Why Thyroid Nodule Evaluation Matters
Thyroid nodules are found incidentally in up to 68% of adults on neck ultrasound. The great majority are benign — only 7–15% of nodules that undergo FNA prove malignant. The clinical challenge is identifying the minority that require surgery while sparing the majority unnecessary procedures. A structured approach using sonographic risk stratification, cytopathology, and, when needed, molecular genomic testing accomplishes this reliably.
Step 1 — Sonographic Risk Stratification (ATA 2015)
The 2015 ATA Management Guidelines classify nodules by their ultrasound appearance into five risk tiers. Each tier carries a different estimated malignancy risk and FNA size threshold.
| ATA Category | Sonographic Features | Malignancy Risk | FNA Threshold |
|---|---|---|---|
| Benign | Purely cystic nodule | <1% | No FNA recommended |
| Very Low Suspicion | Spongiform or partially cystic without any suspicious features | <3% | ≥2 cm (observation without FNA also reasonable) |
| Low Suspicion | Isoechoic or hyperechoic solid; or partially cystic with eccentric solid areas — no microcalcifications, irregular margins, or ETE | 5–10% | ≥1.5 cm |
| Intermediate Suspicion | Hypoechoic solid with smooth margins — no microcalcifications, irregular margins, taller-than-wide, or ETE | 10–20% | ≥1 cm |
| High Suspicion | Solid hypoechoic nodule or hypoechoic component of partially cystic nodule plus ≥1 of: irregular/spiculated margins, microcalcifications, taller-than-wide shape, rim calcification with extrusion, ETE | 70–90% | ≥1 cm |
Step 2 — Fine Needle Aspiration (FNA) & Bethesda Reporting
When a nodule meets the FNA size threshold, ultrasound-guided FNA is performed. The aspirate is evaluated by a cytopathologist and reported using the Bethesda System for Reporting Thyroid Cytopathology (BSRTC). The 2023 third edition updated malignancy risk estimates to reflect NIFTP reclassification and expanded molecular testing data.
The Six Bethesda Categories
| Category | Diagnosis | Malignancy Risk | Typical Management |
|---|---|---|---|
| I — Nondiagnostic | Insufficient cellularity, obscuring blood, or cyst fluid only | 5–10% | Repeat US-guided FNA; if repeatedly nondiagnostic consider core needle biopsy or surgical excision for growing/suspicious nodule |
| II — Benign | Benign follicular nodule, colloid nodule, Hashimoto thyroiditis, granulomatous (de Quervain) thyroiditis, simple cyst | 0–3% | Clinical and sonographic follow-up; repeat ultrasound at 12–24 months for low-risk nodules; no surgery indicated |
| III — AUS / FLUS | Atypia of Undetermined Significance or Follicular Lesion of Undetermined Significance — nuclear atypia beyond benign but not meeting FN or malignant criteria | 13–30% | Molecular testing (AFIRMA or Thyroseq); repeat FNA at 3–6 months; or diagnostic lobectomy. Management individualized. |
| IV — Follicular Neoplasm | Follicular Neoplasm or Suspicious for Follicular Neoplasm (SFN); includes Hürthle cell (oncocytic) variant | 25–45% | Molecular testing strongly recommended; if molecular positive or suspicious → diagnostic lobectomy (may proceed to total thyroidectomy based on findings) |
| V — Suspicious for Malignancy | Suspicious for PTC, MTC, poorly differentiated carcinoma, lymphoma, or metastatic carcinoma | 50–75% | Near-total thyroidectomy or diagnostic lobectomy; molecular testing may inform extent of surgery (avoid completion thyroidectomy if avoidable) |
| VI — Malignant | Papillary thyroid carcinoma (PTC), medullary thyroid carcinoma (MTC), anaplastic carcinoma, primary thyroid lymphoma, metastatic carcinoma to thyroid | 97–99% | Appropriate definitive surgery (usually near-total thyroidectomy ± central or lateral neck dissection); RAI consideration for PTC; calcitonin/CEA for MTC |
Step 3 — Molecular Testing for Indeterminate Nodules (Bethesda III & IV)
Approximately 20–30% of FNAs return as Bethesda III or IV — "indeterminate." Historically, the standard approach was diagnostic lobectomy for all indeterminate nodules, resulting in surgery for up to 75% of patients whose resected specimens proved benign. Molecular genomic testing has transformed this: a benign/negative molecular result identifies patients who can safely avoid surgery with ~97% confidence.
AFIRMA Gene Sequence Classifier (GSC)
The AFIRMA GSC (Veracyte) analyzes whole transcriptome RNA expression from FNA material. It classifies nodules as Benign or Suspicious. Approximately 68% of tested indeterminate nodules receive a Benign call, achieving a negative predictive value of ~97% — meaning fewer than 1 in 20 Benign-called nodules proves malignant at surgery. AFIRMA also offers the Xpression Atlas (AXA) add-on for Suspicious results, identifying specific driver mutations and gene fusions (BRAF, RAS, RET/PTC, NTRK, etc.) to guide surgical planning.
Thyroseq v3 Genomic Classifier
Thyroseq v3 (CBLPath / University of Pittsburgh Medical Center) uses next-generation sequencing of both DNA and RNA to detect point mutations, copy number alterations, and gene fusions across 112 genes. A Negative result reduces the malignancy risk to approximately 3–5% (NPV ~97%). A Positive result identifies the specific alteration, which has important surgical implications: BRAF V600E predicts near-certain PTC and typically guides total thyroidectomy; RAS mutations carry a 20–40% malignancy risk and may still allow lobectomy; high-risk fusions (RET, NTRK) support more aggressive surgery.
| Parameter | AFIRMA GSC | Thyroseq v3 |
|---|---|---|
| Manufacturer | Veracyte (South San Francisco, CA) | CBLPath / UPMC (Rye Brook, NY) |
| Technology | RNA whole transcriptome microarray | DNA + RNA next-gen sequencing (112 genes) |
| Best Bethesda categories | III and IV | III, IV, and V |
| Sensitivity | ~91% | ~94% |
| Specificity | ~68% | ~82% |
| NPV (indeterminate nodule) | ~97% | ~97% |
| PPV (indeterminate nodule) | ~38% | ~66% |
| Benign/Negative call rate | ~68% of tested nodules | ~65% of tested nodules |
| Guides surgery extent? | Limited (AXA add-on) | Yes — alteration type informs lobectomy vs. TT |
| Limitations | Lower specificity; Hürthle cell neoplasms less reliable | Some alterations (RAS) still require surgery; Hürthle cell variants also challenging |
| Insurance | Medicare + most commercial | Medicare + most commercial |
| Turnaround | ~10 business days | ~10 business days |
Which Test to Order?
Both AFIRMA and Thyroseq perform comparably for ruling out malignancy, and the choice is often driven by physician preference, laboratory relationships, and insurance. Thyroseq's mutation-specific reports can provide additional surgical planning detail when positive. AFIRMA's Xpression Atlas (AXA) offers similar driver information on Suspicious calls. For Hürthle cell (oncocytic) variants of Bethesda IV, neither platform performs as well and surgical excision remains the standard in most guidelines.
In-Office FNA at Endocrine & Diabetes Plus
We perform ultrasound-guided FNA in the office. All biopsies are sent to cytopathology. For indeterminate results (Bethesda III or IV), we coordinate AFIRMA or Thyroseq collection from the FNA specimen and discuss results with patients in a dedicated follow-up visit. Our team is experienced in collecting both proprietary FNA preservation media required for molecular testing and in reviewing molecular reports in clinical context.
After the Biopsy: What Happens Next?
Results typically return within 7–10 business days. Our providers review every result and contact patients to discuss findings. Most benign results (Bethesda II) require no further workup beyond scheduled ultrasound surveillance. Indeterminate results are discussed in the context of clinical risk — nodule size, growth rate, patient preference — before recommending molecular testing or surgery. For malignant results, we coordinate surgical referral with experienced thyroid surgeons and facilitate staging workup when indicated.