Diagnostic and Molecular Yield of a Multidisciplinary Team-Guided Combined Endobronchial and Endoscopic Ultrasound-Guided Sampling Strategy for Mediastinal and Hilar Lymphadenopathy

🎯 Background & Rationale

While endobronchial ultrasound (EBUS) is established as a first-line modality for lung cancer staging, its yield is lower in benign conditions (such as sarcoidosis and tuberculosis) and hematologic malignancies. Furthermore, evolving treatment paradigms demand increasingly complex molecular profiling, and traditional EBUS is anatomically limited in reaching certain lymph node stations (e.g., stations 5, 6, 8, and 9). Although adding endoscopic ultrasound (EUS or EUS-B) expands nodal access and tissue acquisition, real-world data evaluating a structured, multidisciplinary team (MDT)-guided strategy using combined endosonographic modalities for mixed cohorts (both malignant and benign etiologies) remain sparse.

👥 Study Design & Population

• Study Type: Retrospective Cohort Study.
• Population (P): 213 consecutive adult patients (mean age 56±14.5 years; 64.3% male) presenting with mediastinal and hilar lymphadenopathy.
• Intervention (I): A protocolized, multidisciplinary team (MDT)-guided endosonographic sampling strategy incorporating EBUS, EUS-B, and EUS, tailored to anatomical accessibility and anticipated molecular needs.
• Comparator ©: Modality selection was individualized rather than randomized; subgroup comparisons were performed based on clinical indication (suspected carcinoma vs. suspected benign/lymphoproliferative disease) and modality used.
• Outcomes (O): Primary outcome was overall diagnostic yield (defined as a definitive pathologic diagnosis consistent with clinical/radiologic context, with benign diagnoses confirmed by ≥3 months of clinical/radiologic follow-up). Secondary outcomes included yield stratified by indication and modality, molecular adequacy, and procedural safety.

📈 Methodology & Rigor

• Statistical Analysis: Continuous variables were reported as means and standard deviations, and categorical variables as counts and percentages. Diagnostic yields were presented with 95% Wilson binomial confidence intervals. Differences across modalities were evaluated using Fisher’s exact or Fisher-Freeman-Halton exact tests. Continuous variables were compared using one-way ANOVA and Mann-Whitney U tests.
• Procedural Rigor: All cases were pre-reviewed by a dedicated thoracic MDT. Standardized staging sequences (contralateral N3 to N1) were utilized. Rapid on-site evaluation (ROSE) was available in 67.6% of cases, and needle selection and pass numbers were adjusted based on clinical suspicion and tumor burden.

🔬 Key Findings [or Planned Endpoints]

• Overall Diagnostic Yield: 92.0% (n=196/213; 95% CI: 87.6%–95.0%).
• Diagnostic Yield by Indication: 91.8% (n=123/134) in patients with suspected carcinoma and 92.4% (n=73/79) in patients with suspected benign or lymphoproliferative disease.
• Yield by Modality: EBUS alone achieved 90.4% (n=142/157), EUS-B achieved 87.5% (n=7/8), EUS alone achieved 100% (n=26/26), and combined procedures achieved 95.4% (n=21/22). Differences across modalities were not statistically significant (P=0.29).
• Molecular Adequacy: Among the 57 malignant cases sent for molecular testing, successful completion was achieved in 89.5% (n=51). PD-L1 testing was adequate in 88% of designated samples.
• Safety: Complications included minor bleeding in 2 patients (0.9%) and mediastinitis in 1 patient (0.5%).
• Node Size: Nondiagnostic nodes had a significantly smaller mean CT short-axis diameter compared to diagnostic nodes (12.7 mm vs. 15.4 mm; P=0.024).

⚖️ Critical Appraisal

• Internal Validity & Bias: The retrospective, non-randomized design introduces selection bias, as modalities (particularly EUS) were preferentially assigned based on clinical suspicion and anatomical suitability. Furthermore, EUS procedures were performed by a single trained operator, potentially limiting internal consistency across multiple endosonographers.
• Statistical Limitations: Small subgroup sizes for EUS-B (n=8) and combined procedures (n=22), as well as a modest number of confirmed lymphoma cases (n=9), limit the statistical power for comparative subgroup analyses.
• Generalizability: The findings reflect practice at a single tertiary referral center with specialized MDT infrastructure and advanced interventional pulmonology expertise, which may restrict direct extrapolation to centers lacking dedicated EUS capabilities or structured multidisciplinary lung boards.

💡 The Clinical Bottom Line

An MDT-driven, individualized endosonographic strategy (incorporating EBUS, EUS-B, and EUS) is safe and highly effective, achieving high diagnostic yields and molecular adequacy across both malignant and benign mediastinal/hilar lymphadenopathy. Clinicians in the bronchoscopy suite should consider a flexible, tailored modality selection—utilizing EUS when architectural preservation or tissue for complex molecular profiling is prioritized—rather than relying strictly on a uniform EBUS-first approach.