First Multicenter, Randomized Phase 3 Study in Patients (Pts) with Relapsed/Refractory (R/R) Peripheral T-Cell Lymphoma (PTCL): Alisertib (MLN8237) Versus Investigator's Choice (Lumiere trial; NCT01482962)

Background PTCL is an aggressive subgroup of NHL. Outcomes in R/R PTCL remain poor and alternative therapies are needed. Alisertib is an investigational, oral, selective inhibitor of Aurora A kinase, a key mitotic regulator that is overexpressed/amplified in various cancers, including lymphomas. Phase 2 data with single-agent alisertib in R/R NHL (Friedberg et al JCO 2014) and R/R PTCL (Barr et al JCO 2015) suggested promising antitumor activity. This international phase 3 trial evaluated single-agent alisertib vs investigator's choice in pts with R/R PTCL, and was the first initiated randomized trial in this setting.

Methods Adult pts with R/R PTCL (WHO criteria) after ≥1 prior conventional systemic cytotoxic therapy, who had measurable disease by 2007 IWG criteria, tumor biopsy for central hematopathology review, and ECOG PS 0-2, were eligible. Pts were randomized 1:1 (stratified by nodal vs extranodal disease, IPI score [0/1/2 vs 3/4/5], and region [North America + EU vs Rest of World]) to receive alisertib 50 mg twice daily as an enteric coated tablet on d 1-7 in 21-d cycles (alisertib arm), or investigator's choice of: pralatrexate 30 mg/m² IV once weekly for 6 weeks in 7-week cycles; romidepsin 14 mg/m² IV on d 1, 8, and 15 in 28-d cycles; or gemcitabine 1000 mg/m² IV on d 1, 8, and 15 in 28-d cycles (comparator arm), until disease progression, unacceptable toxicity, or 2 yrs (extendable if benefit shown). Selected comparator drug could not have been previously received by the pt, and crossover to another study drug was not permitted. The primary endpoints were ORR (CR+PR) and PFS by IWG criteria per central (IRC) assessment. Secondary endpoints included OS, CR rate, DOR, and safety. The study employed an adaptive sample size reassessment approach, with two interim analyses (IA1, futility analysis; IA2) plus a planned final analysis. Sample size was determined to give 80% power to detect a difference in ORR (assumed 55% alisertib vs 30% comparator) at IA2, and ~85% power to detect a reduction in HR for PFS in favor of alisertib at the final analysis with a maximum of 354 randomized pts and a maximum of 261 PFS events. Conditional power calculation based on PFS at IA2 could determine whether the study would be stopped for futility or if sample size would be expanded for the final analysis. Here we report data from IA2 (17Sep2014 data cut-off).

Results 238 pts were randomized across 27 countries (120 alisertib, 118 comparator). Baseline characteristics were balanced between arms (alisertib vs comparator): median age 63 vs 64 yrs; male 68% vs 64%; Ann Arbor stage III/IV 74% vs 72%; bone marrow involvement at study entry 29% vs 35%. Efficacy data are shown in the Table. ORR by IRC with alisertib vs comparator was 33% vs 43% (OR 0.65 [95% CI: 0.34-1.23]), including 16% vs 25% CR. Median duration of follow-up was 9.5 vs 9.2 mos with alisertib vs comparator. Median PFS by IRC was 3.7 vs 3.4 mos (HR 0.939 [95% CI: 0.681-1.293]; Figure) and median OS was 9.9 vs 12.2 mos (HR 0.901 [95% CI: 0.607-1.337]; OS data not mature at current follow-up). Median treatment duration with alisertib vs comparator was 12 vs 10 weeks and 15% vs 5% of pts remained on treatment at data cut-off. With alisertib vs comparator, rates of Gr ≥3 AEs were 85% vs 81%, serious AEs 53% vs 55%, discontinuations due to AEs 9% vs 13%, and on-study treatment-related deaths 2% vs 2%. Gr ≥3 AEs (>20% all pts) were neutropenia 44% vs 27%, thrombocytopenia 29% vs 27%, and anemia 30% vs 11%.

Conclusion While alisertib showed activity in R/R PTCL, there was no significant efficacy benefit vs comparator. Based on IA2, there was a low probability of claiming superiority of alisertib for PFS at the final analysis; therefore, per IDMC recommendation enrollment was stopped at 271 pts and pts deriving benefit continue on treatment. The study was unblinded and final data are pending.

View largeDownload slide

View largeDownload slide

Close modal