Crizotinib in ROS1 and MET Deregulated NSCLC—Response

We thank the authors for their interest in our study and for their comments. Wiesweg and colleagues correctly reported that the minor allele frequency of both p.T1010I and p.R988C MET variants is close to 1% in the European population, but these mutations were described also as somatic (T1010I: COSM707; R988C: COSM1666978; refs. 1, 2). Interpretation of both alterations is still conflicting and debated (ClinVar: p.T1010I-VCV000041624.1 and p.R988C-VCV000041623.2; oncokb.org). Nevertheless, these mutations were associated to a more aggressive cancer phenotype affecting cell morphology, adhesion, and altering MET receptor tyrosine kinase signaling (1, 3). In addition, different prediction algorithms describe p.T1010I and p.R988C mutations as pathogenic or likely pathogenic (i.e., FATHMM prediction; Mutation Tester). The METROS trial included patients with MET amplification or with exon 14 mutations, without any restriction for the type of mutation (4). All enrolled patients fulfilled the inclusion criteria and among mutant patients we observed only two responders, one with skipping mutation and one with the p.T1010I mutation. We agree with Wiesweg and colleagues that the number is too low for any definitive conclusion, but our data are more suggestive for an unsatisfactory MET inhibition rather than a lack of appropriate selection. Our data are in agreement with the results of a recently published French trial showing modest results with crizotinib as MET inhibitor (5).

In general, driver events are mutually exclusive. Nevertheless, cooccurrence of molecular alterations in different genes (including EGFR, ALK, ROS1, and MET) has been reported. In addition, cooccurrence of different drivers has been described in patients with acquired resistance to target therapies, with no data after chemotherapy. We cannot exclude that previous exposure to chemotherapy, one of the main METROS inclusion criterion, could modify the molecular profile of the disease. For all these reasons, cooccurrence of mutations should not be considered as a confirmation of the germline and nonpathogenic nature of a variant; for instance, in our study population we found cases with ROS1 rearrangement and MET amplification cooccurring with KRAS and BRAF mutations.

In conclusion, METROS and other studies showed modest efficacy for crizotinib in MET-altered patients highlighting the urgent need for alternative strategies against this aggressive disease.

See the original Letter to the Editor, p. 1774