Are EUS-FNA and EBUS-TBNA specimens reliable for subtyping non-small cell lung cancer?

With endosonography, the diagnosis and staging of non-small cell lung cancer (NSCLC) increasingly relies on small samples. The discrimination between squamous and non-squamous subtypes is now important for therapy tailoring. We analyzed the agreement between fine needle aspirates obtained by endosonography and matched biopsy samples for subtyping NSCLC. Patients with a positive endoscopic fine needle aspirate and a matched biopsy were identified. The level of diagnostic agreement was estimated with biopsy samples as golden standard. In 951 patients investigated with endosonography, we identified 92 with NSCLC on the positive fine needle aspirate and on the matched biopsy. Squamous cell carcinoma was diagnosed in 34 (37%) and 44 (48%) of fine needle aspirate and biopsy samples; while non-squamous carcinoma was diagnosed in 58 (63%) and 48 (52%) respectively. The agreement between needle aspirate and biopsy for the subtyping of NSCLC was 76% (kappa=0.52). In cases with cell block preparation, the agreement for subtyping was 96% (kappa=0.91) vs 69% (kappa=0.39) in cases without cell blocks. Therefore, the diagnostic agreement between endosonographic fine needle aspirates and biopsy specimens for subtyping NSCLC is moderate with a disagreement in 1 out of 4 patients. However, cell block preparation increased the agreement and thus the reliability of the fine needle specimens obtained during endosonography, for subtyping NSCLC considerably. In conclusion, for patients with NSCLC in whom subtyping is relevant, a diagnostic technique yielding larger samples (FNA with cell block preparation or biopsies) should be preferred.