In this issue of CHEST (see page 1628), Horiike and colleagues1 present us with a demonstration of modern technological magic applied to the diagnosis and treatment of lung cancer. They illustrate how the combination of a sensitive diagnostic technique and a sensitive molecular assay can extract information from small samples that is comparable to that achieved by more conventional approaches. Although bronchoscopic sampling techniques of biopsy, brushing, washing, and BAL have become standard approaches for diagnosis of lung cancers, transbronchial fine-needle aspiration (TBNA) is less commonly used. Its success in obtaining diagnostic material is highly dependent on the skills of the operators, including the person performing the procedure and the person interpreting the slides produced. Horiike and colleagues1– collected TBNA samples from 93 patients with a cytologically confirmed diagnosis of non-small cell carcinoma (established from the TBNA samples) and subjected the TBNA samples to molecular analysis for epidermal growth factor receptor (EGFR) mutations. The latter testing was performed using two methodologies: direct sequencing, which is currently the routine method used for detecting EGFR mutations in tumor samples, and a newer assay (Scorpion Amplification Refractory Mutation System [ARMS]; DxS; Manchester, UK), which employs primers consisting of a specific probe sequence held in a hairpin loop configuration by complementary stem sequences on the 5′ and 3′ sides of the probe.3 Sufficient polymerase chain reaction products were obtained to allow for direct sequencing in samples from 83 patients, while EGFR mutation status could be analyzed in 91 patients using the EGFR Scorpion ARMS method. EGFR mutations were detectable in 31 patients, with detection by both methods in 9 patients, Scorpion ARMS alone in 18 patients, and direct sequencing alone in 4 patients, leading the authors to conclude that the Scorpion ARMS method is more sensitive for detecting the most common EGFR mutations than direct sequencing.