INTRODUCTION: Next-generation sequencing (NGS) of multiple metastases in an advanced cancer patient reveals the evolutional history of the tumor. The evolutionary model is clinically valuable because it reflects the future course of the tumorigenic process and prognosis of the patient. MATERIALS AND METHODS: We experienced two lung cancer patients whose clinical courses were abruptly deteriorating resulting in very poor prognosis. To investigate the evolutionary model of these patients, we performed targeted sequencing covering whole exons of 53 significantly mutated genes associated with lung cancer of multiple metastases by autopsy. We conducted PyClone analysis to infer subclonal archtecture of multi-lesional samples. RESULTS: The NGS analysis revealed both patients harboring multiple clonal driver mutations. In Case.1, KRAS Q61H, KEAP1 G333C, STK11 K312*, RBM10 Q320* and MGA I1429V and in Case.2, TP53 R337L, TP53 Q192*, PTEN W274C, RB1 P29fs and CREBBP P696L with high allele fraction were detected in all lesions. These mutations were clustered and occupied major population across multi-lesional tumor samples. Our data suggested their lung cancers progressed with punctuated and big bang evolutional model. CONCLUSION: We should pay attention to clinical course of lung cancer patients harboring multiple clonal driver mutations in their primary lesions. Their punctuated and big bang evolutionary process could develop systemic clinically undetectable metastases with an unexpected speed.