The Combination of Duvelisib, a PI3K-δ,γ Inhibitor, and Romidepsin Is Highly Active in Relapsed/Refractory Peripheral T-Cell Lymphoma with Low Rates of Transaminitis: Results of Parallel Multicenter, Phase 1 Combination Studies with Expansion Cohorts

Introduction: Most therapies for relapsed/refractory (rel/ref) T-cell lymphomas (TCLs) induce responses in about 25%-30% of pts. Phosphoinositide-3-kinase (PI3K) delta has a role in survival and proliferation of malignant T cells as well as T-cell receptor and cytokine signaling in nonmalignant T cells. Inhibition of PI3K gamma can reprogram macrophages and promote tumor phagocytosis. A phase I study of the PI3K-δ/γ inhibitor duvelisib (D) in pts with rel/ref TCLs showed promising activity, but high rates of grade (Gr) 3/4 ALT elevation at the MTD of 75 mg BID (Horwitz et al, Blood 2018). Based on in vitro evidence of synergy, we initiated a phase I/II study of D combined with either romidepsin (R) or bortezomib (B).

Methods: Pts were enrolled into parallel phase I dose escalation arms utilizing a 3+3 design to define the maximum tolerated dose (MTD) of D combined with either R (Arm A) or B (Arm B) in cycle 1. D was dosed at 25 mg, 50mg, or 75 mg BID on days 1-28 with either R (10 mg/m2 on days 1, 8, & 15) or B (1 mg/m2 on days 1, 4, 8, & 11) each on 28-day cycles. Once the MTD was established with each combination, preplanned expansion cohorts were enrolled to further characterize safety and describe subtype-specific efficacy (PTCL and CTCL). Based upon promising safety and efficacy, a total of 39 pts were treated on Arm A (33 at the MTD, D 75 mg BID + R 10 mg/m2). All pts received prophylaxis against Varicella and Pneumocystis. Response assessments were performed q2 cycles for 6 months and then q3 cycles. To assess biomarkers of response and resistance to single-agent D, 10 pts in Arm A (75 mg BID) and 10 pts in Arm B (25 mg BID) were each treated with 1 cycle of D alone, with pretreatment and on-treatment biopsies. Pts who did not achieve CR on D alone at the end of cycle 1 proceeded to combination therapy.

Results: The MTD was not reached in Arm A (R+D); thus, dose level 3(DL3); (D 75 mg BID + R 10 mg/m2 days 1, 8, & 15) was deemed the MTD and used for expansion. In Arm B there were no cycle 1 DLTs. However, Gr 3 elevations of ALT or AST following cycle 2 were observed in 3 pts at DL2 (D 50mg BID) and 2 pts at DL3 (D 75mg BID) leading to DL1 (D 25mg BID + B 1mg/m2 days 1, 4, 8, & 11) being accepted as MTD for expansion.

Of Arm A pts at the MTD, 21/32 (65%) had adverse events (AEs) ≥Gr 3, possibly related to study drug. Events occurring in ≥10% of pts included: increased ALT/AST (n=5, 15%), neutropenia (n=6, 18%), and hyponatremia (n=4, 12%). Three pts had ≥Gr 3 diarrhea. There were no Gr 5 AEs related to protocol therapy. Strikingly, 4 of 5 pts with elevated transaminases (ALT [4], AST [1]) on combination began on the D-only Lead-In Arm at 75 mg BID. In contrast, only 1 of 22 (4%) pts receiving combination R+D in cycle 1 had Gr 3-4 transaminitis (p=.0242). Of the pts with Gr 3-4 diarrhea, 2 of 3 were on Lead-In (p=.0793).

In Arm A, 35/39 pts were evaluable for response. Overall response rate (ORR) across all DLs was 51% (18/35) and CR rate (CR) was 17% (6/35). PTCL, ORR and CR rates were 55% (12/22) and 27% (6/22) respectively. Among CTCL, ORR was 46% (6/13), no CR. Reponses by histology are detailed in Table 1. Of these responders, 3 proceeded to allogeneic stem cell transplantation (allo SCT) with curative intent. Of note, 4 pts were not evaluable for response, described in Table 1. Median PFS for Arm A (all DL) was 8.8 m (PTCL) and 5.4 m (CTCL). Median follow up was 5.8 m, and median duration of response was 9.1 m.

Of Arm B pts at the MTD, 10/22 (45%) had AEs ≥Gr 3, possibly related to study drug, of these, only neutropenia (n=4, 18%) occurred in ≥10% of pts at the MTD. There was 1 Gr 5 event, Stevens-Johnson syndrome, possibly related to protocol therapy.

In Arm B, the ORR across all DLs was 32% (9/28), the CR rate was 11% (3/28). ORR in PTCL was

36% (5/14), 21% (3/14) achieved CR. ORR in CTCL was 28% (4/14), no CR. Responses by histology are detailed in Table 2. Of these responders, 1 proceeded to allo SCT with curative intent. Median PFS for Arm B (all DL) was 3.5 m (PTCL) and 4.6 m (CTCL). Median follow up was 7.2 m, and median duration of response was 9.3 m.

Conclusion: Duvelisib in combination with romidepsin is highly active in pts with PTCL with tolerable side effects. Duvelisib can be safely combined with romidepsin at a 3-fold higher dose than with bortezomib (75 mg BID vs 25 mg BID) with much lower rate of Gr 3-4 transaminitis than single-agent duvelisib at the same dose. The high response rates and safety of Arm A (Duvelisib + Romidepsin) in PTCL appears to be a potential therapeutic advance and warrants further evaluation in a larger study.

View largeDownload slide

View largeDownload slide

Close modal