ALK Mutations Possible Predictive Marker for Lorlatinib in Previously Treated Patients With ALK+ NSCLC

Alice T. Shaw, MD

Alice T. Shaw, MD, PhD

Significantly higher response rates were seen in patients withALK-positive non—small cell lung cancer (NSCLC) who had been treated with at least 1 prior ALK tyrosine kinase inhibitor (TKI) before receiving lorlatinib (Lorbrena) and who harbored anALKresistance mutation compared with those who did not have anALKmutation, according to the results of an analysis from a phase II trial.¹Findings suggested that patients previously treated with a second-generation ALK TKI whose tumor had anALKmutation may be more sensitive to treatment with lorlatinib than those without a resistance mutation.  

The objective response rate (ORR) with lorlatinib treatment in patients who had prior exposure to a second-generation ALK TKI, either ceritinib (Zykadia) or alectinib (Alecensa), was almost doubled in patients who harbored anALKmutation (69%) compared with those without anALKmutation (27%). This finding was especially increased in patients who underwent tumor tissue genotyping to identifyALKmutations.

“These findings suggest that in patients who have previously received one or more second-generation ALK inhibitors, the presence of anALKmutation based on tissue genotyping may identify a subgroup of patients more likely to derive durable benefit from lorlatinib,” the study authors, led by Alice T. Shaw, MD, PhD, of Massachusetts General Hospital, wrote in their report recently published in theJournal of Clinical Oncology.

Findings from the phase II registrational trial,²which led to the FDA-accelerated approval for lorlatinib as a treatment for patients withALK-positive metastatic NSCLC who have progressed on 1 or more ALK TKIs, demonstrated that the third-generation ALK inhibitor induced an ORR of 69.5% (95% CI, 56.1%-80.8%) in patients who have previously been treated with crizotinib (Xalkori), the first-generation ALK TKI. The ORR in patients who had prior treatment with ≥2 ALK TKIs was 38.7% (95% CI, 29.6%-48.5%). 

The phase II trial enrolled 228 patients, 30 of whom were treatment-naïve and excluded from this analysis. The other 198 patients were divided into expansion cohorts (EXP) 2 through 5 based on their previous treatment exposure, which included: 27 patients who had only received prior crizotinib (EXP2), patients who had received prior crizotinib and chemotherapy (n = 32; EXP3A), those who had 1 prior second-generation ALK TKI with/without chemotherapy (n = 28; EXP3B), those who had 2 prior ALK TKIs with/without chemotherapy (n = 65; EXP4), and patients who had received 3 prior ALK TKIs with/without chemotherapy (n = 46; EXP5).

Molecular profiling assessments were completed at baseline of both cell-free DNA (cfDNA) and tumor tissue. Tumor tissue, whether archival or de novo, was analyzed with a customized next-generation sequencing (NGS) assay of formalin-fixed paraffin-embedded tissue samples for detection ofALKkinase domain mutations from exons 20 to 25 on the Ion Torrent PGM platform. Plasma genotyping was completed with the Guardant360 73-gene cfDNA NGS assay.

Of the previously treated patients, 189 patients (95%) had baseline cfDNA testing completed; of these, 45 patients (24%) had ≥1 ALKmutation by cfDNA testing and 40 (21%) had no detectable cfDNA. Additionally, 191 patients had available tissue samples, 98 of which were de novo, and 164 (86%) of these samples were evaluable for NGS analysis and 40 (24%) showed ≥1 ALKmutation.

The most commonALKmutations discovered by plasma genotyping were G1202R/del in 42%, L1196M in 24%, F1174X in 24%, G1269A in 18%, and I1171X in 11%; and by tumor genotyping: G1202R/del in 40%, F1174X in 20%, I1171X in 13%, and G1269A in 13%. Each of these resistance mutations have demonstrated sensitivity to lorlatinib in preclinical models.

The spectrum and frequency of theALKmutations different according to the various prior ALK TKIs patients had received. Among patients who had received prior crizotinib (EXP2 to EXP3A), the most common resistance mutations were G1269A, F1174X, and L1196M. On the other hand, among patients who had received ≥1 prior second-generation ALK TKI (EXP3B to EXP5), the most common resistance mutation was G1202R/del, which was detected in 53% with cfDNA testing and 55% with tumor tissue testing.

The sensitivity of plasma genotyping forALKmutations was 61% overall (95% CI, 44%-76%) and the specificity was 82% (95% CI, 67%-91%). The agreement of plasma genotyping and de novo tumor tissue genotyping was 73% (95% CI, 61%-82%).

Responses by Mutation Status

Updated findings after a median follow-up of 16.6 months in the previously treated patients (EXP2-5) demonstrated that patients with prior crizotinib exposure (n = 59; EXP2 to EXP3A) had an ORR of 73% (95% CI, 60%-84%). In these patients, the median duration of response (DOR) was not yet reached and the median progression-free survival (PFS) was 11.1 months.

Eleven of these patients (19%) had detectableALKmutations by plasma genotyping, and the ORR was 73% (95% CI, 39%-94%) in these patients compared with 75% (95% CI, 60%-87%) in patients who did not have anALKmutation and had received prior crizotinib treatment. Similar results were found by tumor tissue genotyping as well, with an ORR of 73% (95% CI, 39%-94%) versus 74% (95% CI, 59%-87%) for patients with and withoutALKmutations, respectively.

The median PFS was not reached in patients who had previously received crizotinib and had anALKmutation and was 12.5 months in those without anALKmutation when detected by plasma genotyping (HR, 1.03; 95% CI, 0.39-2.69). Similar findings for PFS were demonstrated with tissue testing (NR vs 12.5 months; HR, 1.38; 95% CI, 0.48-3.98) and for DOR.

The study authors commented that “most crizotinib-resistant tumors, including those without a detectableALKmutation, are still driven by ALK and remain responsive to more potent ALK inhibitors,” making lorlatinib still an effective option.

Of the patients who had previously been treated with ≥1 second-generation ALK TKI (n = 139; EXP3B to EXP5), the overall ORR was 40% (95% CI, 32%-49%), the median DOR was 7.1 months (95% CI, 5.6-24.4) and the median PFS was 6.9 months (95% CI, 5.4-8.2). By plasma genotyping, 32 of these patients (26%) had detectableALKmutations.

In patients who had received prior treatment with ≥1 second-generation ALK TKI and had an ALKmutation detected by plasma testing had an ORR of 62% (95% CI, 44%-78%) compared with 32% (95% CI, 23%-42%) in patients without anALKresistance mutation. By tumor tissue testing, the ORRs were 69% (95% CI, 49%-85%) in patients with detectedALKmutations and 27% (95% CI, 18%-38%) in those without resistance mutations, and 69% and 31%, respectively, according to de novo tissue testing.

By plasma genotyping, the median PFS was 7.3 months (95% CI, 4.1-13.1) in patients with anALKmutation versus 5.5 months (95% CI, 4.1-8.2) in those without (HR, 0.81; 95% CI, 0.50-1.31). But by tumor tissue genotyping, the median PFS was 11.0 months (95% CI, 6.9-NR) and 5.4 months (95% CI, 3.9-6.9) for those with and withoutALKmutations, respectively (HR, 0.47; 95% CI, 0.27-0.83), and the median DOR was 24.4 months versus 4.3 months.

“These results may reflect the current limitations of plasma genotyping and the fact that plasma-negative cases include both true negatives and false negatives, potentially compromising the ability of plasma to identify 2 distinct subgroups with different PFS,” Shaw et al wrote.

The differences were even more pronounced in de novo tissue testing alone, where the median PFS was 11.0 months (95% CI, 6.9-25.6) in patients withALKmutations versus 4.0 months (95% CI, 2.6-5.5) in those without (HR, 0.20; 95% CI, 0.10-0.40).

Reponses by Type of Mutation

Antitumor activity was seen with lorlatinib against all of the 5 most commonALKresistance mutations seen in patients between groups EXP2 and EXP5: G1202R/del, F1174X, L1196M, G1269A, and I1171X.

The ORR in all patients with anALKG1202R/del mutation (n = 28), regardless of prior exposure, was 57% (95% CI, 37%-76%), the median DOR was 7 months (95% CI, 6.1-24.4) and the median PFS was 8.2 months (95% CI, 5.6-25.6).

In patients with an L1196M mutation (n = 12), the ORR was 67% (95% CI, 35%-90%) and both the median DOR and median PFS were not yet reached.

Twelve patients also had an F1174X mutation, these patients demonstrated an ORR of 42% (95% CI, 15%-72%), a median DOR that was not reached (95% CI, 5.7-not reached), and a median PFS of 7.4 months (95% CI, 2.8-not reached) from lorlatinib treatment.

Among patients with anALKG1269A mutation (n = 9), the ORR was 89% (95% CI, 52%-100%) and both the median DOR and median PFS were not yet reached.

Only 8 patients had an I1171X resistance mutation. In these patients, the ORR was 75% (95% CI, 35%-97%), the median DOR was 4.2 months (95% CI, 2.8-4.2), and the median PFS was 5.5 months (95% CI, 4.1-6.9).

Approximately one-third of patients had more than 1ALKmutation, which could be representative of compound or heterogenous resistance mechanisms, the investigators suggested, both of which are seen in heavily pretreated patients. The ORR was higher among patients with only 1 resistance mutation compared with those with ≥1ALKmutation (75% vs 56%) and the DOR was also longer in patients with only 1 mutation (24.4 vs 6.1 months, respectively).

“The association ofALKmutations with clinical response to lorlatinib after failure of a second-generation ALK TKI is reminiscent ofEGFRT790M serving as a predictive biomarker for the third-generation EGFR inhibitor osimertinib after failure of earlier-generation TKIs,” the study authors commented. “Thus, in patients who have failed a second-generation ALK TKI,ALKmutations as a group may be analogous toEGFRT790M in identifying patients who are more responsive to a highly potent third-generation TKI.”

The investigators concluded that further study is needed regarding plasma versus tissue genotyping ofALKmutations and of other genes as potential biomarkers of response and resistance to lorlatinib. One such study looking into this is the National Cancer Institute’s ALK Master Protocol trial which is looking to match patients to ALK TKIs based on resistance mutations in previously treated patients withALK+ NSCLC.

References:

1. Shaw AT, Solomon BJ, Besse B, et al. ALK resistance mutations and efficacy of lorlatinib in advanced anaplastic lymphoma kinase-positive non—small-cell lung cancer [published online March 20, 2019].J Clin Oncol.doi: 10.1200/JCO.18.02236.
2. Solomon BJ, Besse B, Bauer TM, et al. Lorlatinib in patients with ALK-positive non-small-cell lung cancer: results from a global phase 2 study.Lancet Oncol.2018;19(12):1654-1667. doi: 10.1016/S1470-2045(18)30649-1.