Molecular Tests for Thyroid Nodules and Thyroid Cancer Diagnosis

• What each test is built to do:
  • Afirma GSC (± Xpression Atlas): 
    • RNA whole-transcriptome classifier optimized as a rule-out test:
      • XA adds variants / fusions if GSC is Suspicious PMCVeracyte
  • ThyroSeq v3 GC: 
    • DNA / RNA next-gen panel across many genes (mutations, fusions, copy-number alterations, helpful in oncocytic cell lesions / Hürthle cell lesions) designed for rule-out + rule-in and therapy-relevant profiling JAMA NetworkThyroSeq
  • ThyGeNEXT + ThyraMIR v2 (Interpace): 
    • Two-step oncogene panel → reflex miRNA classifier; aims for balanced rule-out with stronger rule-in when driver / miRNA high-risk pattern is present PMCThygenext Thyramir
• Performance (typical ranges in Bethesda III–IV):
  • Afirma GSC: 
    • High sensitivity / NPV in VS and real-world cohorts:
      • PPV modest to moderate
    • Many studies report:
      • NPV ~ 94% to 97%
      • PPV ~ 45% to 65%
    • Real-world series show improved yield vs validation PMCScienceDirect
  • ThyroSeq v3: 
    • Validation and multi-center series show high sensitivity / NPV with higher PPV than Afirma in several cohorts:
      • PPV ~ 60% to 65%
    • Institutional data show sensitivity ~ 92% to 95% (Bethesda III to IV) JAMA NetworkACS Publications
  • ThyGeNEXT/ThyraMIR v2: 
    • Reported NPV ~ 93% to 96%
    • Useful PPV when drivers / miRNA risk present
    • 2025 systematic review suggests high surgical avoidance rates among platforms PMCResearchGate
• Prognostic/management information (actionability):
  • Afirma GSC + XA: 
    • If Suspicious, XA reports fusions / variants (e.g., RET, NTRK, ALK):
      • Which can guide targeted therapy decisions down the line
    • GSC itself primarily aids avoiding surgery Veracyte
  • ThyroSeq v3: 
    • Reports BRAF, RAS, TERT-p, RET / NTRK/ ALK, gene expression and copy-number profiles:
      • Useful for risk stratification (e.g., TERT / BRAF V600E for aggressiveness) and for operative planning (extent, LN assessment) and potential targeted options
      • Particularly helpful in Hürthle-predominant nodules JAMA NetworkThyroSeq
  • ThyGeNEXT/ThyraMIR v2: 
    • Calls out high-risk drivers (BRAF V600E, TERT, ALK, etc.) and refines intermediate results with miRNA pairs:
      • Useful for deciding between surveillance vs diagnostic lobectomy when imaging / clinical risk is equivocal PMCInterpace Biosciences®, Inc.
• Independent comparisons and guidelines:
  • Contemporary reviews / meta-analyses: 
    • Afirma GSC and ThyroSeq v3 both excel as rule-out tests (high sensitivity / NPV), with ThyroSeq often showing higher PPV (rule-in)
    • Interpace’s combined platform performs comparably on NPV with strong rule-in behavior in some studies ScienceDirect+1PMC
• Societal guidance (ETA 2023):
  • Molecular tests can reduce diagnostic surgery and should be selected to match pretest risk and clinical goals rather than “one best test for all.” PMC
• Bottom line (how to choose today):
  • If your primary goal is to avoid surgery (rule-out) in Bethesda III to IV with low–intermediate pretest risk:
    • Afirma GSC or ThyroSeq v3 are both appropriate:
      • Real-world data support high NPV for both PMCJAMA Network
  • If you want both strong rule-out and richer “what kind of cancer is this if positive?” detail to guide extent of surgery and future therapy:
    • ThyroSeq v3 generally provides more granular prognostic / actionable data (TERT/driver profile, CNA burden, Hürthle copy-number signature) and often a higher PPV than Afirma
    • This makes it my usual pick when operative planning may hinge on genotype JAMA NetworkThyroSeq
  • If you favor a stepwise, cost-conscious approach with meaningful rule-in capability when drivers /miRNA are high-risk:
    • ThyGeNEXT + ThyraMIR v2 is reasonable:
      • Pooled data show competitive NPV and good surgical-avoidance rates PMCResearchGate
• Practical take for your clinic:
  • For Bethesda III to IV nodules with indeterminate US where you want to minimize unnecessary operations and inform extent if positive: 
    • ThyroSeq v3 is the most versatile single test today
  • For surgeon’s “rule-out first” workflows with straightforward nodules and low pretest risk: 
    • Afirma GSC is perfectly acceptable:
      • Add XA if suspicious and you want therapy targets.
  • For equivocal cases where a rule-in signal would materially change from surveillance to surgery, or when prior testing is ambiguous: 
    • ThyGeNEXT + ThyraMIR v2 is a solid option

Interpretation & Take-Home

• Best Rule-Out (Highest NPV and Sensitivity):
  • Afirma GSC (real-world meta-analysis: SN ~97%, NPV ~99%) and ThyroSeq v3 (UCLA trial: SN 97%, NPV 99%) are essentially neck-and-neck in ruling out malignancy. Both demonstrate excellent reliability in avoiding unnecessary surgeries.
• Better Rule-In (Higher PPV):
  • ThyroSeq v3 edges ahead slightly with PPV ~64% versus Afirma’s ~57% in comparable settings, meaning a positive result is more likely to indicate true malignancy.
• Manufacturer-Reported MPTX (ThyGeNEXT + ThyraMIR):
  • Claims very strong performance (SN ~ 95%, SP ~ 90%, NPV ~ 97%, PPV ~ 75%):
    • But these results may reflect idealized cohorts with prevalence adjustments
    • Real-world studies show more conservative metrics:
      • SN ~ 76%, SP ~ 75%, NPV ~ 83%, PPV ~ 67%).
• In Summary:
  • For maximum confidence in rule-out:
    • Afirma GSC and ThyroSeq v3 are clearly superior
  • If positive actionable findings (e.g., higher PPV, molecular prognostic detail) are crucial:
    • ThyroSeq v3 offers an advantage
  • MPTX is promising, especially if you value a modular approach, but real-world validation remains less robust than for the other two