Thyroid nodules are common in clinical practice and are increasingly detected because of widespread use of high-resolution ultrasonography. Although most nodules are benign, accurate recognition of malignancy is essential to avoid both delayed cancer treatment and unnecessary thyroid surgery. Fine-needle aspiration cytology has become central to thyroid nodule evaluation because it is minimally invasive, inexpensive and capable of providing clinically actionable risk stratification. The Bethesda System for Reporting Thyroid Cytopathology (TBSRTC) standardizes reporting into six diagnostic categories and links each category to an estimated risk of malignancy and a management pathway [3]. The 2023 third edition retains six categories while simplifying terminology and updating malignancy-risk estimates. Despite its clinical value, FNAC is not infallible. Diagnostic performance varies because of sampling adequacy, operator technique, nodule characteristics, lesion heterogeneity and interpretive overlap. Cytology is particularly limited in follicular-patterned lesions because follicular carcinoma is defined by capsular and/or vascular invasion, features that require histological examination. A large systematic review and meta-analysis published in 2022 included 16,597 patients from 36 studies and found pooled sensitivity of 85.6%, specificity of 71.4%, and an area under the summary ROC curve of 86.1%, confirming good overall performance but substantial heterogeneity [4]. More recent Bethesda-based cytohistological studies continue to report wide variation in sensitivity and specificity, emphasizing the importance of understanding the methodological context in which diagnostic indices are calculated. This systematic review therefore evaluates contemporary evidence on the diagnostic accuracy of FNAC in thyroid nodules with histopathology as the reference standard, with particular emphasis on Bethesda-based reporting, cytology-histology discordance and risk of bias. 2. Objectives The primary objective was to determine the diagnostic accuracy of FNAC in differentiating benign from malignant thyroid nodules using histopathology as the reference standard. 1. Evaluate sensitivity and specificity of FNAC. 2. Evaluate PPV, NPV and overall diagnostic accuracy. 3. Assess cytohistological concordance across Bethesda categories. 4. Summarize risk of malignancy across the Bethesda spectrum. 5. Identify common causes of false-negative and false-positive FNAC results. 6. Evaluate methodological quality using QUADAS-2.

3.1 Reporting Framework The review was structured in accordance with the PRISMA 2020 statement [1]. Because this was a review of published literature, institutional ethics approval and individual patient consent were not required. 3.2 Review Question The diagnostic review question was defined using the population-index test-reference standard-target condition framework. 3.3 Information Sources and Search Strategy Structured searches were performed across PubMed/MEDLINE, Scopus and Web of Science, supplemented by reference-list screening. Search terms combined controlled vocabulary and free-text terms for thyroid nodules, fine-needle aspiration cytology, Bethesda classification, histopathology and diagnostic accuracy. A representative search string was: ("thyroid nodule" OR "thyroid lesion") AND ("fine needle aspiration cytology" OR FNAC OR FNAB) AND (histopathology OR histology OR cytohistological) AND (accuracy OR sensitivity OR specificity OR predictive value OR Bethesda) Searches were updated through June 2026. Reference lists of major systematic reviews and eligible studies were hand-searched for additional primary studies. 3.4 Eligibility Criteria Inclusion Criteria 1. Adult patients with thyroid nodules or thyroid swellings. 2. FNAC performed before definitive surgery or histopathological assessment. 3. Histopathology used as the diagnostic reference standard. 4. Cytological and histological diagnoses available for correlation. 5. Sufficient data to report or derive diagnostic performance, risk of malignancy or cytohistological concordance. 6. Original peer-reviewed diagnostic-accuracy or cytohistological cohort studies. Exclusion Criteria 1. Case reports, editorials, narrative reviews, systematic reviews and meta-analyses from the primary-study pool. 2. Studies without histopathological correlation. 3. Studies focused exclusively on molecular tests, artificial-intelligence classifiers or core-needle biopsy. 4. Paediatric-only cohorts without separately extractable adult data. 5. Duplicate or overlapping cohorts without distinct relevant outcome data. 6. Studies with insufficient diagnostic information. 3.5 Study Selection and PRISMA Flow After duplicate removal, titles and abstracts were screened against the eligibility criteria. Full texts of potentially relevant reports were then assessed. Reasons for exclusion at full-text stage were recorded. The numerical selection process is shown in Figure 1. 3.6 Data Extraction For each included study, the following were extracted where available: author, year, country, study design, number of FNACs, number with histopathological correlation, cytological reporting system, sensitivity, specificity, PPV, NPV, overall diagnostic accuracy, malignancy rates, and reported causes of cytology-histology discordance. 3.7 Diagnostic Outcome Definitions Sensitivity was defined as the proportion of histologically malignant lesions correctly identified by FNAC, while specificity was the proportion of histologically benign lesions correctly identified as non-malignant. PPV and NPV were interpreted according to each study's prespecified cytological threshold. Because studies differed in whether Bethesda III, IV and V were considered positive, indeterminate or excluded, performance estimates were interpreted in the context of the threshold used. 3.8 Risk-of-Bias Assessment Risk of bias was evaluated using QUADAS-2 [2], which assesses four domains: patient selection, index test, reference standard, and flow/timing. Each domain was judged as low, high or unclear risk of bias. Applicability concerns were also considered for patient selection, index test and reference standard. Particular attention was paid to partial verification bias because histopathology is disproportionately available in patients selected for surgery. 3.9 Data Synthesis Because included studies used different thresholds for classifying cytology as positive, varied in the handling of indeterminate Bethesda categories, and differed in the proportion of patients undergoing surgical verification, a de novo meta-analysis was not performed. A structured narrative synthesis was undertaken, and findings were contextualized using the large 2022 diagnostic meta-analysis by Hsiao et al. [4].

Table 1. Diagnostic Review Framework Component Definition Population Adults with thyroid nodules or thyroid swellings undergoing diagnostic evaluation Index test Fine-needle aspiration cytology Reference standard Histopathological examination of surgically excised thyroid tissue Target condition Thyroid malignancy Outcomes Sensitivity, specificity, PPV, NPV, accuracy, risk of malignancy and cytohistological discordance 4.1 PRISMA Study Selection The search identified 728 records. After removal of 186 duplicates, 542 records underwent title and abstract screening. Of these, 471 were excluded. Seventy-one full-text reports were assessed for eligibility, and 61 were excluded: no histopathological reference standard (n=18), insufficient diagnostic-accuracy data (n=14), review/meta-analysis/editorial publication type (n=10), focus on adjunct or core-needle biopsy rather than FNAC accuracy (n=7), paediatric-only population (n=4), duplicate/overlapping cohort (n=4), and unavailable full text (n=4). Ten studies were included in the qualitative synthesis. 4.2 Characteristics of Included Studies The ten primary studies were published between 2018 and 2026 and represented India, Türkiye, Azerbaijan, Lebanon and South Africa. Sample sizes and the proportion undergoing histopathological correlation varied widely. Most studies used a Bethesda-based reporting framework, although diagnostic thresholds differed. Table 2. Characteristics and Diagnostic Performance of Included Studies Study Country FNAC / Histology Correlation Sensitivity Specificity PPV NPV Accuracy Nandedkar et al., 2018 India 606 FNAC; 166 correlated 85.7% 98.6% NR NR 97.7% Anand et al., 2020 India 646 FNAC; 100 histology 72.4% 94.3% 84.0% 89.2% 87.9% Ucak et al., 2021 Türkiye 879 histology-correlated cases 84.7% 81.1% 74.1% 89.2% 82.5% Aliyev et al., 2022 Azerbaijan 738 surgical patients 97.4% 86.1% 96.4% 81.6% 94.8% Sen et al., 2023 India 1,763 FNAC; 444 histology 71.4% 49.5% NR NR NR Jain et al., 2023 India 330 FNAC; 38 histology 100% 60.0% 94.3% 100% 94.7% Osseis et al., 2023 Lebanon 344 thyroidectomy patients 89.3% 48.4% 78.0% 68.9% 75.9% Sureka et al., 2025 India 387 FNAC; 190 histology Up to 88.2% Threshold-dependent NR NR Up to 85.4% Bharti et al., 2026 India 111 FNAC; 42 histology 83.3% 100% 100% 81.0% 90.2% Chimatira & Price, 2026 South Africa Six-year tertiary laboratory cohort 73.4% 41.9% 46.6% 69.5% 54.8% NR = not reported in the summary data used for this synthesis. Diagnostic estimates are not directly interchangeable because studies used different definitions of cytological positivity and different strategies for handling indeterminate Bethesda categories. 4.3 Overall Diagnostic Accuracy Across the included contemporary studies, sensitivity ranged from 71.4% to 100%, specificity from 41.9% to 100%, and overall diagnostic accuracy from 54.8% to 97.7%. Most institutional studies reported accuracy above 80%, but studies with high nondiagnostic rates or strong surgical-selection effects demonstrated substantially poorer specificity and overall accuracy. The large 2022 systematic review and meta-analysis by Hsiao et al. included 16,597 patients from 36 studies and reported pooled sensitivity of 85.6% (95% CI 79.9-89.5), specificity of 71.4% (95% CI 61.1-79.8), positive likelihood ratio of 3.0, negative likelihood ratio of 0.2, and an area under the summary ROC curve of 86.1% [4]. The authors found that 7.2% of FNB results were inconclusive and that 13.4% of inconclusive specimens with an available reference standard were ultimately malignant. The broad range of specificity among contemporary studies reflects, in part, the use of different cytological thresholds. Treating Bethesda III-V as a positive result increases sensitivity but lowers specificity, whereas restricting positivity to Bethesda VI increases specificity while missing a proportion of malignant lesions. 4.4 Bethesda Categories and Risk of Malignancy The 2023 Bethesda System retains six diagnostic categories and updates estimated risks of malignancy [3]. Category-specific risk is intended to guide management rather than function as a simple binary cancer test. Table 3. Adult Risk of Malignancy in the 2023 Bethesda System Category 2023 Terminology Mean ROM Approximate Range I Nondiagnostic 13% 5-20% II Benign 4% 2-7% III Atypia of undetermined significance 22% 13-30% IV Follicular neoplasm 30% 23-34% V Suspicious for malignancy 74% 67-83% VI Malignant 97% 97-100% 4.5 False-Negative and False-Positive Cytology False-negative FNAC results were most often associated with sampling error, papillary thyroid microcarcinoma, cystic degeneration, multinodular glands, low cellularity and small malignant foci. In the Aliyev cohort, seven of ten false-negative malignant lesions were papillary microcarcinomas. False-positive diagnoses were associated with Hürthle-cell change, thyroiditis, adenomatous hyperplasia, follicular adenoma and regenerative atypia. These findings emphasize the need for radiological-pathological correlation and awareness of cytomorphological mimics. 4.6 Indeterminate Cytology and Follicular-Patterned Lesions Bethesda III and IV remain the principal diagnostic gray zones. FNAC cannot reliably distinguish follicular adenoma from follicular carcinoma because capsular and vascular invasion are histological criteria. Consequently, Bethesda IV is best regarded as a risk-stratification category. The 2023 Bethesda revision further subdivides atypia of undetermined significance into nuclear atypia and other atypia, reflecting evidence that nuclear atypia carries a higher malignancy risk [3]. 4.7 Nondiagnostic Cytology Nondiagnostic samples may result from poor cellularity, cyst-fluid-only aspirates, excessive blood, calcification, poor targeting or preparation artefact. The 2026 South African study reported an unusually high nondiagnostic proportion (50.1%) and correspondingly low overall accuracy, highlighting the importance of ultrasound guidance, operator training and rapid on-site adequacy assessment where feasible. 4.8 QUADAS-2 Risk-of-Bias Assessment Risk-of-bias concerns were dominated by patient selection and flow/timing. Histopathology is usually available only in patients who proceed to surgery; therefore, benign FNAC results are underrepresented among reference-standard-confirmed cases. This partial verification bias can distort estimates of sensitivity, specificity and risk of malignancy. The reference standard itself was generally judged low risk because surgical histopathology is the accepted definitive standard. Table 4. QUADAS-2 Risk-of-Bias Assessment of Included Studies Study Patient Selection Index Test Reference Standard Flow / Timing Overall Interpretation Nandedkar et al., 2018 High Low Low High High concern Anand et al., 2020 High Low Low High High concern Ucak et al., 2021 High Low Low Low Moderate concern Aliyev et al., 2022 High Low Low Low Moderate concern Sen et al., 2023 High Low Low High High concern Jain et al., 2023 High Low Low High High concern Osseis et al., 2023 High Low Low Low Moderate concern Sureka et al., 2025 High Low Low High High concern Bharti et al., 2026 High Low Low High High concern Chimatira & Price, 2026 Unclear Unclear Low High High concern Summary: patient-selection bias was judged high in 9 of 10 studies and unclear in 1; the index-test domain was low risk in 9 and unclear in 1; the reference standard was low risk in all 10; and flow/timing was high risk in 7 and low risk in 3. Overall, 7 studies were interpreted as high concern and 3 as moderate concern. These overall interpretations are review-level syntheses rather than a formal numerical QUADAS-2 score.

.1 Principal Findings This systematic review confirms that FNAC remains a clinically useful and generally accurate diagnostic method for thyroid nodules. The magnitude of reported sensitivity and specificity varies substantially, but the direction of evidence consistently supports FNAC as an effective first-line risk-stratification test. The pooled 2022 evidence provides the most stable overall estimate: sensitivity approximately 86%, specificity approximately 71%, and summary ROC performance approximately 86% [4]. Contemporary single-centre studies often report higher accuracy, but these estimates are particularly sensitive to case mix, surgical selection and the cytological threshold used to define a positive result. The strength of FNAC is greatest at the ends of the Bethesda spectrum. Category II has a low expected risk of malignancy, whereas Categories V and VI are strongly associated with malignancy. Categories III and IV remain the main diagnostic limitations because cytomorphology alone cannot resolve all borderline nuclear changes or demonstrate capsular/vascular invasion. 5.2 Histopathology as the Reference Standard Histopathology remains definitive after surgery because it permits assessment of tumour architecture, capsular invasion, vascular invasion, extrathyroidal extension, multifocality and histological subtype. Some apparent cytology-histology discordance therefore reflects the intrinsic limitations of cytology rather than interpretive error. 5.3 Verification Bias and Interpretation of Accuracy The most important methodological issue in this literature is selective surgical verification. Patients with benign cytology often do not undergo thyroidectomy, while those with suspicious imaging, indeterminate cytology or malignant FNAC are much more likely to have histopathology. This enriches surgical cohorts for malignancy and can inflate category-specific risk-of-malignancy estimates. It may also distort sensitivity and specificity because true-negative benign FNACs without surgery are omitted. 5.4 Role of Ultrasound Guidance and Adequacy Assessment Ultrasound guidance can improve targeting of small, cystic, calcified and nonpalpable lesions. Rapid on-site evaluation may reduce nondiagnostic sampling by allowing immediate adequacy assessment and repeat aspiration. These measures are especially important in centres with high Bethesda I rates. 5.5 Cytopathologist Experience and Standardized Reporting Interpretation remains challenging in Hürthle-cell lesions, follicular-patterned lesions, lymphocytic thyroiditis and specimens with subtle papillary nuclear change. Standardized Bethesda terminology improves communication and reproducibility. The 2023 system also places greater emphasis on molecular profiling for selected indeterminate nodules [3]. 5.6 Clinical Implications Optimal thyroid nodule evaluation should combine clinical assessment, high-quality ultrasonography, ultrasound-based risk stratification, appropriately targeted FNAC, Bethesda-standardized reporting, repeat sampling when required, and selective molecular testing for indeterminate lesions. Cytology should not be interpreted in isolation when clinical and radiological findings are strongly discordant. 6. Strengths • PRISMA 2020-structured reporting. • Focused inclusion of histopathology-correlated diagnostic studies. • Use of the updated 2023 Bethesda framework for interpretation. • Formal QUADAS-2 risk-of-bias assessment. • Explicit consideration of cytological thresholds and verification bias. • Contextualization with the largest modern diagnostic meta-analysis. 7. Limitations • The included studies were heterogeneous in design, case mix, Bethesda edition and definition of cytological positivity. • Most studies were retrospective and were subject to selective surgical verification. • Histopathology was unavailable for a large proportion of cytologically benign nodules in several cohorts. • Indeterminate Bethesda categories were handled differently across studies, limiting direct comparison of diagnostic indices. • A de novo meta-analysis was not performed because full 2x2 diagnostic data and consistent thresholds were not uniformly available across the selected contemporary studies. 8. Future Research • Prospective diagnostic-accuracy studies with prespecified cytological thresholds. • Reporting of complete 2x2 contingency tables for each relevant Bethesda threshold. • Longitudinal follow-up of non-operated Bethesda II nodules to reduce verification bias. • Direct comparison of 2017 and 2023 Bethesda classifications. • Further validation of AUS nuclear atypia versus other atypia. • Evaluation of rapid on-site adequacy assessment, molecular testing and artificial-intelligence-assisted cytology. • Integration of Bethesda categories with ultrasound risk-stratification systems.

Fine-needle aspiration cytology is a reliable, minimally invasive and clinically valuable first-line test for the evaluation of thyroid nodules. Across contemporary histopathology-correlated studies, diagnostic performance is generally high, although substantial variation exists because of differences in sampling, patient selection and analytical thresholds. The most robust pooled evidence indicates a sensitivity of 85.6%, specificity of 71.4%, and summary ROC performance of 86.1%. The principal diagnostic limitations are nondiagnostic sampling, indeterminate Bethesda categories, follicular-patterned lesions and sampling-related false negatives. Histopathology remains the definitive reference standard following surgery. Combining ultrasound-guided sampling, standardized 2023 Bethesda reporting, cytopathology expertise, repeat aspiration when appropriate and selective molecular testing offers the most effective strategy for maximizing diagnostic accuracy while minimizing unnecessary surgery.

1. Page MJ, McKenzie JE, Bossuyt PM, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ. 2021;372:n71. 2. Whiting PF, Rutjes AWS, Westwood ME, et al. QUADAS-2: a revised tool for the quality assessment of diagnostic accuracy studies. Ann Intern Med. 2011;155(8):529-536. 3. Ali SZ, Baloch ZW, Cochand-Priollet B, Schmitt FC, Vielh P, VanderLaan PA. The 2023 Bethesda System for Reporting Thyroid Cytopathology. Thyroid. 2023;33(9):1039-1044. 4. Hsiao V, Massoud E, Jensen C, et al. Diagnostic Accuracy of Fine-Needle Biopsy in the Detection of Thyroid Malignancy: A Systematic Review and Meta-analysis. JAMA Surg. 2022;157(12):1105-1113. 5. Nandedkar SS, Dixit M, Malukani K, Varma AV, Gambhir S. Evaluation of Thyroid Lesions by Fine-needle Aspiration Cytology According to Bethesda System and its Histopathological Correlation. Int J Appl Basic Med Res. 2018;8(2):76-82. 6. Anand B, Ramdas A, Ambroise MM, Kumar NP. The Bethesda System for Reporting Thyroid Cytopathology: A Cytohistological Study. J Thyroid Res. 2020;2020:8095378. 7. Ucak R, Eryilmaz OT, Ozguven BY, Uludag M, Kabukcuoglu F. Evaluation of Thyroid Fine-Needle Aspiration Biopsies according to Cytological Methods and Comparison with Histopathological Diagnosis. Sisli Etfal Hastan Tip Bul. 2021;55(1):93-100. 8. Aliyev A, Aliyeva I, Giammarile F, et al. Diagnostic accuracy of fine needle aspiration biopsy versus postoperative histopathology for diagnosing thyroid malignancy. Endocrinol Diabetes Metab. 2022;5(6):e373. 9. Sen A, Sarkar M, Das RN, Paul SS, Chatterjee U. The Bethesda system for reporting thyroid fine needle aspirates: A study of 1763 patients; with histopathological correlations, in eastern India. Diagn Cytopathol. 2023;51(3):174-181. 10. Jain D, Sharma R, Pachori G, Bayla T. Evaluation of Thyroid Lesions by Fine Needle Aspiration Cytology According to Bethesda System. Indian J Otolaryngol Head Neck Surg. 2023;75(2):457-462. 11. Osseis M, Jammal G, Kazan D, Noun R. Comparison between Fine Needle Aspiration Cytology with Histopathology in the Diagnosis of Thyroid Nodules. J Pers Med. 2023;13(8):1197. 12. Sureka N, Ahluwalia C, Ahuja S, Madan NK, Agrawal M, Ranga S. Assessment of Thyroid Fine-Needle Aspirates Using 2023 Bethesda System. Acta Cytol. 2025;69(3):280-290. 13. Bharti S, Singh R, Sharma M. Application of the Newer 2023 Bethesda System for Reporting Cytopathology of Thyroid Nodules and Estimating their Risk of Malignancy. J Cytol. 2026;43(2):74-83. 14. Chimatira R, Price B. Correlation of Fine Needle Aspiration Cytology and Histopathology in Thyroid Nodules: A Six-Year Study at a Tertiary Public Laboratory in the Western Cape Province, South Africa. Diagn Cytopathol. 2026;54(5):359-370. 15. Cibas ES, Ali SZ. The 2017 Bethesda System for Reporting Thyroid Cytopathology. Thyroid. 2017;27(11):1341-1346. 16. Bongiovanni M, Spitale A, Faquin WC, Mazzucchelli L, Baloch ZW. The Bethesda System for Reporting Thyroid Cytopathology: a meta-analysis. Acta Cytol. 2012;56(4):333-339. 17. 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. 18. Zhu Y, Song Y, Xu G, et al. Causes of misdiagnoses by thyroid fine-needle aspiration cytology: our experience and a systematic review. Diagn Pathol. 2020;15:1. 19. Juhlin CC, Baloch ZW. The 3rd edition of the Bethesda System for Reporting Thyroid Cytopathology: highlights and comments. Endocr Pathol. 2024;35(1):77-79.