Recent advances in molecular biology have revealed that lung cancer is not a single disease and that there are subsets of non-small cell lung cancer (NSCLC) with specific genetic alterations that are critical to the growth and survival of cancer cells. Alterations of the EGFR, ALK and ROS1 gene, which are present in a mutually exclusionary fashion, are representative driver oncogene mutations. Targeted drugs against each driver oncogene usually result in dramatic tumor shrinkage and prolongation of progression free survival (PFS) compared with conventional platinum doublet chemotherapy. However, there is only a weak association between WHO pathologic classification 2015 and type of driver oncogenes. Therefore, it is of utmost importance to identify who are likely to benefit from targeted drugs by performing molecular tests for each lung cancer patient who is a candidate for drug therapy. A list of driver oncogenes is further expanding; BRAF, RET, MET, HER2, NTRK1 are being recognized as new drivers that can be exploited in the clinic. It is getting more practical to screen these molecular alterations by use of next generation sequencing technology, rather than to detect each gene alterations one by one using different platforms. We have also known that not all the tumors with mutations of the same gene behave similarly. For example, while deletional mutation in exon 19 and L858R in exon 21 are two representative mutations that sensitize cancer cells to EGFR-tyrosine kinase inhibitors (TKI), G719X in exon 18 has an intermediate sensitivity and insertional mutation in exon 20 or de novo T790M are known to be resistant. It has been shown that there is a heterogeneity in efficacy of EGF-TKIs depending on the class of mutation. For example, afatinib is active among other EGFR-TKIs for exon 18 mutations. Furthermore, a certain molecular context is known to be associated with primary resistance even within lung cancers with the same EGFR mutations. For example, it is reported that mutations in the PI3K/AKT/mTOR pathway (AKT1, PIK3CA, STK11, PTEN) or TP53 mutations are more frequent in non-responders and are associated with shorter PFS. This context dependence may present in other driver oncogenes, too. Acquired resistance is almost inevitable in the treatment of lung cancer with targeted drug. Mechanisms of this resistance has been extensively studied and now we know there are at least 3 types of mechanisms; i.e., 1) target modification by the secondary mutation that alters the affinity between the drug and the target relative to the affinity between ATP and the target (e.g., T790M in EGFR, L1196M in ALK), 2) accessory pathway activation that bypass the inhibitory effect of the drug(e.g., Met amplification in EGFR), and histologic transformation, such as small cell lung cancer transformation and epithelial-mesenchymal transition. We are now able to use the newer generation of TKIs to treat some of the resistance due to the secondary mutation of the target gene. Osimertinib has recently been shown to prolong PFS of patients who acquired resistance to EGFR-TKI through T790M mutation compared with platinum-pemetrexed in the AURA 3 trial. Therefore, detection of this mutation which accounts for about 50∼60 % of the acquired resistance against EGFR-TKI is important. However, re-biopsy is sometimes more challenging compared with that in the first-line setting, and therefore detection of T790M in cell-free DNA in plasma has been rapidly developed and is now approved in regulatory authorities in several countries. There is another issue which should be taken into consideration when treating patients with acquired resistance. When there are multiple metastatic lesions, resistance mechanisms may vary from one tumor to another. Hence, it can happen that while one tumor shrinks but others increase in size. It may be reasonable and thus beneficial for patients when treatment is planned according to most prevalent mechanism of resistance in the plasma as a sum of total resistant mechanism. In this talk, I would like to overview recent advances of molecular diagnosis in targeted therapy of lung cancer and also like to discuss future perspectives in this field.