Imatinib for melanomas harboring mutationally activated or amplified KIT arising on mucosal, acral, and chronically sun-damaged skin

Abstract

Purpose: Amplifications and mutations in the KIT proto-oncogene in subsets of melanomas provide therapeutic opportunities.

Patients and methods: We conducted a multicenter phase II trial of imatinib in metastatic mucosal, acral, or chronically sun-damaged (CSD) melanoma with KIT amplifications and/or mutations. Patients received imatinib 400 mg once per day or 400 mg twice per day if there was no initial response. Dose reductions were permitted for treatment-related toxicities. Additional oncogene mutation screening was performed by mass spectroscopy.

Results: Twenty-five patients were enrolled (24 evaluable). Eight patients (33%) had tumors with KIT mutations, 11 (46%) with KIT amplifications, and five (21%) with both. Median follow-up was 10.6 months (range, 3.7 to 27.1 months). Best overall response rate (BORR) was 29% (21% excluding nonconfirmed responses) with a two-stage 95% CI of 13% to 51%. BORR was significantly greater than the hypothesized null of 5% and statistically significantly different by mutation status (7 of 13 or 54% KIT mutated v 0% KIT amplified only). There were no statistical differences in rates of progression or survival by mutation status or by melanoma site. The overall disease control rate was 50% but varied significantly by KIT mutation status (77% mutated v 18% amplified). Four patients harbored pretreatment NRAS mutations, and one patient acquired increased KIT amplification after treatment.

Conclusion: Melanomas that arise on mucosal, acral, or CSD skin should be assessed for KIT mutations. Imatinib can be effective when tumors harbor KIT mutations, but not if KIT is amplified only. NRAS mutations and KIT copy number gain may be mechanisms of therapeutic resistance to imatinib.

Trial registration: ClinicalTrials.gov NCT00424515.

Conflict of interest statement

Authors' disclosures of potential conflicts of interest and author contributions are found at the end of this article.

Figures

Fig 1.

Kaplan-Meier curves for clinical efficacy by KIT mutational status. Time to progression (TTP) analysis. (A) All patients showed a median TTP of 3.7 months (95% CI, 2.6 to 5.6 months). (B) Subset TTP analysis showed medians of 3.9 months (95% CI, 2.6 to 6.6 months) in those with mutations and 3.4 months (95% CI, 1.0 to 5.7 months) in those with amplifications. Overall survival (OS) analysis for (C) all patients showed a median of 12.5 months (95% CI, 8.8 to 18.0 months) with (D) subsets showing medians of 12.9 months (95% CI, 5.5 to 24.3 months) for those with mutations and 11.9 months (95% CI, 4.5 to 16.2 months) for those with amplifications.

Fig 2.

Representative computed tomography and positron emission tomography (PET) study of a responding patient. Cross-sectional axial slices from (A,B) computed tomography images and (C,D) [18F]fluorodeoxyglucose PET ([18F]FDG-PET) images at baseline (left) and after 1 month of therapy (right) show resolution of FDG uptake and decrease in the size of a right inguinal lymph node (gold arrow) and a left perirectal soft tissue nodule (blue arrow).

Fig A1.

Examples of fluorescent in situ hybridization analyses. Probes were labeled by random octamer priming and hybridized after microwaving and pepsin digestion. Adjacent sections were stained by hematoxylin and eosin to identify tumor-containing regions. The KIT locus on the chromosome 4 (Chr4) long arm was evaluated by using a mini-contig of bacterial artificial chromosome (BAC) probes (digoxigenin-labeled BACs C00-840L10 and RP11-586A2) cohybridized with a biotinylated chromosome 2 (Chr2) centromeric probe, the latter serving as a copy number control, given that the chromosome 2 centromeric region is rarely deleted or gained in melanoma. Signal detection was strepavidin-Alexa598 (red) for the chromosome 2 centromere, and fluorescein isothiocyanate-antidigoxigenin (green) for KIT. Two hundred nuclei were analyzed by using guidelines proposed by Hopman et al (Curr Protoc Cytom Chapter 8:Unit 8.5, 2001). (A,B: patient 102) Both chromosome 4 and KIT were increased (low-level gain of a major part of chromosome 4). Chromosome 2 centromere was not increased (better copy number was a neutral indicator of genome ploidy). A: Chr2, red; KIT, green; B: Chr4, red; KIT, green. (C,D; patient 116) High-level/localized KIT amplification, not extending to chromosome 4 centromere. Both chromosome 2 and chromosome 4 centromeres were informative for copy number reference. C: Chr2, red; KIT, green; D: Chr4, red; KIT, green. (E) Examples of KIT/chromosome 2 and KIT/chromosome 4 ratios.