Routine Systematic Sampling vs. Selective Targeted Sampling of lymph nodes during mediastinal staging: A feasibility randomized controlled trial

Abstract

Background: The standard of care for mediastinal staging during endobronchial ultrasound (EBUS) is Systematic Sampling (SS) where a minimum of 3 lymph node (LN) stations are biopsied, even if they appear normal on imaging. When LNs appear normal on PET and CT, the Canada Lymph Node Score can also identify if they appear normal on EBUS. For these Triple Normal LNs, the pretest probability of malignancy is < 6%, and routine biopsy may not be required. This preliminary study introduced Selective Targeted Sampling (STS), which omits biopsy of Triple Normal LNs and compared it firsthand to SS. Methods: A prospective, feasibility RCT was conducted to determine whether the progression of a definitive trial was warranted. Primary outcomes and their progression criterium were recruitment rate (70% acceptable minimum); procedure length (no overlap between sampling methods’ 95%CIs); and missed nodal metastasis (overlap between sampling methods’ diagnostic accuracy 95%CIs and crossing of the null for the percent difference in diagnosis). cN0-N1 NSCLC patients undergoing EBUS were randomized to the STS or SS arm. Patients in the STS arm were then crossed over to the SS arm to receive standard of care. Wilson’s CI method and McNemar’s test of paired proportions were used for statistical comparison. Surgical pathology was the reference standard. Results: Thirty-eight patients met the eligibility criteria, and all were recruited (100%; 95%CI: 90.82 to 100.00%). The median procedure lengths, in minutes, for STS and SS were 3.07 (95%CI: 2.33 to 5.52) and 19.07 (95%CI: 15.34 to 20.05) respectively. STS had a diagnostic accuracy of 100% (95%CI: 74.65% to 100.00%), whereas SS was 93.75% (95%CI: 67.71% to 99.67%) with the inclusion of inconclusive results. Percent difference in diagnosis between sampling method was 5.35% (95%CI: -0.54% to 11.25%). Conclusion: With the progression criteria successfully met, a subsequent multicentered, non-inferiority crossover trial comparing STS to SS is warranted.