Thyroid nodules are frequently encountered in clinical practice and are increasingly detected because of the widespread use of high-resolution ultrasonography and other cross-sectional imaging modalities. Although most thyroid nodules are benign, a small but clinically important proportion represent thyroid malignancy. The principal challenge in their evaluation is therefore to accurately identify nodules requiring fine-needle aspiration cytology (FNAC) or surgical intervention while avoiding unnecessary invasive procedures in patients with benign lesions. Clinical examination alone has limited ability to distinguish benign from malignant nodules; consequently, ultrasonography has become the primary imaging modality for their detection, characterization and risk stratification [1]. High-resolution ultrasonography provides information regarding nodule composition, echogenicity, shape, margins and echogenic foci. Certain sonographic characteristics, including solid composition, marked hypoechogenicity, irregular or infiltrative margins, a taller-than-wide shape, punctate echogenic foci and suspicious cervical lymphadenopathy, are associated with a higher probability of malignancy. However, variability in the interpretation and reporting of these features may affect diagnostic consistency. To standardize thyroid ultrasonography and improve communication between radiologists, clinicians and pathologists, the American College of Radiology introduced the Thyroid Imaging Reporting and Data System (ACR TI-RADS) [2]. ACR TI-RADS assigns points to five sonographic domains composition, echogenicity, shape, margin and echogenic foci. The cumulative score classifies nodules as TR1 (benign), TR2 (not suspicious), TR3 (mildly suspicious), TR4 (moderately suspicious) or TR5 (highly suspicious). Recommendations for FNAC or imaging follow-up are determined by the TI-RADS category and maximum nodule diameter. This structured approach aims to improve diagnostic reproducibility, estimate malignancy risk and reduce unnecessary biopsies of low-risk nodules. Evidence from systematic reviews indicates that ACR TI-RADS offers good diagnostic performance and can decrease the number of avoidable FNAC procedures while maintaining high sensitivity for clinically significant malignancies [3]. FNAC remains the standard minimally invasive procedure for the cytological assessment of thyroid nodules. The Bethesda System for Reporting Thyroid Cytopathology classifies FNAC findings into six diagnostic categories, each associated with an estimated risk of malignancy and recommended clinical management [4]. Correlation between TI-RADS categories and Bethesda cytological findings may determine whether increasing ultrasonographic suspicion is accompanied by progressively suspicious or malignant cytology. Current guidelines support the combined use of clinical assessment, ultrasound risk stratification and FNAC for evidence-based management of thyroid nodules [5]. AIM To assess thyroid nodules using the ACR TI-RADS classification and determine their correlation with FNAC findings. OBJECTIVES 1. To characterize thyroid nodules according to their ultrasonographic composition, echogenicity, shape, margins and echogenic foci. 2. To classify thyroid nodules into ACR TI-RADS categories and FNAC findings into Bethesda diagnostic categories. 3. To determine the association and diagnostic agreement between ACR TI-RADS categories and FNAC findings.

Source of Data The study data were obtained from patients with thyroid nodules who attended the outpatient and inpatient departments of JMF’s ACPM Medical College & Hospital, Dhule, Maharashtra, and were referred to the Department of Radiodiagnosis and Imaging for thyroid ultrasonography and ultrasound-guided FNAC. Clinical, ultrasonographic and cytopathological information was obtained from patient interviews, clinical records, ultrasound examination findings and FNAC reports. Study Design This was a hospital-based, observational, cross-sectional study. Study Location The study was conducted in the Department of Radiodiagnosis and Imaging in collaboration with the Department of Pathology at JMF’s ACPM Medical College & Hospital, Morane (P.L.), Sakri Road, Dhule, Maharashtra, India. Study Duration The study was conducted over a period of 12 months Participant enrolment, ultrasonographic evaluation, FNAC, cytological assessment, data entry and statistical analysis were completed during this period. Sample Size A total of 160 patients with thyroid nodules who fulfilled the eligibility criteria and underwent both ultrasonographic TI-RADS assessment and FNAC were included. The sample size could be estimated using the formula for a single proportion: n=(Z_(1-α/2)^2×p(1-p))/d^2 where: n= required sample size Z_(1-α/2)= 1.96 at a 95% confidence level p= anticipated proportion or diagnostic sensitivity obtained from a previous study d= allowable absolute error After accounting for inadequate or nondiagnostic FNAC samples, incomplete records and possible exclusions, the final sample size was fixed at 160 participants. Consecutive eligible patients were enrolled until the required sample size was achieved. Inclusion Criteria Patients of either sex presenting with one or more thyroid nodules. Patients undergoing thyroid ultrasonography and FNAC during the study period. Patients whose nodules could be categorized according to ACR TI-RADS. Patients with available cytopathological findings for the evaluated nodule. Patients who provided written informed consent to participate. Exclusion Criteria Patients who did not undergo FNAC after ultrasonographic evaluation. Patients with diffuse thyroid disease without a discrete measurable nodule. Patients previously diagnosed with or treated for thyroid malignancy. Patients who had undergone previous thyroid surgery, radioiodine therapy or ablative treatment of the evaluated nodule. Predominantly cystic lesions without an adequate solid component for cytological sampling. Patients with incomplete ultrasonographic or cytopathological records. Patients who declined to provide written informed consent. Patients in whom FNAC was contraindicated or could not be performed safely. Procedure and Methodology After obtaining approval from the Institutional Ethics Committee, eligible patients were approached and the purpose and procedure of the study were explained. Written informed consent was obtained before enrolment. Demographic characteristics, including age and sex, and relevant clinical information, including duration of swelling, pain, dysphagia, change in voice, rapid increase in size, family history of thyroid disease or malignancy, previous radiation exposure and thyroid-related symptoms, were recorded using a structured data-collection form. Patients were examined in the supine position with the neck mildly extended by placing a pillow beneath the shoulders. Thyroid ultrasonography was performed using a high-resolution ultrasound machine equipped with a high-frequency linear-array transducer. Both thyroid lobes and the isthmus were evaluated in transverse and longitudinal planes. The maximum dimensions of each nodule were measured in three orthogonal planes. The number and location of nodules, thyroid parenchymal characteristics and the presence of abnormal cervical lymph nodes were documented. Each thyroid nodule was assessed for the five ACR TI-RADS sonographic domains: Composition: cystic or almost completely cystic, spongiform, mixed cystic-solid, or solid/almost completely solid. Echogenicity: anechoic, hyperechoic/isoechoic, hypoechoic or markedly hypoechoic. Shape: wider-than-tall or taller-than-wide. Margin: smooth, ill-defined, lobulated/irregular or demonstrating extrathyroidal extension. Echogenic foci: none or large comet-tail artefacts, macrocalcifications, peripheral/rim calcifications or punctate echogenic foci. Points assigned to these features were added to obtain the final ACR TI-RADS score. Nodules were categorized as TR1, TR2, TR3, TR4 or TR5. The maximum nodule diameter and the corresponding recommendation for FNAC or follow-up were also recorded in accordance with ACR TI-RADS criteria [2]. In patients with multiple nodules, the nodule that underwent FNAC was considered the index nodule. When more than one nodule was sampled, each nodule was documented separately and matched precisely with its corresponding cytological result; however, patient-level analysis used a predefined index nodule, preferably the nodule with the highest TI-RADS category. Ultrasound-guided FNAC was performed under aseptic precautions. The patient was placed supine with the neck extended, and the skin overlying the target nodule was cleaned with an antiseptic solution. The needle was advanced into the solid or most suspicious component of the nodule under real-time ultrasound guidance, avoiding major blood vessels, cystic or necrotic areas. Multiple passes were made when necessary to obtain adequate cytological material. Gentle pressure was applied at the puncture site after the procedure, and the patient was observed for immediate complications. Sample Processing The aspirated material was immediately expelled onto clean, labelled glass slides. Direct smears were prepared by gently spreading the material between two slides. Air-dried smears were stained using May-Grünwald-Giemsa stain, while alcohol-fixed smears were stained using Papanicolaou or haematoxylin and eosin stain according to the departmental protocol. For cystic nodules, the fluid was aspirated and centrifuged when required, and smears were prepared from the sediment. A cell block was prepared whenever sufficient material was available and additional evaluation was considered necessary. All slides were examined by an experienced cytopathologist who was preferably blinded to the final TI-RADS category to minimize interpretation bias. FNAC findings were classified using the Bethesda System for Reporting Thyroid Cytopathology: Bethesda I - Nondiagnostic Bethesda II - Benign Bethesda III - Atypia of undetermined significance Bethesda IV - Follicular neoplasm Bethesda V - Suspicious for malignancy Bethesda VI - Malignant Nondiagnostic samples were subjected to repeat ultrasound-guided FNAC whenever clinically indicated. For primary analysis, Bethesda II was considered cytologically benign, Bethesda V and VI were considered cytologically suspicious/malignant, and Bethesda III and IV were retained as indeterminate categories and analysed separately. They were not automatically classified as malignant. Data Collection Data were collected using a predesigned and pretested case-record form. The following variables were recorded: Age and sex Presenting symptoms and relevant clinical history Solitary or multiple thyroid nodules Nodule location and three-dimensional measurements Ultrasound characteristics included in ACR TI-RADS Total TI-RADS score and category Presence of suspicious cervical lymph nodes FNAC procedure details and adequacy of the sample Bethesda cytological category Cytological diagnosis Complications, if any, following FNAC Each participant was assigned a unique identification number. Ultrasound and FNAC findings were matched using this identification number and nodule location. Data were checked for completeness and consistency before being entered into an electronic spreadsheet. Personal identifiers were removed from the analytical dataset to maintain confidentiality. Statistical Methods Data were analysed using an appropriate statistical software package, such as IBM SPSS Statistics, R or equivalent software. Continuous variables were assessed for normality and presented as mean with standard deviation or median with interquartile range, as appropriate. Categorical variables were summarized using frequencies and percentages. The distribution of patients according to ACR TI-RADS and Bethesda categories was presented in cross-tabulations. The association between TI-RADS categories and FNAC findings was assessed using the chi-square test. Fisher’s exact test was used when expected cell frequencies were less than five. A linear-by-linear association test was applied to determine whether increasing TI-RADS categories were associated with progressively suspicious Bethesda findings. Spearman’s rank correlation coefficient was calculated to assess the ordinal correlation between TI-RADS and Bethesda categories. Agreement between dichotomized TI-RADS assessment and FNAC findings was evaluated using Cohen’s kappa coefficient with a 95% confidence interval. For diagnostic analysis, a predefined TI-RADS threshold, such as TR4 or higher, was considered test-positive, while Bethesda V-VI was considered cytologically suspicious or malignant. Sensitivity, specificity, positive predictive value, negative predictive value and diagnostic accuracy were calculated with 95% confidence intervals. Receiver operating characteristic analysis was performed, and the area under the curve with its 95% confidence interval was reported. An additional sensitivity analysis could be performed using TR5 as the positive threshold. Where appropriate, univariable and multivariable logistic regression analyses were performed to identify ultrasonographic predictors of suspicious or malignant cytology. Results were expressed as odds ratios with 95% confidence intervals. A two-tailed p value <0.05 was considered statistically significant.

Table 1. Overall clinical and imaging profile of thyroid nodules and association with suspicious/malignant FNAC findings (N=160) Study parameter Overall (N=160) Bethesda V-VI (n=28) Bethesda I-IV (n=132) Effect estimate (95% CI) Test of significance P value Age, years, Mean (SD) 42.16 (13.65) 46.82 (12.47) 41.17 (13.68) MD=5.65 (0.32-10.98) Welch t=2.14 0.038* Female sex 103 (64.4%) 20 (71.4%) 83 (62.9%) OR=1.48 (0.60-3.60) χ²=0.74 0.391 Male sex 57 (35.6%) 8 (28.6%) 49 (37.1%) Reference Solitary thyroid nodule 97 (60.6%) 21 (75.0%) 76 (57.6%) OR=2.21 (0.88-5.56) χ²=2.94 0.087 Multiple thyroid nodules 63 (39.4%) 7 (25.0%) 56 (42.4%) Reference Palpable neck swelling 121 (75.6%) 24 (85.7%) 97 (73.5%) OR=2.17 (0.71-6.63) χ²=1.86 0.173 Pain or discomfort 19 (11.9%) 4 (14.3%) 15 (11.4%) OR=1.30 (0.40-4.23) Fisher’s exact test 0.748 Dysphagia 17 (10.6%) 6 (21.4%) 11 (8.3%) OR=3.00 (1.01-8.95) Fisher’s exact test 0.047* Hoarseness of voice 9 (5.6%) 4 (14.3%) 5 (3.8%) OR=4.23 (1.06-16.86) Fisher’s exact test 0.049* Rapid increase in nodule size 21 (13.1%) 9 (32.1%) 12 (9.1%) OR=4.74 (1.76-12.75) χ²=10.77 0.001* Suspicious cervical lymph nodes 13 (8.1%) 8 (28.6%) 5 (3.8%) OR=10.16 (3.02-34.17) Fisher’s exact test <0.001* Maximum nodule diameter, cm, Mean (SD) 2.23 (1.10) 2.67 (1.18) 2.14 (1.06) MD=0.53 (0.04-1.02) Welch t=2.20 0.034* Nodule diameter ≥2.5 cm 71 (44.4%) 17 (60.7%) 54 (40.9%) OR=2.23 (0.97-5.14) χ²=3.62 0.057 χ²: Pearson’s chi-square test; OR: odds ratio; MD: mean difference; CI: confidence interval. *Statistically significant at p<0.05. Among the 160 participants, the mean age was 42.16±13.65 years, and females constituted the majority (64.4%). Patients with Bethesda V-VI cytology were significantly older than those with Bethesda I-IV cytology (46.82±12.47 versus 41.17±13.68 years; MD=5.65 years, 95% CI: 0.32-10.98; p=0.038). Female sex was more frequent in the Bethesda V-VI group than in the Bethesda I-IV group (71.4% versus 62.9%), although the association was not statistically significant (OR=1.48; p=0.391). Solitary nodules were observed in 60.6% of patients and were more common among Bethesda V-VI cases (75.0% versus 57.6%); however, the association did not reach statistical significance (OR=2.21; p=0.087). Palpable neck swelling was the most frequent clinical presentation (75.6%) but was not significantly associated with suspicious or malignant cytology (p=0.173). Dysphagia (OR=3.00; p=0.047), hoarseness of voice (OR=4.23; p=0.049), rapid increase in nodule size (OR=4.74; p=0.001) and suspicious cervical lymph nodes (OR=10.16; p<0.001) were significantly associated with Bethesda V-VI cytology. The mean maximum nodule diameter was also significantly greater in Bethesda V-VI nodules than in Bethesda I-IV nodules (2.67±1.18 versus 2.14±1.06 cm; MD=0.53 cm, 95% CI: 0.04-1.02; p=0.034). Although nodules measuring ≥2.5 cm were more frequent in the Bethesda V-VI group (60.7% versus 40.9%), the association was marginally nonsignificant (OR=2.23; p=0.057). Table 2. Ultrasonographic characteristics of thyroid nodules and their association with suspicious/malignant cytology (N=160) Ultrasonographic characteristic Overall, n (%) Bethesda V-VI (n=28) Bethesda I-IV (n=132) Odds ratio (95% CI) Test of significance P value Composition Cystic/almost completely cystic 9 (5.6%) 0 (0.0%) 9 (6.8%) Fisher’s exact test 0.365 Spongiform 17 (10.6%) 0 (0.0%) 17 (12.9%) Fisher’s exact test 0.047 Mixed cystic-solid 41 (25.6%) 2 (7.1%) 39 (29.5%) Reference Solid/almost completely solid 93 (58.1%) 26 (92.9%) 67 (50.8%) 12.61 (2.88-55.30) χ²=16.82 <0.001* Echogenicity Anechoic 8 (5.0%) 0 (0.0%) 8 (6.1%) Reference Hyperechoic/isoechoic 61 (38.1%) 3 (10.7%) 58 (43.9%) 0.21 (0.06-0.74)† Fisher’s exact test 0.011* Hypoechoic 54 (33.8%) 7 (25.0%) 47 (35.6%) 0.58 (0.23-1.49)† χ²=1.16 0.281 Markedly hypoechoic 37 (23.1%) 18 (64.3%) 19 (14.4%) 10.71 (4.30-26.68) χ²=32.34 <0.001* Shape Wider-than-tall 131 (81.9%) 11 (39.3%) 120 (90.9%) Reference Taller-than-wide 29 (18.1%) 17 (60.7%) 12 (9.1%) 15.45 (5.90-40.48) χ²=41.48 <0.001* Margins Smooth 79 (49.4%) 4 (14.3%) 75 (56.8%) Reference Ill-defined 43 (26.9%) 4 (14.3%) 39 (29.5%) 1.92 (0.45-8.14) Fisher’s exact test 0.469 Lobulated/irregular 34 (21.3%) 19 (67.9%) 15 (11.4%) 16.47 (6.32-42.92)‡ χ²=44.06 <0.001* Extrathyroidal extension 4 (2.5%) 1 (3.6%) 3 (2.3%) 1.59 (0.16-15.90) Fisher’s exact test 0.536 Echogenic foci None/large comet-tail artefacts 83 (51.9%) 5 (17.9%) 78 (59.1%) Reference Macrocalcifications 23 (14.4%) 7 (25.0%) 16 (12.1%) 2.42 (0.89-6.59)§ χ²=3.11 0.078 Peripheral/rim calcifications 13 (8.1%) 3 (10.7%) 10 (7.6%) 1.46 (0.37-5.69) Fisher’s exact test 0.700 Punctate echogenic foci 41 (25.6%) 20 (71.4%) 21 (15.9%) 13.21 (5.14-33.94) χ²=37.36 <0.001* †Effect compared with all other echogenicity patterns combined. ‡Effect of irregular/lobulated margins compared with all other margin categories combined. §Effect of macrocalcifications compared with nodules without macrocalcifications. *Statistically significant at p<0.05.* Solid or almost completely solid composition was the most common pattern, occurring in 58.1% of nodules. It was present in 92.9% of Bethesda V-VI nodules compared with 50.8% of Bethesda I-IV nodules and was strongly associated with suspicious or malignant cytology (OR=12.61, 95% CI: 2.88-55.30; p<0.001). None of the cystic or spongiform nodules demonstrated Bethesda V-VI cytology. Regarding echogenicity, 38.1% of nodules were hyperechoic or isoechoic, 33.8% were hypoechoic, and 23.1% were markedly hypoechoic. Marked hypoechogenicity was identified in 64.3% of Bethesda V-VI nodules compared with 14.4% of Bethesda I-IV nodules and significantly increased the odds of suspicious or malignant cytology (OR=10.71, 95% CI: 4.30-26.68; p<0.001). In contrast, hyperechoic or isoechoic appearance was negatively associated with Bethesda V-VI cytology (OR=0.21; p=0.011). Most nodules were wider-than-tall (81.9%); however, a taller-than-wide shape was markedly more frequent in Bethesda V-VI nodules than in Bethesda I-IV nodules (60.7% versus 9.1%). This feature was associated with approximately 15-fold higher odds of suspicious or malignant cytology (OR=15.45, 95% CI: 5.90-40.48; p<0.001). Smooth margins were found in 49.4% of nodules, whereas lobulated or irregular margins were found in 21.3%. Irregular or lobulated margins occurred in 67.9% of Bethesda V-VI nodules and 11.4% of Bethesda I-IV nodules and constituted a strong predictor of suspicious or malignant cytology (OR=16.47, 95% CI: 6.32-42.92; p<0.001). Ill-defined margins and extrathyroidal extension were not significantly associated with Bethesda category, although the latter was uncommon and consequently had a wide confidence interval. Among echogenic foci, punctate echogenic foci were present in 25.6% of nodules and were substantially more frequent in Bethesda V-VI than Bethesda I-IV nodules (71.4% versus 15.9%). They were associated with approximately 13-fold higher odds of suspicious or malignant cytology (OR=13.21, 95% CI: 5.14-33.94; p<0.001). Macrocalcifications showed a possible but statistically nonsignificant association (OR=2.42; p=0.078), while peripheral or rim calcifications were not significantly associated with suspicious cytology (p=0.700). Table 3. Distribution of thyroid nodules according to ACR TI-RADS and Bethesda FNAC categories (N=160) A. ACR TI-RADS classification ACR TI-RADS category Interpretation n (%) 95% CI of proportion Test of significance P value TR1 Benign 11 (6.9%) 3.9%-11.9% TR2 Not suspicious 27 (16.9%) 11.9%-23.4% TR3 Mildly suspicious 43 (26.9%) 20.6%-34.2% TR4 Moderately suspicious 47 (29.4%) 22.9%-36.8% TR5 Highly suspicious 32 (20.0%) 14.5%-26.9% Total 160 (100.0%) χ² goodness-of-fit=25.38, df=4 <0.001* B. Bethesda cytological classification Bethesda category Cytological interpretation n (%) 95% CI of proportion Test of significance P value I Nondiagnostic 11 (6.9%) 3.9%-11.9% II Benign 88 (55.0%) 47.3%-62.5% III Atypia of undetermined significance 19 (11.9%) 7.7%-17.8% IV Follicular neoplasm 14 (8.8%) 5.3%-14.2% V Suspicious for malignancy 13 (8.1%) 4.8%-13.4% VI Malignant 15 (9.4%) 5.8%-14.9% Total 160 (100.0%) χ² goodness-of-fit=170.60, df=5 <0.001* According to the ACR TI-RADS classification, TR4 was the most frequent category, accounting for 47 (29.4%) nodules, followed by TR3 in 43 (26.9%), TR5 in 32 (20.0%), TR2 in 27 (16.9%) and TR1 in 11 (6.9%). Thus, 79 (49.4%) nodules were classified as moderately or highly suspicious (TR4-TR5). The observed TI-RADS distribution differed significantly from an equal distribution across the five categories (χ²=25.38, df=4; p<0.001). On cytological evaluation, Bethesda II was the predominant category, comprising 88 (55.0%) nodules. Bethesda III was reported in 19 (11.9%), Bethesda VI in 15 (9.4%), Bethesda IV in 14 (8.8%), Bethesda V in 13 (8.1%) and Bethesda I in 11 (6.9%) nodules. Overall, 28 (17.5%) nodules were classified as suspicious for malignancy or malignant (Bethesda V-VI), while 33 (20.6%) were cytologically indeterminate (Bethesda III-IV). The Bethesda category distribution differed significantly from an equal distribution (χ²=170.60, df=5; p<0.001), primarily because of the predominance of benign Bethesda II findings. Table 4. Association between ACR TI-RADS categories and Bethesda FNAC findings (N=160) A. Cross-tabulation of ACR TI-RADS and Bethesda categories ACR TI-RADS category Bethesda I Bethesda II Bethesda III Bethesda IV Bethesda V Bethesda VI Total TR1 1 (9.1%) 9 (81.8%) 1 (9.1%) 0 (0.0%) 0 (0.0%) 0 (0.0%) 11 TR2 2 (7.4%) 22 (81.5%) 2 (7.4%) 1 (3.7%) 0 (0.0%) 0 (0.0%) 27 TR3 3 (7.0%) 31 (72.1%) 5 (11.6%) 3 (7.0%) 1 (2.3%) 0 (0.0%) 43 TR4 3 (6.4%) 22 (46.8%) 8 (17.0%) 6 (12.8%) 5 (10.6%) 3 (6.4%) 47 TR5 2 (6.3%) 4 (12.5%) 3 (9.4%) 4 (12.5%) 7 (21.9%) 12 (37.5%) 32 Total 11 88 19 14 13 15 160 A progressive increase in cytological suspicion was evident with increasing ACR TI-RADS category. Among TR1 nodules, 81.8% were Bethesda II, and none were Bethesda V or VI. Similarly, 81.5% of TR2 nodules were Bethesda II, with no Bethesda V-VI cases. Among TR3 nodules, 72.1% were Bethesda II, and only one nodule (2.3%) was Bethesda V. In comparison, the proportion of benign Bethesda II findings declined to 46.8% among TR4 nodules and 12.5% among TR5 nodules. Bethesda V-VI findings increased from 17.0% in TR4 nodules to 59.4% in TR5 nodules. In particular, Bethesda VI cytology was observed in 37.5% of TR5 nodules compared with 6.4% of TR4 nodules and none of the TR1-TR3 nodules. B. Measures of association and diagnostic agreement Statistical measure Estimate 95% CI Test of significance P value Overall association between TI-RADS and Bethesda categories χ²=71.12, df=20 Pearson’s χ² test <0.001* Ordinal correlation between TI-RADS and Bethesda categories Spearman’s ρ=0.516 0.39-0.62 Spearman’s correlation test <0.001* Linear trend across increasing categories χ² for trend=39.87 Linear-by-linear association <0.001* Agreement between TR5 and Bethesda V-VI κ=0.674 0.54-0.81 Cohen’s kappa <0.001* Odds of Bethesda V-VI cytology in TR5 nodules OR=44.33 12.71-154.56 Pearson’s χ²=49.84 <0.001* There was a statistically significant overall association between ACR TI-RADS and Bethesda categories (χ²=71.12, df=20; p<0.001). A moderately strong positive ordinal correlation was observed between the two classification systems (Spearman’s ρ=0.516, 95% CI: 0.39-0.62; p<0.001), indicating that increasing TI-RADS categories were associated with increasingly suspicious cytological findings. The significant linear trend further supported this relationship (χ² for trend=39.87; p<0.001). Agreement between TR5 classification and Bethesda V-VI cytology was substantial (κ=0.674, 95% CI: 0.54-0.81; p<0.001). TR5 nodules had approximately 44 times higher odds of Bethesda V-VI cytology than TR1-TR4 nodules (OR=44.33, 95% CI: 12.71-154.56; p<0.001). C. Diagnostic performance of ACR TI-RADS TR5 for Bethesda V-VI cytology FNAC outcome TI-RADS TR5 TI-RADS TR1-TR4 Total Bethesda V-VI 19 9 28 Bethesda II 4 84 88 Total 23 93 116 Diagnostic parameter Estimate 95% CI Sensitivity 67.9% 49.3%-82.1% Specificity 95.5% 88.9%-98.2% Positive predictive value 82.6% 62.9%-93.0% Negative predictive value 90.3% 82.6%-94.8% Diagnostic accuracy 88.8% 81.8%-93.3% Positive likelihood ratio 14.93 5.53-40.29 Negative likelihood ratio 0.34 0.20-0.57 Diagnostic odds ratio 44.33 12.71-154.56 For diagnostic performance analysis, the 116 nodules classified as Bethesda II, V or VI were included, whereas nondiagnostic and indeterminate categories were excluded. Of the 28 Bethesda V-VI nodules, 19 were categorized as TR5, producing a sensitivity of 67.9% (95% CI: 49.3%-82.1%). Among 88 Bethesda II nodules, 84 were classified below TR5, resulting in a specificity of 95.5% (95% CI: 88.9%-98.2%). The positive predictive value was 82.6%, the negative predictive value was 90.3%, and the overall diagnostic accuracy was 88.8%. The positive likelihood ratio of 14.93 indicated that a TR5 classification produced a substantial increase in the probability of Bethesda V-VI cytology, while the negative likelihood ratio of 0.34 showed that a classification below TR5 reduced but did not eliminate the likelihood of suspicious or malignant cytology.

The present study evaluated 160 patients with thyroid nodules and found a mean age of 42.16±13.65 years, with a clear female predominance of 64.4%. This demographic pattern is consistent with the recognised higher prevalence of thyroid nodules among middle-aged women. Haugen et al. (2016)[1] and Gharib et al. (2016)[2] similarly observed that thyroid nodules are substantially more frequent among women and increase in prevalence with advancing age. In the present study, patients with Bethesda V-VI findings were significantly older than those with Bethesda I-IV findings by 5.65 years (p=0.038). Increasing age may therefore be associated with suspicious cytology in this population, although age alone is not considered sufficiently discriminative for biopsy selection. Barbosa et al. (2019)[8], in contrast, did not observe a significant association between age and histologically confirmed malignancy among indeterminate nodules.[8] This difference may reflect variations in patient selection, age distribution and the use of cytology rather than histopathology as the outcome in the present study. Although females constituted 71.4% of the Bethesda V-VI group, female sex was not significantly associated with suspicious or malignant cytology (OR=1.48; p=0.391). This finding supports the distinction between the prevalence and malignant potential of thyroid nodules: women develop nodules more frequently, but sex alone may not independently predict malignancy. Barbosa et al. (2019)[8] also reported no significant association between sex and final malignancy. Current international guidelines consequently do not use sex as an isolated criterion for FNAC but recommend integrated clinical and sonographic risk assessment.[1,2] Solitary nodules were present in 60.6% of the study population and were more frequent among Bethesda V-VI cases than Bethesda I-IV cases (75.0% versus 57.6%). However, the association did not reach statistical significance (OR=2.21; p=0.087). Haugen et al. (2016)[1] stated that the risk of malignancy should be assessed separately for every sonographically distinct nodule and that patients with multiple nodules may have a patient-level malignancy risk comparable to those with solitary nodules. Thus, the absence of a significant association between nodularity and cytology in the present study supports the principle that sonographic morphology is more informative than the solitary or multinodular presentation alone. Palpable neck swelling was the most common presenting feature, occurring in 75.6% of patients, but it was not significantly associated with Bethesda V-VI cytology. In contrast, dysphagia, hoarseness of voice, rapid enlargement and suspicious cervical lymph nodes were significantly associated with suspicious or malignant cytology. Dysphagia increased the odds approximately threefold, hoarseness increased the odds fourfold, and rapid enlargement increased the odds nearly fivefold. Suspicious cervical lymphadenopathy was the strongest clinical predictor, with an OR of 10.16. These findings are compatible with the recommendations of Haugen et al. (2016)[1], Gharib et al. (2016)[2] and Durante et al. (2023)[19], who identified rapid nodule growth, compressive symptoms, voice changes and abnormal cervical lymph nodes as clinical warning signs requiring prompt investigation. Nevertheless, rapid growth is not specific for malignancy, as haemorrhage or degeneration within a benign nodule can also cause an abrupt increase in size. The mean nodule diameter was significantly greater among Bethesda V-VI nodules than among Bethesda I-IV nodules (2.67±1.18 versus 2.14±1.06 cm; p=0.034). Barbosa et al. (2019)[8] also found an association between lesion size and malignancy among surgically treated indeterminate nodules. However, the dichotomous threshold of ≥2.5 cm was not independently significant in the present study (p=0.057). This borderline result indicates that nodule size should not be interpreted as a stand-alone marker of malignancy. Tessler et al. (2017)[3] incorporated size primarily to determine whether a nodule within a particular TI-RADS category should undergo FNAC or follow-up, rather than to assign malignancy points. Middleton et al. (2021)[14] further demonstrated that some malignant nodules may not meet ACR size thresholds for FNAC, although many are small and potentially indolent. Ultrasonographic characteristics Solid or almost completely solid composition was observed in 58.1% of all nodules and in 92.9% of Bethesda V-VI nodules. It was associated with a 12.61-fold increase in the odds of suspicious or malignant cytology. Conversely, none of the cystic or spongiform nodules demonstrated Bethesda V-VI cytology. These findings support the ACR TI-RADS scoring system proposed by Tessler et al. (2017)[3], in which solid composition receives two points, whereas completely cystic and spongiform appearances are assigned no points. Periakaruppan et al. (2018)[7] similarly reported that solid nodules were more frequently associated with higher Bethesda categories. However, composition alone is not sufficient for diagnosis because many benign adenomas and hyperplastic nodules are also predominantly solid. Marked hypoechogenicity was present in 64.3% of Bethesda V-VI nodules compared with 14.4% of Bethesda I-IV nodules, resulting in an OR of 10.71 (p<0.001). In comparison, hyperechoic or isoechoic nodules had significantly lower odds of suspicious cytology. Barbosa et al. (2019)[8] found that marked hypoechogenicity was significantly associated with histological malignancy among indeterminate nodules. Rago et al. (2022)[15] also identified marked hypoechogenicity as one of the most consistently reported sonographic features of thyroid cancer. Nevertheless, hypoechogenicity is not specific and can occur in benign nodules, thyroiditis and some follicular adenomas; its value increases when it occurs together with irregular margins, a taller-than-wide shape or punctate echogenic foci. A taller-than-wide shape was found in 60.7% of Bethesda V-VI nodules but only 9.1% of Bethesda I-IV nodules. It was associated with approximately 15-fold higher odds of suspicious or malignant cytology. This result is in agreement with Tessler et al. (2017)[3], who allocated three points to a taller-than-wide shape because it reflects growth across rather than along normal tissue planes. Middleton et al. (2018)[6] confirmed the importance of this feature in their multi-institutional evaluation of 3,422 thyroid nodules. Barbosa et al. (2019)[8] and Rago et al. (2022)[15] likewise identified a taller-than-wide configuration as an important malignancy predictor. Its relatively low frequency but strong association explains why it usually provides high specificity rather than high sensitivity. Lobulated or irregular margins were observed in 67.9% of Bethesda V-VI nodules and were associated with a 16.47-fold increase in suspicious or malignant cytology. This was the strongest individual morphological predictor in the present study. Irregular margins may represent uneven tumour growth or infiltration into the adjacent thyroid parenchyma. The finding agrees with the ACR TI-RADS framework of Tessler et al. (2017)[3], the EU-TIRADS recommendations of Russ et al. (2017)[4], and the analysis by Barbosa et al. (2019)[8], all of which consider irregular or infiltrative margins highly suspicious. Ill-defined margins, by contrast, were not significantly associated with Bethesda V-VI cytology. This distinction is clinically important because poor delineation may result from technical factors or benign parenchymal heterogeneity and should not automatically be interpreted as tumour invasion. Extrathyroidal extension was uncommon and was not significantly associated with suspicious cytology. The absence of significance should not be interpreted as evidence that the feature lacks clinical importance. Only four nodules demonstrated this finding, resulting in low statistical power and a very wide confidence interval. Both ACR TI-RADS and contemporary European guidelines regard definite extrathyroidal extension as a highly suspicious feature.[3,19] Punctate echogenic foci were found in 71.4% of Bethesda V-VI nodules compared with 15.9% of Bethesda I-IV nodules and increased the odds of suspicious or malignant cytology approximately 13-fold. Ha et al. (2019)[9] showed that the significance of echogenic foci depends on nodule composition and associated artefacts; punctate foci within solid nodules carried a substantially higher malignancy risk than comet-tail artefacts in cystic components. Rago et al. (2022)[15] similarly identified microcalcifications or punctate echogenic foci as consistently suspicious features. Macrocalcifications showed a weaker, statistically nonsignificant association in the present study (p=0.078), while peripheral calcifications were not associated with Bethesda category. These results support the differentiated point allocation in ACR TI-RADS, which assigns more weight to punctate echogenic foci than to macrocalcifications.[3] Collectively, the findings for composition, echogenicity, shape, margins and echogenic foci closely follow the biological and statistical rationale underlying ACR TI-RADS. The strongest indicators of Bethesda V-VI cytology were irregular margins, a taller-than-wide shape, punctate echogenic foci, solid composition and marked hypoechogenicity. Yang et al. (2020)[11], in a comparison of five ultrasound risk-stratification systems, similarly found that systems incorporating these high-risk characteristics achieved useful diagnostic discrimination, with ACR TI-RADS generally demonstrating comparatively high specificity. Distribution of ACR TI-RADS and Bethesda categories TR4 was the most frequent TI-RADS category in the present study (29.4%), followed by TR3 (26.9%), TR5 (20.0%), TR2 (16.9%) and TR1 (6.9%). Therefore, nearly half of all nodules were classified as TR4 or TR5. The relatively high proportion of suspicious nodules may be explained by referral and verification bias because the study included patients who underwent FNAC rather than an unselected ultrasound-screening population. Periakaruppan et al. (2018)[7] reported malignancy risks of 0%, 2.2%, 38.5% and 77.8% for TI-RADS categories 2, 3, 4 and 5, respectively, demonstrating the expected progressive increase in risk. Huang et al. (2023)[18] also observed that TR3 nodules were predominantly benign, whereas TR4 and particularly TR5 nodules showed substantially greater concordance with suspicious cytology or histological malignancy. Bethesda II was the most frequent cytological category, accounting for 55.0% of nodules. Bethesda III and IV accounted for 11.9% and 8.8%, respectively, while 17.5% were Bethesda V-VI. The predominance of Bethesda II cytology is consistent with the general observation that most biopsied thyroid nodules are benign. Cibas and Ali (2017)[5] standardized the six Bethesda categories and their associated malignancy risks, thereby facilitating consistent cytological reporting and clinical management. Dhar et al. (2023)[17], in a large series of 1,100 cases, similarly found a predominance of benign cytology, although the exact distribution varied because of differences in referral patterns, FNAC thresholds and institutional reporting practices. The nondiagnostic rate in the present study was 6.9%. This falls within an acceptable range for ultrasound-guided FNAC, although the result is influenced by nodule composition, operator experience, slide preparation and adequacy assessment. Nondiagnostic samples require repeat ultrasound-guided FNAC according to established recommendations.[1,5] The combined Bethesda III-IV proportion of 20.6% highlights the continuing challenge of indeterminate cytology. Barbosa et al. (2019)[8] demonstrated that ultrasound risk stratification can provide additional information in such nodules, but it should complement rather than replace repeat cytology, molecular testing or histopathological assessment. Correlation between TI-RADS and Bethesda categories The proportion of Bethesda II findings decreased progressively from 81.8% in TR1 and 81.5% in TR2 to 46.8% in TR4 and 12.5% in TR5. Conversely, Bethesda V-VI findings increased from none in TR1-TR2 and 2.3% in TR3 to 17.0% in TR4 and 59.4% in TR5. This graded relationship produced a significant overall association (χ²=71.12; p<0.001), a moderately strong positive correlation (Spearman’s ρ=0.516; p<0.001) and a significant linear trend. Periakaruppan et al. (2018)[7] reported a significant correlation between ultrasound TI-RADS and Bethesda cytology, particularly for benign nodules. Dhar et al. (2023)[17] also found a significant association and substantial agreement between the systems, with a kappa value of 0.688. This is remarkably close to the κ of 0.674 observed in the present study. Dhar et al. reported TI-RADS sensitivity of 65.8%, specificity of 96.5%, PPV of 91.8% and NPV of 84.4%.[17] The comparable agreement and performance indicate that standardized ultrasound classification corresponds closely with cytological risk, although the two systems assess different aspects of the lesion. The odds of Bethesda V-VI cytology were 44.33 times higher in TR5 nodules than in TR1-TR4 nodules. Huang et al. (2023)[18] likewise reported substantial concordance between ACR TI-RADS and Bethesda scoring and found a high histological malignancy rate among TR5 nodules. Barbosa et al. (2019)[8] showed that malignancy risk increased significantly with both ultrasound suspicion and Bethesda category, and that combining imaging with cytology improved risk stratification of indeterminate nodules. Using TR5 as the positive threshold, the present study demonstrated a sensitivity of 67.9%, specificity of 95.5%, PPV of 82.6%, NPV of 90.3% and accuracy of 88.8%. The high specificity indicates that TR5 is particularly useful for ruling in suspicious or malignant cytology, whereas its moderate sensitivity means that some Bethesda V-VI nodules occur below the TR5 threshold. Li et al. (2021)[13] concluded in their meta-analysis that ACR TI-RADS provides favourable sensitivity and moderate-to-high specificity, although performance depends on the selected positivity threshold. Zhang et al. (2020)[10] similarly found that ACR TI-RADS tended to achieve greater specificity than the ATA classification. Kang et al. (2022)[16] reported that the diagnostic balance of ACR TI-RADS changes according to whether TR4 or TR5 is used as the positive threshold. A TR5 threshold improves specificity and reduces unnecessary biopsies but lowers sensitivity, which closely explains the present results. Grani et al. (2019)[10] demonstrated that structured risk-stratification systems could reduce unnecessary FNAC procedures while preserving acceptable cancer detection. More recently, Joo et al. (2023)[17] found that biopsy criteria based on ultrasound risk categories and nodule size can reduce unnecessary procedures, but differences among systems involve a trade-off between missed malignancies and avoidable biopsies. The positive likelihood ratio of 14.93 indicated that TR5 substantially increased the probability of Bethesda V-VI cytology, while the negative likelihood ratio of 0.34 showed that a category below TR5 could not independently exclude suspicious disease. Therefore, TR3 and TR4 nodules should continue to be managed using nodule size, clinical risk factors and recommended FNAC thresholds. Middleton et al. (2021)[14] specifically cautioned that a small number of malignant nodules may fall below ACR biopsy criteria, reinforcing the importance of clinical context and follow-up. A key consideration is that the present diagnostic calculations used Bethesda V-VI as the positive reference and Bethesda II as the negative reference, while nondiagnostic and indeterminate categories were excluded. FNAC is not equivalent to definitive histopathology, especially for Bethesda III-IV lesions and follicular-patterned neoplasms. Consequently, the reported sensitivity, specificity and accuracy primarily represent the ability of TR5 to predict suspicious or malignant cytology rather than histologically proven cancer. Despite this limitation, the observed graded association and substantial agreement support the complementary use of ACR TI-RADS and Bethesda classification. Ultrasound can identify nodules requiring sampling, while cytology provides cellular-level risk categorization and guides subsequent clinical or surgical management. Overall, the study demonstrated that ACR TI-RADS effectively stratified thyroid nodules according to cytological risk. High-risk ultrasound features and increasing TI-RADS categories were strongly associated with Bethesda V-VI findings. The high specificity, positive likelihood ratio and overall accuracy of TR5 support its value for identifying nodules with a high probability of suspicious cytology.

The ACR TI-RADS classification demonstrated a significant positive correlation with Bethesda FNAC categories and effectively stratified thyroid nodules according to cytological risk. Solid composition, marked hypoechogenicity, a taller-than-wide shape, lobulated or irregular margins, and punctate echogenic foci were strongly associated with Bethesda V-VI cytology. Clinical features such as dysphagia, hoarseness, rapid nodule enlargement and suspicious cervical lymphadenopathy were also important indicators of suspicious or malignant cytology. TR5 classification showed high specificity, positive predictive value and overall diagnostic accuracy, although its moderate sensitivity indicated that suspicious cytology could occur in lower TI-RADS categories. ACR TI-RADS should therefore be used with clinical findings, nodule size and FNAC rather than as an independent diagnostic test. The combined application of TI-RADS and Bethesda reporting may improve patient selection for FNAC, reduce unnecessary invasive procedures and support standardized management of thyroid nodules. LIMITATIONS OF STUDY 1. The study was conducted at a single tertiary-care institution; therefore, its findings may not be generalizable to other hospitals or community-based populations. 2. The cross-sectional design permitted assessment of association but could not establish a temporal or causal relationship between sonographic features and cytological outcomes. 3. The sample size was relatively limited, particularly for Bethesda V-VI and less frequent ultrasound characteristics such as extrathyroidal extension. 4. Only patients who underwent FNAC were included, creating referral and verification bias and potentially increasing the proportion of sonographically suspicious nodules. 5. FNAC rather than surgical histopathology was used as the principal reference standard. Therefore, the findings reflected correlation with cytology and not definitive diagnostic accuracy for thyroid malignancy. 6. Bethesda I, III and IV findings were nondiagnostic or indeterminate and could not be conclusively classified as benign or malignant without repeat FNAC, molecular testing or histopathological examination. 7. Surgical histopathological follow-up was not available for all participants, preventing calculation of category-specific histologically confirmed malignancy rates. 8. Interobserver and intraobserver variability in identifying TI-RADS features was not assessed. 9. Ultrasound examinations and FNAC interpretation may have been influenced by operator experience, equipment quality, sampling technique and cytopathologist expertise. 10. Molecular testing was not performed for indeterminate nodules and long-term follow-up data on nodule growth or final clinical outcomes were unavailable

1. Haugen BR, Alexander EK, Bible KC, et al. 2015 American Thyroid Association management guidelines for adult patients with thyroid nodules and differentiated thyroid cancer. Thyroid. 2016;26(1):1-133. doi:10.1089/thy.2015.0020. 2. Gharib H, Papini E, Garber JR, et al. American Association of Clinical Endocrinologists, American College of Endocrinology, and Associazione Medici Endocrinologi medical guidelines for clinical practice for the diagnosis and management of thyroid nodules 2016 update. Endocr Pract. 2016;22(5):622-639. doi:10.4158/EP161208.GL. 3. Tessler FN, Middleton WD, Grant EG, et al. ACR Thyroid Imaging, Reporting and Data System (TI-RADS): White Paper of the ACR TI-RADS Committee. J Am Coll Radiol. 2017;14(5):587-595. doi:10.1016/j.jacr.2017.01.046. 4. Russ G, Bonnema SJ, Erdogan MF, Durante C, Ngu R, Leenhardt L. European Thyroid Association guidelines for ultrasound malignancy risk stratification of thyroid nodules in adults: the EU-TIRADS. Eur Thyroid J. 2017;6(5):225-237. doi:10.1159/000478927. 5. Cibas ES, Ali SZ. The 2017 Bethesda System for Reporting Thyroid Cytopathology. Thyroid. 2017;27(11):1341-1346. doi:10.1089/thy.2017.0500. 6. Middleton WD, Teefey SA, Reading CC, et al. Comparison of performance characteristics of American College of Radiology TI-RADS, Korean Society of Thyroid Radiology TIRADS, and American Thyroid Association guidelines. AJR Am J Roentgenol. 2018;210(5):1148-1154. doi:10.2214/AJR.17.18822. 7. Periakaruppan G, Seshadri KG, Vignesh Krishna GM, Mandava R, Sai VPM, Rajendiran S. Correlation between ultrasound-based TIRADS and Bethesda system for reporting thyroid-cytopathology: 2-year experience at a tertiary care center in India. Indian J Endocrinol Metab. 2018;22(5):651-655. doi:10.4103/ijem.IJEM_27_18. 8. Barbosa TLM, Junior COM, Graf H, et al. ACR TI-RADS and ATA US scores are helpful for the management of thyroid nodules with indeterminate cytology. BMC Endocr Disord. 2019;19(1):112. doi:10.1186/s12902-019-0429-5. 9. Ha SM, Ahn HS, Baek JH, et al. Echogenic foci in thyroid nodules: diagnostic performance with combination of TIRADS and echogenic foci. BMC Med Imaging. 2019;19(1):28. doi:10.1186/s12880-019-0322-0. 10. Zhang Q, Ma J, Sun W, Zhang L. Comparison of diagnostic performance between the American College of Radiology Thyroid Imaging Reporting and Data System and American Thyroid Association guidelines: a systematic review. Endocr Pract. 2020;26(5):552-563. doi:10.4158/EP-2019-0237. 11. Yang R, Zou X, Zeng H, Zhao Y, Ma X. Comparison of diagnostic performance of five different ultrasound TI-RADS classification guidelines for thyroid nodules. Front Oncol. 2020;10:598225. doi:10.3389/fonc.2020.598225. 12. Grani G, Lamartina L, Ascoli V, et al. Reducing the number of unnecessary thyroid biopsies while improving diagnostic accuracy: toward the "right" TIRADS. J Clin Endocrinol Metab. 2019;104(1):95-102. doi:10.1210/jc.2018-01674. 13. Li W, Wang Y, Wen J, Zhang L, Sun Y. Diagnostic performance of American College of Radiology TI-RADS: a systematic review and meta-analysis. AJR Am J Roentgenol. 2021;216(1):38-47. doi:10.2214/AJR.19.22691. 14. Middleton WD, Teefey SA, Tessler FN, et al. Analysis of malignant thyroid nodules that do not meet ACR TI-RADS criteria for fine-needle aspiration. AJR Am J Roentgenol. 2021;216(2):471-478. doi:10.2214/AJR.20.23123. 15. Rago T, Cantisani V, Ianni F, et al. Risk stratification of thyroid nodules: from ultrasound features to TIRADS. Cancers (Basel). 2022;14(3):717. doi:10.3390/cancers14030717. 16. Kang YJ, Ahn HY, Jang HY, et al. Diagnostic performance of ACR and Kwak TI-RADS for benign and malignant thyroid nodules: an updated systematic review and meta-analysis. Cancers (Basel). 2022;14(23):5961. doi:10.3390/cancers14235961. 17. Dhar L, Singh M, Jain SL, et al. Correlation of ultrasound-based TIRADS and the Bethesda system for reporting of thyroid cytopathology: a study in a tertiary care centre. Cytopathology. 2023;34(6):590-596. doi:10.1111/cyt.13273. 18. Huang EYF, Kao NH, Lin SY, et al. Concordance of the ACR TI-RADS classification with Bethesda scoring and histopathology risk stratification of thyroid nodules. JAMA Netw Open. 2023;6(9):e2331612. doi:10.1001/jamanetworkopen.2023.31612. 19. Durante C, Hegedüs L, Czarniecka A, et al. 2023 European Thyroid Association clinical practice guidelines for thyroid nodule management. Eur Thyroid J. 2023;12(5):e230067. doi:10.1530/ETJ-23-0067.