Abstract
Background: Thyroid nodules are a common clinical finding, and accurate preoperative risk stratification is essential to avoid unnecessary surgery while ensuring timely detection of malignancy. The Bethesda System for Reporting Thyroid Cytopathology (TBSRTC) and the American College of Radiology Thyroid Imaging Reporting and Data System (ACR TI-RADS) are widely used but are frequently applied in isolation. We evaluated their correlation, agreement, and combined diagnostic utility in nodular thyroid disease.
Methods: This hospital-based cross-sectional study enrolled 80 consecutive patients with nodular thyroid lesions at a tertiary care centre over 18 months. Each nodule underwent ultrasonographic ACR TI-RADS categorisation and fine-needle aspiration cytology (FNAC) reported per TBSRTC (2023 edition); serum triiodothyronine (T3), thyroxine (T4), and thyroid-stimulating hormone (TSH) were measured. Agreement (Cohen’s kappa), diagnostic accuracy, receiver operating characteristic (ROC) analysis, and multivariable logistic regression were performed.
Results: The mean age was 46 ± 16.4 years with marked female predominance (86.25%). Bethesda II (87.5%) and TI-RADS 2 (57.5%) were the commonest categories. The risk of malignancy rose progressively across TI-RADS categories (0% in TR1-TR2 to 50% in TR5). Overall cyto-radiological concordance was 85% (κ = 0.42, moderate). For predicting malignancy, TI-RADS showed sensitivity 77.8%, specificity 90.3%, negative predictive value (NPV) 93.3%, and accuracy 87.5% (area under the curve (AUC) 0.91), while Bethesda cytology showed specificity 87.3% and NPV 95.4%. Bethesda category (odds ratio (OR) 3.8), TI-RADS category (OR 2.6), and elevated serum T3 (OR 1.9) were independent predictors of malignancy. Concordant high-risk categorisation was uniformly malignant.
Conclusion: ACR TI-RADS and the Bethesda system are complementary. Their integrated application improves malignancy risk stratification, reliably excludes malignancy in low-risk nodules, and may reduce unnecessary diagnostic interventions.
Keywords: bethesda system, fine-needle aspiration cytology, risk of malignancy, thyroid nodule, thyroid ultrasonography, tirads
Introduction
Thyroid nodules are among the most frequent endocrine findings. They are palpable in roughly 5% of adults but are detectable by high-resolution ultrasonography in up to 50-67%, whereas only approximately 5-15% prove malignant [1,2]. The central clinical challenge is to identify this small malignant fraction reliably without subjecting the large benign majority to unnecessary surgery. Fine-needle aspiration cytology (FNAC) remains the cornerstone investigation, and the Bethesda System for Reporting Thyroid Cytopathology (TBSRTC) provides a standardised six-tier framework linking cytology to defined malignancy risk and management [3,4]. In parallel, the American College of Radiology Thyroid Imaging Reporting and Data System (ACR TI-RADS) stratifies sonographic malignancy risk and guides the decision to aspirate [5,6].
Although both systems are individually validated, they are frequently interpreted in isolation, and discordance is well recognised, particularly for intermediate-risk nodules where imaging and cytology may diverge [6,7]. Serum thyroid hormones have additionally been explored as adjunctive markers: thyroid-stimulating hormone (TSH) is a recognised growth factor for follicular epithelium, and higher serum TSH, even within the normal range, has been associated with an increased risk of differentiated thyroid cancer [8,9], providing a rationale for examining the biochemical profile alongside imaging and cytology. Data integrating the updated 2023 Bethesda classification with ACR TI-RADS and biochemistry in Indian tertiary-care populations remain limited. We therefore evaluated the cyto-radiological correlation between ACR TI-RADS and TBSRTC and quantified their agreement. Because histopathological confirmation was not available in this cross-sectional cohort, cytology necessarily served as a surrogate reference standard, and the performance metrics reported here describe agreement with cytology rather than accuracy against tissue diagnosis.
Materials and methods
Study design, setting, and duration
This hospital-based, cross-sectional observational study was conducted in the Department of Pathology, Kanti Devi Medical College, Hospital and Research Centre (KDMCH&RC), Mathura, Uttar Pradesh, in collaboration with the Department of Radiology, over 18 months. Reporting follows the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement.
Participants and selection
Consecutive patients with nodular thyroid lesions who were referred for FNAC and provided written informed consent were enrolled, irrespective of age or sex; those unwilling or unable to consent were excluded. Because enrolment was confined to nodules already selected for aspiration, the sample is enriched for clinically or sonographically suspicious lesions; this spectrum effect is acknowledged in the Limitations.
Sample size
An a priori estimate using n = Z²p(1−p)/d² (Z = 1.96, assumed malignancy prevalence p = 0.10, absolute precision d = 0.067) yielded 77, and 80 patients were enrolled. This formula sizes the study to estimate a single prevalence proportion; it does not directly power the agreement (κ) or the sensitivity/specificity objectives, whose precision depends on the number of high-risk nodules. As only nine high-risk aspirates were obtained, this event count is the binding precision constraint and is reflected in the wide confidence intervals reported below.
Radiological evaluation
All patients underwent grey-scale thyroid ultrasonography as part of routine clinical care in the Department of Radiology. Nodules were categorised using the ACR TI-RADS system, in which points across five feature categories - composition, echogenicity, shape, margin, and echogenic foci - are summed to assign categories TR1 (0 points), TR2 (2 points), TR3 (3 points), TR4 (4-6 points), and TR5 (≥7 points). The ACR TI-RADS category for each nodule was obtained from the corresponding routine clinical ultrasound report issued by the department; scanning and reporting were undertaken by the department's radiologists during standard care rather than by a single investigator dedicated to the study, and the imaging and cytology assessments were not formally blinded to one another. FNAC was performed and reported independently by the Department of Pathology.
Cytological evaluation
FNAC was palpation-guided using a 23-gauge needle, with ultrasound guidance for non-palpable, deep, or initially non-diagnostic nodules; cystic aspirates were centrifuged and smears prepared from the sediment. At least four smears per patient were wet-fixed in 95% ethanol for Papanicolaou and haematoxylin-eosin staining or air-dried for May-Grünwald-Giemsa staining, applying the Bethesda adequacy threshold of six groups of well-visualised follicular cells. Cytology was reported per TBSRTC, 2023 edition.
Biochemical assessment
Serum free triiodothyronine (T3), free thyroxine (T4) (both pmol/L), and TSH (mIU/L) were measured for all patients on a Cobas e 411 fully automated electrochemiluminescence immunoassay (ECLIA) analyser (Roche Diagnostics GmbH, Mannheim, Germany), using institutional reference intervals of 3.10-6.80 pmol/L, 12-22 pmol/L, and 0.27-4.20 mIU/L respectively.
Outcome definitions
As histopathological confirmation was not available, FNAC served as the surrogate reference standard and ACR TI-RADS as the index test. A single dichotomisation was applied throughout: Bethesda IV-VI denoted cytologically high-risk and Bethesda II low-risk; ACR TI-RADS was dichotomised as TR4-TR5 (high risk) versus TR1-TR3 (low risk). The one nondiagnostic aspirate (Bethesda I) was excluded from all classification analyses, as category I is nondiagnostic rather than benign and cannot be assigned a risk stratum; no atypia-of-undetermined-significance (Bethesda III) cases occurred. Thus 79 nodules entered the correlation, agreement, and performance analyses. Category-wise risk of malignancy (ROM) was defined separately as the proportion of Bethesda V-VI among all nodules within a given TI-RADS category. Because Bethesda served as the reference standard, its own accuracy was not assessed and it was not entered as a predictor in any model, to avoid circularity.
Statistical analysis
Analyses used SPSS v27.0 (IBM Corp., Armonk, NY, USA). Normality was assessed with the Shapiro-Wilk test; all three biochemical variables were non-normal (T3 W = 0.327; T4 W = 0.430; TSH W = 0.276; all p < 0.001), so non-parametric methods were used. Categorical variables are reported as n/N (%) with 95% CIs by the Wilson score method, and continuous variables as median (IQR) or mean ± SD. TI-RADS-cytology associations were tested by Pearson's chi-square with degrees of freedom, substituting Fisher's exact (2×2) or the Fisher-Freeman-Halton exact test (5×2) where more than 20% of cells had expected counts below five. Ordered trend across ascending TI-RADS categories was assessed by the Cochran-Armitage test. Serum hormones were compared between high-risk and low-risk nodules by the Mann-Whitney U test, reporting U, the standardised z, the exact two-sided p, and the rank-biserial correlation (r) as an effect size. Agreement was quantified by Cohen's κ with a 95% CI from the asymptotic standard error (Landis-Koch interpretation). Receiver operating characteristic (ROC) analysis used the ordinal ACR TI-RADS category (TR1-TR5) as the test variable and cytologically high-risk status (Bethesda IV-VI) as the state variable, with the area under the curve (AUC) and its 95% CI by the Hanley-McNeil method, tested against 0.50. A two-sided p < 0.05 was considered significant.
Ethics
The study was approved by the Institutional Ethics Committee of KDMCH&RC (approval no. KDMCHRC/PG/IEC/2024/37), and written informed consent was obtained from all participants.
Results
Demographic and clinical profile
Among 80 enrolled patients, the mean age was 42.5 ± 16.3 years (median 39; range 12-80), the largest group being 21-40 years (36/80, 45.0%). Sixty-nine of 80 (86.3%) were female (female-to-male ratio ≈ 6.3:1). Neck swelling was universal (80/80, 100%) and dysphagia was reported by 13/80 (16.3%). Solitary nodules predominated (73/80, 91.3%) and most were firm (70/80, 87.5%) (Table 1).
Table 1. Baseline demographic and clinical characteristics (N = 80).
| Characteristic | Category | n/N (%) |
| Age, years | mean ± SD (range) | 42.5 ± 16.3 (12-80) |
| median | 39 | |
| Age group | ≤20 years | 6/80 (7.5) |
| 21-40 years | 36/80 (45.0) | |
| 41-60 years | 24/80 (30.0) | |
| >60 years | 14/80 (17.5) | |
| Sex | Female | 69/80 (86.3) |
| Male | 11/80 (13.8) | |
| Presenting complaint | Neck swelling | 80/80 (100.0) |
| Dysphagia | 13/80 (16.3) | |
| Nodule type | Solitary | 73/80 (91.3) |
| Multinodular | 7/80 (8.8) | |
| Consistency | Firm | 70/80 (87.5) |
| Hard | 10/80 (12.5) |
Cytological and radiological distribution
No Bethesda III (atypia of undetermined significance) cases occurred. Bethesda II was the commonest category (70/80, 87.5%), comprising colloid goitre (42/70, 60.0%), colloid goitre with cystic change (16/70, 22.9%), multinodular goitre with cystic degeneration (10/70, 14.3%), and adenomatoid nodule (2/70, 2.9%). Higher categories were infrequent: Bethesda V (5/80, 6.3%), IV (2/80, 2.5%), VI (2/80, 2.5%), and I (1/80, 1.3%). Sonographically, TR2 predominated (46/80, 57.5%), followed by TR3 (11/80, 13.8%), TR4 (11/80, 13.8%), TR1 (8/80, 10.0%), and TR5 (4/80, 5.0%) (Table 2). Representative FNAC photomicrographs illustrating the observed cytological categories are provided in the Appendix.
Table 2. Cross-tabulation of ACR TI-RADS against Bethesda categories, with category-wise high-risk cytology and risk of malignancy (ROM).
B = Bethesda category. ACR TI-RADS = American College of Radiology Thyroid Imaging Reporting and Data System. High-risk cytology = Bethesda IV–VI. ROM = risk of malignancy, defined as Bethesda V–VI among all nodules in that TI-RADS category. 95% CIs by the Wilson score method.
| TIRADS | n/N (%) | B-I | B-II | B-IV | B-V | B-VI | High-risk cytology (B IV–VI), n/N (%) (95% CI) | ROM (B V–VI), n/N (%) (95% CI) |
| TR1 | 8/80 (10.0) | 0 | 8 | 0 | 0 | 0 | 0/8 (0.0) (0.0–32.4) | 0/8 (0.0) (0.0–32.4) |
| TR2 | 46/80 (57.5) | 0 | 46 | 0 | 0 | 0 | 0/46 (0.0) (0.0–7.7) | 0/46 (0.0) (0.0–7.7) |
| TR3 | 11/80 (13.8) | 0 | 8 | 0 | 3 | 0 | 3/11 (27.3) (9.7–56.6) | 3/11 (27.3) (9.7–56.6) |
| TR4 | 11/80 (13.8) | 1 | 7 | 1 | 2 | 0 | 3/11 (27.3) (9.7–56.6) | 2/11 (18.2) (5.1–47.7) |
| TR5 | 4/80 (5.0) | 0 | 1 | 1 | 0 | 2 | 3/4 (75.0) (30.1–95.4) | 2/4 (50.0) (15.0–85.0) |
| Total | 80 (100) | 1 | 70 | 2 | 5 | 2 | 9/80 (11.3) | 7/80 (8.8) |
Cyto-radiological correlation
Cytological risk category was strongly associated with TIRADS category (Fisher-Freeman-Halton exact p < 0.001). High-risk cytology (Bethesda IV-VI) was absent in TR1 (0/8; 95% CI 0.0-32.4) and TR2 (0/46; 95% CI 0.0-7.7), and present in 3/11 (27.3%; 9.7-56.6) TR3, 3/11 (27.3%; 9.7-56.6) TR4, and 3/4 (75.0%; 30.1-95.4) TR5 nodules, with a significant ordered trend (Cochran-Armitage z = 4.84; χ²(1) = 23.42; p < 0.001) (Figure 1).
Figure 1. Cyto-radiological correlation between ACR TI-RADS and Bethesda categories.
ACR TI-RADS = American College of Radiology Thyroid Imaging Reporting and Data System
Risk of malignancy
The category-wise ROM (Bethesda V-VI) was 0/8 (0.0%) in TR1, 0/46 (0.0%) in TR2, 3/11 (27.3%; 9.7-56.6) in TR3, 2/11 (18.2%; 5.1-47.7) in TR4, and 2/4 (50.0%; 15.0-85.0) in TR5. Although the overall trend was significant (z = 3.82; χ²(1) = 14.58; p < 0.001), the series was not strictly monotonic, dipping from TR3 to TR4; with only 11 nodules per intermediate category the confidence intervals are wide and overlapping, and these estimates are compatible with an underlying monotonic gradient rather than evidence against it.
Performance of ACR TI-RADS against cytology
Using cytology as the surrogate reference, TR4-TR5 showed sensitivity 6/9 (66.7%; 95% CI 35.4-87.9), specificity 62/71 (87.3%; 77.6-93.2), positive predictive value 6/15 (40.0%; 19.8-64.3), negative predictive value 62/65 (95.4%; 87.3-98.4), and overall accuracy 68/80 (85.0%; 75.6-91.2). The association was significant (χ²(1) = 15.28, p < 0.001; Fisher's exact p = 0.001; unadjusted odds ratio 13.78). ROC analysis using the ordinal TI-RADS category yielded an AUC of 0.913 (95% CI 0.784-1.000; z = 6.26, p < 0.001) (Figure 2, Table 3). Consistent with the small number of high-risk nodules, the AUC interval extends to the boundary value of 1.000, indicating an unstable estimate that should be read as directional rather than precise.
Table 3. Diagnostic performance of ACR TI-RADS (TR4-TR5) against cytology as reference standard, with agreement (N = 80).
Reference standard: fine-needle aspiration cytology (FNAC) (Bethesda IV-VI = high risk; I-II = low risk). χ²(1) = 15.28, p < 0.001; Fisher's exact p = 0.001; odds ratio 13.78. Area under the curve (AUC) 95% CI by Hanley–McNeil method; κ 95% CI from asymptotic standard error (SE 0.155). Proportion CIs by Wilson score method. ACR TI-RADS = American College of Radiology Thyroid Imaging Reporting and Data System
| Parameter | n/N (%) | 95% CI |
| Sensitivity | 6/9 (66.7) | 35.4-87.9 |
| Specificity | 62/71 (87.3) | 77.6-93.2 |
| Positive predictive value | 6/15 (40.0) | 19.8-64.3 |
| Negative predictive value | 62/65 (95.4) | 87.3-98.4 |
| Overall accuracy | 68/80 (85.0) | 75.6-91.2 |
| Area under ROC curve | 0.913 | 0.784-1.000 |
| Cohen's κ (agreement) | 0.418 | 0.115-0.722 |
Figure 2. Receiver operating characteristic (ROC) curve for ACR-TIRADS predicting cytologically high-risk status.
AUC = area under the curve. ACR TI-RADS = American College of Radiology Thyroid Imaging Reporting and Data System
Agreement
Cohen's κ between dichotomised TI-RADS and cytological risk was 0.418 (95% CI 0.115-0.722; SE 0.155), indicating moderate agreement, with 85.0% observed versus 74.2% chance-expected agreement. Concordance was 68/80 (85.0%; 75.6-91.2) and discordance 12/80 (15.0%; 8.8-24.4), comprising nine TR4-TR5 nodules with low-risk cytology and three TR3 nodules with Bethesda V cytology.
Biochemical parameters
Serum T3 was higher in cytologically high-risk than low-risk nodules (median 8.50 [IQR 5.70-8.60] vs 5.10 [3.15-6.40] pmol/L; U = 502.0, z = 2.78, p = 0.006, r = +0.571). TSH was marginally higher (3.20 [1.89-4.80] vs 2.40 [1.46-3.04] mIU/L; U = 449.0, z = 1.97, p = 0.049, r = +0.405), whereas free T4 did not differ (13.50 [10.50-15.00] vs 12.00 [9.64-14.40] pmol/L; U = 344.5, z = 0.38, p = 0.709). Serum T3 exceeded the reference range in 6/9 (66.7%; 35.4-87.9) high-risk versus 10/71 (14.1%; 7.8-24.0) low-risk nodules, and TSH in 4/9 (44.4%; 18.9-73.3) versus 4/71 (5.6%; 2.2-13.6). The high-risk T3 distribution reflected a cluster of three nodules at 8.5-8.6 pmol/L together with two markedly elevated values (15.9 pmol/L each), so the group median is sensitive to individual cases (Table 4).
Table 4. Serum thyroid hormone profile by cytological risk category (N = 80).
U = Mann-Whitney U statistic; p = exact two-sided p value; r = rank-biserial correlation (effect size). Reference intervals: T3 3.10-6.80 pmol/L; T4 12-22 pmol/L; TSH 0.27-4.20 mIU/L. Values in square brackets are 95% CIs (Wilson score method). T3 = triiodothyronine. T4 = thyroxine. TSH = thyroid-stimulating hormone.
| Parameter | High-risk (B IV-VI), n=9, median (IQR) | Low-risk (B I-II), n=71, median (IQR) | U | p | r |
| T3 (pmol/L) | 8.50 (5.70-8.60) | 5.10 (3.15-6.40) | 502 | 0.0056 | 0.571 |
| T4 (pmol/L) | 13.50 (10.50-15.00) | 12.00 (9.64-14.40) | 344.5 | 0.709 | 0.078 |
| TSH (mIU/L) | 3.20 (1.89-4.80) | 2.40 (1.46-3.04) | 449 | 0.049 | 0.405 |
| T3 above range | 6/9 (66.7) (35.4-87.9) | 10/71 (14.1) (7.8-24.0) | - | - | - |
| TSH above range | 4/9 (44.4) (18.9-73.3) | 4/71 (5.6) (2.2-13.6) | - | - | - |
Discussion
In this cross-sectional cohort, ACR TI-RADS category was strongly and monotonically associated with cytological risk on trend testing: high-risk cytology was absent from TR1-TR2 and rose across ascending categories. Agreement between the dichotomised systems was, however, only moderate (κ = 0.418), so the two are best regarded as complementary rather than interchangeable. Throughout, cytology served as a surrogate reference in the absence of histopathology, and the metrics below therefore describe agreement with cytology, not accuracy against tissue.
The demographic profile - middle-aged predominance and a female-to-male ratio of about 6.3:1 - matches the recognised epidemiology of nodular thyroid disease and comparable Indian and international series [10,11]. The high proportion of Bethesda II and low-risk TI-RADS categories reflects the benign-weighted population of a tertiary pathology service, in line with George et al. and Reuters et al. [12,13].
The graded rise in cytological risk across TI-RADS categories reproduces the pattern of Periakaruppan et al. and others [10,14]. The dip in the ROM series from TR3 (27.3%) to TR4 (18.2%) is best read as sampling noise: with 11 nodules per intermediate category the intervals overlap almost completely, and the formal trend across all five categories remained significant. Bethesda V-VI carried the highest cytological risk, consistent with their established malignancy rates [15,16].
The performance of ACR TI-RADS - specificity 87.3%, negative predictive value 95.4%, and AUC 0.913 - accords with published series and identifies its principal value as an exclusion and triage instrument rather than a rule-in test [14,17,18]. The high negative predictive value is the clinically useful property: a low-risk sonographic category reliably identifies nodules with low-risk cytology and supports conservative follow-up. This is in agreement with benign cytology, not exclusion of histological malignancy; a benign (Bethesda II) aspirate itself retains an approximately 2-7% residual false-negative risk, so a small malignant fraction may still be missed [3,19]. The modest positive predictive value (40.0%) means a high-risk sonographic category should prompt aspiration rather than surgery.
Discordance in 15.0% of cases reflects the different substrates the modalities interrogate: TI-RADS captures structural sonographic suspicion, whereas Bethesda reflects sampled cellular morphology. Sampling error in heterogeneous or cystically degenerated nodules can yield benign cytology despite suspicious imaging, matching the predominance of TR4 nodules with benign cytology among our discordant cases [20]. Importantly, Bethesda IV (follicular neoplasm) is inherently indeterminate: cytology cannot distinguish follicular adenoma from carcinoma, which depends on histological capsular or vascular invasion, so these nodules mandate diagnostic surgical excision, with intra-operative frozen section or formal histopathology as the arbiter [21]. For this reason category IV was treated as high-risk cytology requiring tissue rather than as confirmed malignancy. Interobserver variability in ultrasound interpretation contributes further, particularly across TR3-TR4 [6,22]; the moderate κ observed here is consistent with this literature.
Serum T3 was higher in cytologically high-risk nodules, but this observation should be interpreted with caution and regarded as hypothesis-generating. The high-risk group is small (n = 9), and its elevated T3 is influenced substantially by two nodules with markedly high values; moreover, the group showed higher T3 and higher TSH simultaneously, which is not readily reconciled with normal hypothalamic-pituitary-thyroid feedback, and autonomously hyperfunctioning (high-T3) nodules are classically benign. The available prior evidence concerns TSH rather than T3: higher serum TSH, even within the normal range, has been linked to differentiated thyroid cancer [8,9,23], whereas the TSH difference we observed was only marginal (p = 0.049) and there is no direct prior support for the T3 association. These biochemical signals are therefore adjunctive at most and require prospective, histology-anchored validation before any clinical use. Overall, our findings support an integrated model in which TI-RADS rationalises the decision to aspirate and Bethesda delivers the cytological verdict, an approach reported to improve specificity when the systems are combined [14,24,25], while histopathology remains the necessary reference for definitive validation.
Set against comparable series, our 85.0% concordance and moderate agreement are consistent with other single-centre studies combining the two systems [20,25], including analyses confined to nodules smaller than 4 cm [26] and those correlating individual sonographic descriptors rather than composite scores with cytology [27]; across these reports the recurring pattern is that imaging and cytology agree best at the extremes of risk and least in intermediate categories. Adoption of the 2023 Bethesda revision, which refined the risk strata of the 2017 edition [28], should further improve cross-study comparability.
Strengths
The study applied the current 2023 Bethesda edition alongside standardised ACR TI-RADS scoring in a consecutively sampled cohort. A single dichotomisation of cytological risk was applied uniformly; the index test was kept methodologically distinct from the reference standard to avoid circularity. Reporting follows STROBE, and complete statistical reporting-test statistics, degrees of freedom, U values, effect sizes, and 95% confidence intervals for all proportions, the AUC, and Cohen's κ-permits independent appraisal. Radiological, cytological, and biochemical data were complete for all analysed nodules.
Limitations
The principal limitation is the absence of histopathological confirmation: FNAC is an imperfect reference standard with its own false-negative and indeterminate-rate limitations, so the reported sensitivity, specificity, and ROC describe agreement with cytology rather than accuracy against tissue, and the two Bethesda IV nodules could not be resolved as benign or malignant. Enrolment was restricted to nodules already referred for aspiration, introducing spectrum bias that constrains the range of disease against which TI-RADS was tested and limits generalisability. The cohort yielded only nine high-risk nodules, so estimates are imprecise (sensitivity CI 35.4-87.9; AUC upper bound truncated at 1.000; κ CI 0.115-0.722) and category-specific risks for TR3-TR5 rest on very small denominators; for the same reason no multivariable model was fitted. Ultrasonography and cytology were not formally blinded to one another, which may inflate the observed agreement. The study was single-centre and single-state; interobserver variability among sonologists was not quantified; and molecular testing (e.g., BRAFV600E) was not performed.
An additional limitation concerns the source of the imaging data. Thyroid ultrasonography was performed and reported as part of routine clinical care, and the ACR TI-RADS categories analysed here were taken from these routine radiology reports rather than assigned by a single, study-dedicated radiologist re-reading archived images under standardised, blinded conditions. As a result, the assigned categories could not be independently re-verified, and study-specific ultrasound images were not archived and therefore cannot be presented. This records-based approach reflects real-world practice but limits standardisation of the index test; prospective evaluation with dedicated, blinded image review and retained images is required for confirmation.
Conclusions
ACR TI-RADS category correlated strongly with cytological risk on trend testing, while categorical agreement between the two systems was moderate; low-risk sonographic categories were reliably associated with low-risk cytology. ACR TI-RADS therefore appears a promising exclusion and triage tool, pending histology-anchored validation, whereas its modest positive predictive value indicates that a high-risk category should prompt aspiration rather than direct surgical decision-making. Because cytology, not histopathology, was the reference standard, these findings describe agreement with a cytological surrogate. The biochemical associations were exploratory and inconclusive. Prospectively, integrated cyto-radiological assessment may help rationalise the decision to aspirate; whether it reduces unnecessary intervention was not measured here and requires purpose-designed, histology-anchored study.
Acknowledgments
The authors thank the Department of Radiology, Kanti Devi Medical College, Hospital and Research Centre, Mathura, for performing and reporting the thyroid ultrasound examinations from which the ACR TI-RADS categories in this study were obtained. During the preparation of this manuscript, the authors used Claude (Anthropic, PBC, San Francisco, CA, USA) for language editing and formatting. All authors reviewed and verified all outputs and take full responsibility for the integrity and accuracy of the work.
Appendices
Figure 3. Colloid goiter-sheets of thyroid follicular cells.
May-Grünwald-Giemsa stain, 40x
Figure 4. Colloid goiter with cystic changes, scattered follicular epithelial cells cystic macrophages are seen along with colloid.
May-Grünwald-Giemsa stain, 40x
Figure 5. Follicular neoplasm- thyroid fine-needle aspiration (FNA) shows small follicles composed of cells with nuclear enlargement.
May-Grünwald-Giemsa stain, 40x
Figure 6. Crowding seen in follicular neoplasm .
May-Grünwald-Giemsa stain, 40x
Figure 7. Suspicious of papillary carcinoma of thyroid.
May-Grünwald-Giemsa stain, 10x
Figure 8. Suspicious of papillary carcinoma of thyroid.
May-Grünwald-Giemsa stain, 40x
Figure 9. Papillary thyroid carcinoma- mix of flat sheets and rounded, papillary-like fragments with fibrovascular cores.
May-Grünwald-Giemsa stain, 10x
Disclosures
Human subjects: Informed consent for treatment and open access publication was obtained or waived by all participants in this study. Institute Ethics Committee, Kanti Devi Medical College and Research Centre, Mathura, Uttar Pradesh, India issued approval KDMCHRC/PG/IEC/2024/37.
Animal subjects: All authors have confirmed that this study did not involve animal subjects or tissue.
Conflicts of interest: In compliance with the ICMJE uniform disclosure form, all authors declare the following:
Payment/services info: All authors have declared that no financial support was received from any organization for the submitted work.
Financial relationships: All authors have declared that they have no financial relationships at present or within the previous three years with any organizations that might have an interest in the submitted work.
Other relationships: All authors have declared that there are no other relationships or activities that could appear to have influenced the submitted work.
Author Contributions
Concept and design: Juhi M. Singh, Praneeta J. Singh
Acquisition, analysis, or interpretation of data: Juhi M. Singh, Praneeta J. Singh, Madhusudan P. Singh
Drafting of the manuscript: Juhi M. Singh, Madhusudan P. Singh
Critical review of the manuscript for important intellectual content: Juhi M. Singh, Praneeta J. Singh, Madhusudan P. Singh
Supervision: Juhi M. Singh, Praneeta J. Singh
References
- 1.Zamora EA, Khare S, Cassaro S. StatPearls. Treasure Island: StatPearls Publishing; 2025. Thyroid nodule. [PubMed] [Google Scholar]
- 2.Clinical practice. The thyroid nodule. Hegedüs L. N Engl J Med. 2004;351:1764–1771. doi: 10.1056/NEJMcp031436. [DOI] [PubMed] [Google Scholar]
- 3.The Bethesda System for Reporting Thyroid Cytopathology: a cytohistological study. Anand B, Ramdas A, Ambroise MM, Kumar NP. J Thyroid Res. 2020;2020:8095378. doi: 10.1155/2020/8095378. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.The 2023 Bethesda System for Reporting Thyroid Cytopathology. Ali SZ, Baloch ZW, Cochand-Priollet B, Schmitt FC, Vielh P, VanderLaan PA. Thyroid. 2023;33:1039–1044. doi: 10.1089/thy.2023.0141. [DOI] [PubMed] [Google Scholar]
- 5.ACR Thyroid Imaging, Reporting and Data System (TI-RADS): white paper of the ACR TI-RADS committee. Tessler FN, Middleton WD, Grant EG, et al. J Am Coll Radiol. 2017;14:587–595. doi: 10.1016/j.jacr.2017.01.046. [DOI] [PubMed] [Google Scholar]
- 6.Interobserver agreement of various thyroid imaging reporting and data systems. Grani G, Lamartina L, Cantisani V, Maranghi M, Lucia P, Durante C. Endocr Connect. 2018;7:1–7. doi: 10.1530/EC-17-0336. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.2015 American Thyroid Association management guidelines for adult patients with thyroid nodules and differentiated thyroid cancer: what is new and what has changed? Haugen BR. Cancer. 2017;123:372–381. doi: 10.1002/cncr.30360. [DOI] [PubMed] [Google Scholar]
- 8.Higher serum thyroid stimulating hormone level in thyroid nodule patients is associated with greater risks of differentiated thyroid cancer and advanced tumor stage. Haymart MR, Repplinger DJ, Leverson GE, Elson DF, Sippel RS, Jaume JC, Chen H. J Clin Endocrinol Metab. 2008;93:809–814. doi: 10.1210/jc.2007-2215. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Could serum TSH levels predict malignancy in euthyroid patients affected by thyroid nodules with indeterminate cytology? Cappelli C, Pirola I, Gandossi E, et al. Int J Endocrinol. 2020;2020:7543930. doi: 10.1155/2020/7543930. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Correlation between ultrasound-based TIRADS and Bethesda System for Reporting Thyroid-cytopathology: 2-year experience at a tertiary care center in India. Periakaruppan G, Seshadri KG, Vignesh Krishna GM, Mandava R, Sai VP, Rajendiran S. Indian J Endocrinol Metab. 2018;22:651–655. doi: 10.4103/ijem.IJEM_27_18. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Histological, cytological, and radiological correlation in thyroid lesions. Adil Z, Choudhary R, Sulya M, Koshariya M, Tekam M. Cureus. 2025;17:0. doi: 10.7759/cureus.98286. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Correlation of TIRADS and Bethesda scoring systems with final histopathology of thyroid nodules — an institutional experience. George NA, Suresh S, Jiji V, et al. Indian J Otolaryngol Head Neck Surg. 2022;74:5753–5758. doi: 10.1007/s12070-021-02380-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Bethesda classification and cytohistological correlation of thyroid nodules in a Brazilian thyroid disease center. Reuters KB, Mamone MC, Ikejiri ES, et al. Eur Thyroid J. 2018;7:133–138. doi: 10.1159/000488104. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.The combination of ACR-Thyroid Imaging Reporting and Data system and The Bethesda System for Reporting Thyroid Cytopathology in the evaluation of thyroid nodules-an institutional experience. Sakthisankari S, Vidhyalakshmi S, Shanthakumari S, Devanand B, Nagul U. Cytopathology. 2021;32:472–481. doi: 10.1111/cyt.12967. [DOI] [PubMed] [Google Scholar]
- 15.The Bethesda system of reporting cytopathology of thyroid nodules in correlation with radiological and clinicopathological findings. Najeeb DF, Ali HH. https://www.google.com/url?sa=t&source=web&rct=j&opi=89978449&url=https://sciencescholar.us/journal/index.php/ijhs/article/download/12361/8877/8989&ved=2ahUKEwi9hNfl5c-VAxWszzgGHTN5DUoQFnoECBcQAQ&usg=AOvVaw2A--CJURMyRCIrONwp70aK Int J Health Sci. 2022;6:957–967. [Google Scholar]
- 16.Correlation study of cytology and radiology (ultrasound) in diagnosing thyroid nodules using Bethesda and TIRADS. Rohilla L, Baghla A, Yadav A, Garg A. https://impactfactor.org/PDF/IJPCR/16/IJPCR,Vol16,Issue8,Article238.pdf Int J Pharm Clin Res. 2024;16:1473–1477. [Google Scholar]
- 17.Association between TI-RADS ultrasound categories and BETHESDA cytology categories of thyroid lesions. Kumari S, Suresh TN, Sakalecha AK, Azeem Mohiyuddin SM. Indian J Cancer. 2024;61:775–780. doi: 10.4103/ijc.ijc_479_21. [DOI] [PubMed] [Google Scholar]
- 18.Accuracy of ultrasound scans as compared to fine needle aspiration cytology in the diagnosis of thyroid nodules. Salman MT, AlGhazzawi MS, Al-Kamil EA, Al-Salmi S, Yousuf MS, Abdulla TS. Cureus. 2023;15:0. doi: 10.7759/cureus.35108. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Diagnostic accuracy of fine needle aspiration cytology of thyroid nodules. Erkinuresin T, Demirci H. Diagnosis (Berl) 2020;7:61–66. doi: 10.1515/dx-2019-0039. [DOI] [PubMed] [Google Scholar]
- 20.Correlation between Thyroid Imaging Reporting and Data System and Bethesda System of Reporting of Thyroid Cytopathology of Thyroid Nodule: a single center experience. Biswas A, Basu K, De S, Karmakar S, De D, Sengupta M, Ghosh S. J Cytol. 2020;37:193–199. doi: 10.4103/JOC.JOC_57_19. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Can the American Thyroid Association, K-TIRADS, and ACR-TIRADS ultrasound systems predict malignancy in Bethesda category IV nodules? Yang W, Fananapazir G, LaRoy J, Wilson M, Campbell MJ. Endocr Pract. 2020;26:945–952. doi: 10.4158/EP-2020-0024. [DOI] [PubMed] [Google Scholar]
- 22.Interobserver variability in ultrasound assessment of thyroid nodules. Alyami J, Almutairi FF, Aldoassary S, Albeshry A, Almontashri A, Abounassif M, Alamri M. Medicine (Baltimore) 2022;101:0. doi: 10.1097/MD.0000000000031106. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Serum thyrotropin concentration as a novel predictor of malignancy in thyroid nodules investigated by fine-needle aspiration. Boelaert K, Horacek J, Holder RL, Watkinson JC, Sheppard MC, Franklyn JA. J Clin Endocrinol Metab. 2006;91:4295–4301. doi: 10.1210/jc.2006-0527. [DOI] [PubMed] [Google Scholar]
- 24.Accuracy of the combination of TI-RADS and Bethesda regarding histopathology of thyroid malignancy. Ciputra E, Syamsu S, Smaradania N, Hamid F, Faruk M. Asian Pac J Cancer Biol. 2025;10:245–249. [Google Scholar]
- 25.Clinico-pathological correlation of thyroid nodule ultrasound and cytology using the TIRADS and Bethesda classifications. Singaporewalla RM, Hwee J, Lang TU, Desai V. World J Surg. 2017;41:1807–1811. doi: 10.1007/s00268-017-3919-5. [DOI] [PubMed] [Google Scholar]
- 26.The diagnostic utility of the ultrasound TIRADS and Bethesda System of reporting thyroid cytology in thyroid nodule smaller than 4.0 cm: a retrospective analysis. Hagag HM, Ismail KA, Bakhuraysah MM, et al. Medicine (Baltimore) 2025;104:0. doi: 10.1097/MD.0000000000043264. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Individual ultrasonographic characteristics of thyroid nodules and their cytopathological correlation to determine malignancy risk. Castilla Villanueva MÁ, Solis Cano DG, Amador Martínez A, Téliz Meneses MA, Baquera-Heredia J, Vallin Orozco CE, Loya Ceballos M. Cureus. 2024;16:0. doi: 10.7759/cureus.63918. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.The 2017 Bethesda System for Reporting Thyroid Cytopathology. Cibas ES, Ali SZ. Thyroid. 2017;27:1341–1346. doi: 10.1089/thy.2017.0500. [DOI] [PubMed] [Google Scholar]