Rationale: Effective treatment for lung cancer requires accuracy in subclassification of carcinoma subtypes.
Objectives: To identify microRNAs in bronchial brushing specimens for discriminating small cell lung cancer (SCLC) from non-small cell lung cancer (NSCLC) and for further differentiating squamous cell carcinoma (SQ) from adenocarcinoma (AC).
Methods: Microarrays were used to screen 723 microRNAs in laser-captured, microdissected cancer cells from 82 snap-frozen surgical lung specimens. Quantitative reverse-transcriptase polymerase chain reaction was performed on 153 macrodissected formalin-fixed, paraffin-embedded (FFPE) surgical lung specimens to evaluate seven microRNA candidates discovered from microarrays. Two microRNA panels were constructed on the basis of a training cohort (n = 85) and validated using an independent cohort (n = 68). The microRNA panels were applied as differentiators of SCLC from NSCLC and of SQ from AC in 207 bronchial brushing specimens.
Measurements and main results: Two microRNA panels yielded high diagnostic accuracy in discriminating SCLC from NSCLC (miR-29a and miR-375; area under the curve [AUC], 0.991 and 0.982 for training and validation data set, respectively) and in differentiating SQ from AC (miR-205 and miR-34a; AUC, 0.977 and 0.982 for training and validation data set, respectively) in FFPE surgical lung specimens. Moreover, the microRNA panels accurately differentiated SCLC from NSCLC (AUC, 0.947) and SQ from AC (AUC, 0.962) in bronchial brushing specimens.
Conclusions: We found two microRNA panels that accurately discriminated between the three subtypes of lung carcinoma in bronchial brushing specimens. The identified microRNA panels may have considerable clinical value in differential diagnosis and optimizing treatment strategies based on lung cancer subtypes.