Is surgery indicated for a thyroid nodule with atypical follicular cells of undetermined significance (Bethesda category III)?

Management of Bethesda Category III Thyroid Nodules

Surgery is NOT automatically indicated for Bethesda category III (AUS/FLUS) thyroid nodules; management should be guided by ultrasound risk stratification and consideration of molecular testing, with active surveillance being an appropriate option for low-risk nodules.

Risk-Stratified Approach

The most recent NCCN guidelines 1 explicitly restructured their algorithm to deemphasize immediate surgery for AUS/FLUS nodules and allow for active surveillance when molecular diagnostics are not performed. This represents a significant shift away from reflexive surgical intervention.

Key Decision Points:

1. Ultrasound Risk Assessment (Primary Determinant)

The ultrasound appearance using standardized risk stratification systems (K-TIRADS) is the most powerful predictor of malignancy in Bethesda III nodules:

• High suspicion features (K-TIRADS 5): Malignancy rate approaches 88-100% 2, 3

  • These nodules warrant surgical consideration
  • Nearly all are papillary thyroid carcinomas 2

• Intermediate suspicion features: Malignancy rate ~83-88% 2

  • Consider molecular testing or surgery based on clinical context

• Low suspicion features: Malignancy rate ~50-60% 2

  • Active surveillance is appropriate rather than immediate surgery 1, 2
  • If malignant, more likely to be follicular carcinoma (which K-TIRADS predicts poorly) 2

2. Molecular Testing (Secondary Tool)

The NCCN panel 1 softened recommendations for molecular testing (changing from mandatory to "consider"), reflecting uncertainty about its clinical utility. However, when used:

• BRAF V600E mutation: Strong predictor of malignancy (odds ratio 4.54) 3
• Predicted malignancy risk ≤5% (comparable to benign nodules) supports active surveillance 1
• Important caveat: Molecular diagnostics perform poorly for Hürthle cell neoplasms 1

3. Repeat FNA: Limited Value

Recent evidence challenges the utility of repeat FNA:

• 38.5% remain AUS/FLUS on repeat 4
• Among nodules with two consecutive AUS/FLUS diagnoses, 26.3% are malignant 4
• Repeat FNA did not appear in decision tree models for reducing unnecessary surgery 3
• The conclusive result rate increases with higher K-TIRADS categories (58.3% for low suspicion vs 88.8% for high suspicion) 2

Critical Context: Actual Malignancy Rates

The traditional 5-15% malignancy risk for Bethesda III is outdated. Multiple recent studies demonstrate significantly higher rates:

• Overall malignancy rate: 26.6-37.8% in surgical series 4
• Varies by institution and patient population 5, 6
• AUS subcategory: 94.2% malignancy rate 2
• FLUS subcategory: 40% malignancy rate 2

The 2023 Bethesda System 7 has updated risk estimates, though specific percentages vary by clinical setting.

Recommended Algorithm

1. Classify ultrasound risk using K-TIRADS or equivalent system

2. For K-TIRADS low suspicion nodules:

   • Consider active surveillance 1, 2
   • Molecular testing optional to further refine risk
   • Avoid immediate surgery unless concerning clinical features present

3. For K-TIRADS intermediate/high suspicion nodules:

   • Consider molecular testing if available
   • If molecular testing suggests high risk OR unavailable → proceed to surgery
   • If molecular testing suggests low risk → active surveillance acceptable

4. Additional high-risk clinical features favoring surgery:

   • Younger patient age 4, 6
   • Family history of thyroid cancer
   • History of head/neck irradiation
   • Concerning ultrasound features: hypoechogenicity, microcalcifications, irregular margins 8, 9

Important Caveats

• Institutional variation matters: Malignancy rates differ significantly between community and tertiary cancer centers 4
• AUS vs FLUS distinction: AUS carries substantially higher malignancy risk than FLUS 2
• Follicular carcinoma limitation: Ultrasound and molecular testing predict papillary carcinoma well but perform poorly for follicular carcinoma 2
• Incidental cancers: Even in patients not undergoing surgery for the index nodule, incidental cancers are found in 22.3% 4

The ESMO guidelines 8 emphasize that many thyroid cancers (particularly NIFTP and small papillary carcinomas) have excellent prognosis with <1% recurrence risk, supporting conservative management when appropriate. The shift toward risk-tailored strategies reflects growing recognition of thyroid cancer overdiagnosis and overtreatment 8.