Residency · Residency · Pathology
Thyroid Cytopathology and the Bethesda System
Overview
The Bethesda System for Reporting Thyroid Cytopathology (TBSRTC) is a standardized six-tier classification system that provides a uniform framework for communicating thyroid FNA results. Each category carries an implied risk of malignancy (ROM) and linked management recommendations, enabling clinicians and patients to make informed decisions about surveillance, molecular testing, or surgery. The third edition (2023) incorporates updates reflecting the reclassification of NIFTP and the growing role of molecular testing in clinical decision-making.
The Bethesda System: Six Diagnostic Categories
| Category | Diagnosis | Risk of Malignancy (3rd ed.) | Management |
|---|---|---|---|
| I | Nondiagnostic / Unsatisfactory | 5–10% | Repeat FNA |
| II | Benign | 0–3% | Clinical/US follow-up |
| III | AUS / FLUS | 10–30% | Repeat FNA, molecular testing, or lobectomy |
| IV | Follicular Neoplasm / SFN | 25–40% | Molecular testing or lobectomy |
| V | Suspicious for Malignancy | 50–75% | Lobectomy or total thyroidectomy |
| VI | Malignant | 97–99% | Surgery (extent based on type/stage) |
Category I: Nondiagnostic/Unsatisfactory
A specimen is considered nondiagnostic when it contains fewer than 6 groups of well-preserved follicular cells, with at least 10 cells per group, on at least 2 slides. An important exception exists: any specimen demonstrating atypia is considered adequate regardless of overall cellularity, as the atypical cells themselves are diagnostic material. Cyst fluid-only specimens are classified as nondiagnostic by default unless atypical cells are identified. The risk of malignancy for this category is 5-10% in the 3rd edition. Management consists of repeat FNA with ultrasound guidance in 1-3 months; persistently nondiagnostic aspirates from solid nodules should prompt consideration of surgical excision. Common causes of inadequacy include obscuring blood, poor smearing technique, scant cellularity, and air-drying artifact.
Category II: Benign
The benign category encompasses several entities. Benign follicular nodules (adenomatoid or colloid nodules) show sheets of uniform follicular cells with abundant background colloid. Lymphocytic (Hashimoto) thyroiditis demonstrates Hurthle cell clusters, lymphoid cells, and giant cells. Granulomatous (De Quervain) thyroiditis shows granulomas with multinucleated giant cells and a mixed inflammatory background. The risk of malignancy is 0-3%, and management is clinical and ultrasound follow-up without surgery. This category accounts for approximately 60-70% of all thyroid FNA diagnoses.
Category III: Atypia of Undetermined Significance / Follicular Lesion of Undetermined Significance (AUS/FLUS)
This category applies when cytologic or architectural atypia exceeds what is expected for a benign lesion but is insufficient for classification as Bethesda IV, V, or VI. Subtypes include nuclear atypia, architectural atypia, oncocytic atypia, and other patterns. As a quality metric, this category should not exceed 10% of all thyroid FNA diagnoses. The risk of malignancy ranges from 10-30%, with the nuclear atypia subtype carrying higher risk. Management options include repeat FNA, molecular testing, or diagnostic lobectomy. Nuclear atypia raises concern for papillary carcinoma, while architectural atypia overlaps with follicular neoplasm.
Category IV: Follicular Neoplasm / Suspicious for a Follicular Neoplasm (FN/SFN)
This category is applied to cellular specimens showing microfollicular or trabecular architecture with absent or scant colloid and predominantly follicular cells in repetitive microfollicular arrangements. It includes the Hurthle cell (oncocytic) neoplasm subtype. A fundamental limitation of cytology at this stage is that it cannot distinguish follicular adenoma from follicular carcinoma, because that distinction requires histologic assessment of capsular and vascular invasion on the excised specimen. The risk of malignancy is 25-40%. Management involves molecular testing or diagnostic lobectomy, with total thyroidectomy considered if molecular results suggest high-risk features.
Category V: Suspicious for Malignancy
This category is used when cytologic features are highly suggestive of malignancy but fall short of the definitive criteria required for Category VI. Most commonly, the specimen is suspicious for papillary carcinoma, showing nuclear enlargement, grooves, and rare pseudoinclusions that are quantitatively insufficient for a definitive diagnosis. This category also includes specimens suspicious for medullary carcinoma, lymphoma, or metastatic tumor. The risk of malignancy is 50-75%. Management is near-total thyroidectomy or lobectomy, and molecular testing may be performed to further inform surgical extent.
Category VI: Malignant
This category is reserved for specimens with cytologic features conclusive for malignancy. Papillary thyroid carcinoma is the most common diagnosis, recognized by its characteristic nuclear features: enlarged overlapping nuclei, powdery (ground-glass) chromatin, nuclear grooves, intranuclear pseudoinclusions, and psammoma bodies. Medullary carcinoma shows plasmacytoid or spindle cells with eccentric nuclei, salt-and-pepper chromatin, and background amyloid, confirmed by calcitonin immunohistochemistry. Anaplastic carcinoma demonstrates pleomorphic or bizarre cells with necrosis. Lymphoma presents as a monomorphic lymphoid population requiring flow cytometry confirmation. The risk of malignancy is 97-99%, and management is surgical, with extent determined by tumor type and staging.
Molecular Testing in Thyroid Cytopathology
Indications
Molecular testing is primarily applied to Bethesda III and IV nodules to improve risk stratification. The goal is twofold: to reduce unnecessary diagnostic lobectomies for ultimately benign nodules (the rule-out application) and to identify malignancy preoperatively so that appropriate initial surgery can be performed (the rule-in application).
Major Molecular Platforms
The Afirma Genomic Sequencing Classifier (GSC) is a gene expression classifier that functions primarily as a rule-out test. It classifies nodules as benign or suspicious. A benign call carries a high negative predictive value of 96%, allowing patients to avoid surgery, while suspicious nodules proceed to surgical excision. The platform includes ancillary panels for BRAF, RAS, RET/PTC, PAX8-PPARG, and a medullary thyroid carcinoma classifier.
ThyroSeq v3 is a targeted next-generation sequencing panel that analyzes both DNA and RNA, functioning as both a rule-in and rule-out test. It tests 112 genes for mutations, gene fusions, copy number alterations, and gene expression changes. A negative result carries a 97% negative predictive value. Positive results are stratified by mutation type: BRAF V600E has near 100% positive predictive value for papillary carcinoma, while RAS mutations carry a lower PPV of approximately 50% as they may be found in either adenomas or carcinomas.
ThyGeNEXT/ThyraMIR combines a mutation panel with a microRNA classifier for complementary risk assessment.
Key Molecular Alterations
BRAF V600E is highly specific for papillary thyroid carcinoma with greater than 99% positive predictive value. It is associated with aggressive histologic features and loss of radioiodine avidity. RAS mutations (NRAS, HRAS, KRAS) are found in follicular adenoma, follicular carcinoma, follicular variant of papillary carcinoma, and NIFTP, giving them lower specificity for malignancy. RET/PTC rearrangements are found in papillary carcinoma, with RET/PTC1 being most common and RET/PTC3 associated with tall cell variant and radiation exposure. PAX8-PPARG fusion occurs in both follicular carcinoma and follicular adenoma. RET point mutations are characteristic of medullary thyroid carcinoma and require germline testing for MEN2 evaluation. TERT promoter mutations indicate aggressive behavior in differentiated thyroid carcinoma and are associated with distant metastases and increased mortality. TP53 mutations characterize anaplastic thyroid carcinoma.
Impact of NIFTP
Definition
Non-invasive follicular thyroid neoplasm with papillary-like nuclear features (NIFTP) was reclassified from what was previously called non-invasive encapsulated follicular variant of papillary thyroid carcinoma. Its strict diagnostic criteria include encapsulation or well-demarcated borders, follicular growth pattern, papillary carcinoma-type nuclear features, absence of invasion, absence of true papillae, less than 30% solid or trabecular growth, no tumor necrosis, and low mitotic activity.
Effect on Bethesda System
Because NIFTP is now classified as a neoplasm rather than a carcinoma, its reclassification has decreased the risk of malignancy across all Bethesda categories. The 3rd edition provides ROM figures both with and without NIFTP counted as malignancy, allowing laboratories and clinicians to use whichever framework is most appropriate for their management algorithms.
Quality Metrics
Quality assurance in thyroid cytopathology requires monitoring several metrics. The nondiagnostic rate should be below 10% (excluding cyst fluid-only specimens). The AUS/FLUS (Category III) rate should not exceed 10% of all thyroid FNA diagnoses, serving as a check against overuse of this indeterminate category. The malignancy rate within each category should approximate published ROM ranges. Cytohistologic correlation -- comparing the cytologic diagnosis with the final surgical pathology diagnosis -- is the gold standard quality assurance measure.
<image>A medical illustration showing representative cytologic appearances for each of the six Bethesda categories of thyroid FNA. Panel I (Nondiagnostic): A sparsely cellular smear with a few scattered follicular cells against a background of red blood cells, falling below the six-group threshold. Panel II (Benign): Flat honeycomb sheets of uniform follicular cells with abundant thin colloid in the background (colloid nodule). Panel III (AUS/FLUS): Follicular cells showing focal nuclear enlargement and pallor suggestive of but insufficient for papillary carcinoma. Panel IV (Follicular Neoplasm): Highly cellular smear composed almost exclusively of microfollicular arrangements with scant colloid. Panel V (Suspicious for Malignancy): Clusters showing nuclear enlargement, grooves, and chromatin pallor but lacking definitive pseudoinclusions. Panel VI (Malignant - Papillary Carcinoma): Papillary tissue fragments with enlarged oval nuclei, powdery chromatin, prominent nuclear grooves, intranuclear pseudoinclusions (arrow), and a psammoma body.</image>
<image>A decision algorithm flowchart for management of thyroid nodules using the Bethesda System with molecular testing integration. Starting from thyroid nodule identification, showing ultrasound risk stratification (TI-RADS), FNA indication criteria, the six Bethesda categories with their respective ROM percentages, and branching pathways: Category I leads to repeat FNA; Category II to surveillance; Categories III and IV to molecular testing (with Afirma GSC benign result returning to surveillance and suspicious result proceeding to surgery, or ThyroSeq negative returning to surveillance and positive proceeding to surgery); Category V to surgery with possible preoperative molecular testing; Category VI to surgery with extent determined by staging.</image>
Clinical Pearls
BRAF V600E on molecular testing is essentially diagnostic of papillary thyroid carcinoma and may allow total thyroidectomy as the initial surgery, avoiding a two-stage approach of lobectomy followed by completion thyroidectomy. RAS mutations are found in both benign and malignant thyroid lesions, so a RAS-positive result should not automatically trigger total thyroidectomy. The AUS/FLUS category functions as a "wastebasket" diagnosis, and tracking institutional rates and subtype distribution is essential for quality assurance. In the setting of Hashimoto thyroiditis, Hurthle cells with reactive nuclear atypia can mimic papillary carcinoma on FNA, making correlation with clinical and ultrasound features critical. Cyst fluid-only specimens classified as Category I should prompt re-aspiration targeting the solid component under ultrasound guidance. Medullary thyroid carcinoma may be misclassified as follicular neoplasm on cytology due to its plasmacytoid appearance; serum calcitonin and calcitonin immunohistochemistry on cell block should be considered when plasmacytoid cells or amyloid are present. TERT promoter mutations combined with BRAF V600E identify the highest-risk papillary thyroid carcinomas with significant mortality risk.
References
- Ali SZ, Baloch ZW, et al. The Bethesda System for Reporting Thyroid Cytopathology. 3rd ed. Springer; 2023.
- Cibas ES, Ali SZ. The 2017 Bethesda System for Reporting Thyroid Cytopathology. Thyroid. 2017;27(11):1341-1346.
- Steward DL, et al. Performance of a multigene genomic classifier in thyroid nodules with indeterminate cytology: A prospective blinded multicenter study. JAMA Oncol. 2019;5(2):204-212.
- Nikiforov YE, et al. Impact of the multi-gene ThyroSeq next-generation sequencing assay on cancer diagnosis in thyroid nodules with atypia of undetermined significance/follicular lesion of undetermined significance cytology. Thyroid. 2015;25(11):1217-1223.
- Nikiforov YE, et al. Nomenclature Revision for Encapsulated Follicular Variant of Papillary Thyroid Carcinoma: A Paradigm Shift to Reduce Overtreatment of Indolent Tumors. JAMA Oncol. 2016;2(8):1023-1029.
