feature on ultrasound which is important. Rather the presence of at least two of the USG features is more accurate in diff erentiating a benign nodule and a highrisk nodule for malignancy 11,12 .
In our study, we found that more than the half of thyroid nodules (56%) belonged to TIRADS 2 having 0% chances of malignancy. So we can take a more informed decision about subsequent FNAC which although is a readily available and inexpensive test but also minimally invasive at the same time. But it is a prerequisite that the radiologist performing ultrasound and guided FNACs has had a good learning curve and has audited his/her results in comparison to FNAC.
Comparing our results with previous studies, female predisposition 13 was similar to the studies of G. Periakaruppan et al. 14 and K P Gupta et al. 15 where 84% and 74% patients were females respectively. Th is might be explained by the eff ect of estrogen and progesterone, as pregnancy has been shown to be related to increased nodule size and new nodule development 16 .
Comparative evaluation of risk of malignancy for each TIRADS grade was done with previous studies and is shown in Table 3.
Th e fi ndings are similar in all except for slightly reduced percentage of malignancy in TIRADS 4 and 5 of Singaporewalla et al. 17 as here only Bethesda 5 and 6 were included in malignant and not Bethesda 4 unlike others.  Table 3. Comparative evaluation of risk of malignancy amongst various TIRADS grades