Background: Crizotinib is an effective treatment for patients with ALK rearranged NSCLC but, unfortunately, patients progress with a median of 9-10 months. Resistance could be acquired through ALK point mutations, amplification or activation of alternative pathways i.e. KIT, EGFR and KRAS [1]. Second-generation TKIs demonstrated an enhanced activity in crizotinib-resistant patients that retain ALK addiction. However, rebiopsy represents a critical issue in lung cancer and analysis of circulating tumor DNA (cftDNA) has a promising role for the identification of resistance to treatment. Methods: Sixteen NSCLC patients with ALK rearrangement were enrolled after progression to ALK-TKI. Blood was collected at several time points during and after TKI treatment. CftDNA was extracted from plasma using QIAamp circulating nucleic acid kit (Qiagen) and ALK and KRAS codon 12 mutations were tested using the Digital Droplet PCR (ddPCR - BioRad). Results: All patients were stage IV adenocarcinoma; 15 patients received crizotinib and 1 ceritinib. ALK-TKIs were administered mainly as second-line, in 2 cases as first and in the remaining as third-line therapy. The best responses to treatment were 12 partial responses (PRs), 3 stable diseases (SDs) and 1 progressive disease (PD). The median PFS was 8 months. ALK point mutations alone were identified in 2 patients; one showed both p.L1196M and p.G1269A mutations. Six patients were carriers of a KRAS mutation at crizotinib PD, 2 of them presented both ALK and KRAS mutations. CftDNA was monitored during second generation ALK-TKI therapy and the amount of both ALK or KRAS mutations decreased along with tumor response. Interestingly, one additional extra-patient analysed before treatment with crizotinib was carrying the ALK p.F1174L mutation and showed primary refractoriness. Conclusions: KRAS and ALK mutations induce acquired resistance to TKIs targeting ALK rearrangement. The detection of KRAS mutation in circulating tumor DNA could represent a predictive marker of acquired resistance. Moreover, the quantification of mutant alleleS burden could be useful to monitor treatment response. 1. Katayama R, et al. Sci Transl Med. 2012 Feb;4(120):120ra17