While intertumoral heterogeneity is well recognised in many solid tumours including NSCLC, intratumoral heterogeneity has only recently gained attention. Heterogeneity of tumor morphology, protein expression, gene expression, epigenetic or genetic alterations has the potential to impact optimal biopsy strategies, diagnostic assessment, treatment decisions and clinical outcome. Sequencing of NSCLC from multiple sites of disease shows frequent evidence of intratumoral heterogeneity in terms of genetic mutations, translocations and copy number alterations, although not to the same extent as seen in other tumor types, such as clear cell renal cell carcinoma. NSCLC studies have demonstrated a common pattern of intratumoral heterogeneity with main clonal driver mutations and branched evolutionary acquired mutations. Of clinical relevance, mutations in known lung cancer driver oncogenes (such as EGFR, BRAF and MET) are generally present in all tumor regions in keeping with early evolutionary events. This finding is consistent with the high response rates to tyrosine kinase inhibitors that target these genetic alterations, across multiple sites of disease. Later subclonal driver mutations are found commonly in NSCLC and include alterations in genes such as PIK3CA and NF1. Metastatic sites can exhibit mutational profiles closely related to specific spatial regions of the primary tumor demonstrating that subclones can determine the course of systemic disease resulting in subclonal diversification. Clonal evolution is driven by multiple factors including selective pressure from targeted therapies and adaptive mechanisms due to interaction with immune cells and the microenvironment. Treatment resistance can occur due to acquisition and/or selection of clones and contributes to temporal heterogeneity. The hierarchy of genetic alterations can be used to trace clonal intratumoral heterogeneity although adequate sequencing depth is required to accurately assess for subclonal mutations. Reassuringly, sequencing of a single region of a tumor should be sufficient to identify known targetable driver mutations as they generally occur early in the evolutionary course of the disease. The exact clinical significance of various subclonal mutations is less well understood. Intratumoral heterogeneity can potentially lead to sampling errors when single sites of disease are sampled for mutational events that may only exist in another metastatic site. For this reason, testing for genetic markers of treatment resistance may be more appropriately performed on circulating tumour DNA as the ctDNA may derive from multiple metastatic deposits, although lower sensitivity limits the effectiveness of this approach. Liquid biopsy approaches also have the advantage of providing a contemporaneous sample, more likely to reflect impact of most recent therapy. Further investigation of spatial and temporal tumoral heterogeneity by comprehensive deep sequencing of multiple spatially discrete sites of disease at different time points may assist in understanding the complexity of intratumoral heterogeneity and could potentially impact optimal biopsy and treatment strategies, particularly when assessing for drug resistance. genetic heterogeneity NSCLC