Endosonography for lung cancer staging: predictors for false-negative outcomes

Objectives Non-small cell lung cancer (NSCLC) guidelines recommend endosonography (endobronchial [EBUS] and/or transesophageal ultrasound [EUS]) as the initial step for mediastinal tissue staging. Identifying predictors for false negative results could help establish which patients should undergo confirmatory surgical staging. Materials and methods 775 NSCLC patients staged negative by EBUS, EUS or combined EUS/EBUS were retrospectively analyzed. Predictors of false-negative outcomes were identified by logistic regression analysis. Results and conclusion Three predictors for false-negative outcomes were identified: central location of the lung tumor (OR 3.7/4.5/3.6 for EBUS, EUS and EUS/EBUS respectively, p<0.05), nodal enlargement on CT (OR 3.2/2.5/4.9 for EBUS, EUS and EUS/EBUS respectively, p<0.05) and FDG-avidity of N2/N3 lymph node stations on PET (OR 4.2/4.0/7.5 for EBUS, EUS and EUS/EBUS respectively, p<0.05). One subgroup (peripheral lung tumor, nodal enlargement on CT without FDG-avidity for N2/N3) had a low predicted probability (7.8%) for false-negative EUS. For combined EUS/EBUS, two subgroups were identified: peripheral located tumor with nodal enlargement on CT but without FDG-avidity for N2/N3 (predicted probability 4.7%) and centrally located tumor without affected lymph nodes on CT or PET (predicted probability 3.4%). In conclusion, for specific well-defined subsets of NSCLC patients the low predicted probability of metastasis after negative endosonography might justify omitting confirmatory surgical staging.