In this exploratory Phase I/II biomarker study (as outlined by the Early Detection Research Network (EDRN) (http://edrn.nci.nih.gov
)) we screened serum from patients with no history of cancer, and patients with NSCLC in an effort to identify miRNAs that can be used as biomarkers for the detection of early stage lung cancer. We identified a miRNA pair miR-15b/miR-27b that was able to distinguish between serum from NSCLC patients and cancer-free healthy controls, and with a high degree of sensitivity. Our results support the findings of recent studies that have shown that circulating miRNAs profiles may be useful in screening for NSCLC, 
however, our study included substantially more patients (180 total) than either of these previous studies. One drawback of these biomarkers is the low specificity of 84% in the test set. The relatively high false positive rate for these tests could be considered unacceptable for screening the general population. However, in a population of smokers at high risk for lung cancer, unnecessary screening would be mitigated by the ability of this test, with its negative predictive value of 100%, to be able to exclude a large number of patients from going on to more expensive screening modalities, such as helical chest CT. Although these data are compelling, further testing in a large, prospective cohort as a Phase III biomarker study is required to assess the clinical utility of these miRNA markers as a first line screening test for NSCLC.
The miRNAs markers identified in this study have previously been implicated in human malignancies. miR-15a and miR-15b have been shown to be de-regulated in human lung cancer 
. Both miR-15a and miR-15b have been shown to have a diagnostic and prognostic value in chronic lymphocytic leukemia 
. miR-27b was found to be down-regulated in lung cancer tissues compared to non-cancerous lung tissue 
. In addition, miR-27b expression levels have been correlated with invasiveness of breast cancer 
and with regulation of angiogenesis 
. A study involving a total of 86 NSCLC and 57 controls recently revealed a four miRNA panel in plasma including miR-126 that distinguished NSCLC from healthy controls with a sensitivity and specificity of 73% and 96% respectively 
. Using whole blood, Keller and coworkers 
showed that miR-126 and miR-98 were among the top miRNAs that could distinguish NSCLC from healthy controls. miR-126 is highly expressed in lung tissue and is involved in the regulation of vascular cell adhesion molecule 1 (VCAM-I) 
. Aberrant expression of miR-126 has been implicated in the pathogenesis of NSCLC 
. In this study, several pairs involving miR-126 were among the several candidates identified in the training set (Figure S1
), however upon further testing in additional 130 samples, all the miRNA-126 candidate pairs exhibited sensitivity greater than 75% and the specificity was less than 75%. Other recent studies by Foss and coworkers 
showed serum miR-1254 and miR-574-5p were differentially expressed between NSCLC (n
33) and healthy controls (n
42). It is important to note that several factors affect the outcome of miRNA studies in biofluids, including variations in sample type e.g. whole blood, plasma or serum, sample numbers, study design, sample collection, RNA isolation, patient characteristics, number of miRNA examined and technologies used in miRNA profiling e.g. solexa sequencing, RT-qPCR or microarray technologies. In addition, standardization of isolation methods, normalization and data analysis methods is needed in order to demonstrate a clear clinical utility of these putative markers.
Although these data are promising, the test set included a relatively small number of samples. Ultimately, these miRNA biomarkers require further validation on larger prospective cohorts such as a Phase III biomarker study in order to validate these results. Incorporating blood-based miRNA markers in spiral CT studies may aid in exploring the utility of miRNAs in screening of lung cancer. Although this is an exploratory phase I/II trial, the patients were selected primarily from surgical clinics, and are weighted towards early stage disease (60% Stage I, 24% Stage II, 12% Stage III and 4% Stage IV). This skew towards early stage disease supports the investigation of these markers in a phase III or IV trial aimed at defining the performance of these markers in a prospective manner in early stage detection.
The recent dissemination of the utility of screening helical chest CT scans for reduction in mortality from lung cancer from the NLST trial places a premium on identification of high risk individuals who could benefit from screening. Adjunctive serum based testing may be performed in a highly cost effective manner compared to imaging, and may be helpful to identify high risk populations that may benefit from chest CT, or to be used in combination with imaging to identify early lung cancers.
There is great need for improved screening for lung cancer given the large number of people affected each year and the high mortality rate of the disease when diagnosed in its later stages. There are currently numerous clinical trials being conducted to test the efficacy of novel therapies for NSCLC, however, the majority of these are Phase II trials and recently a number of Phase III trials have failed to meet their primary end points 
To date, improved screening to provide early detection is the most promising avenue to reduce mortality from NSCLC. Our study further strengthens the argument that serum miRNA have the potential to be used as a cost effective, non-invasive diagnostic test for NSCLC, and could potentially be used as a first line screen to help risk stratify patients for further, more expensive or invasive screening regimens.