Abstract

INTRODUCTION:

Stem-cell transplantation (SCT) is standard treatment of relapsed lymphomas. Patients in refractory relapse or relapsed after autoSCT are often considered for an alloSCT, aiming at a graft-vs.-lymphoma (GVL) effect. While reduced intensity conditioning regimens have substantially contributed safety to alloSCT, aggressive lymphomas often progress rapidly afterwards, partly due to tumor cell growth outpacing GVL. Outcomes are particularly poor for patients with active disease at the time of alloSCT. Thus, development of more efficacious conditioning regimens for aggressive lymphomas, avoiding mucotoxic drugs (such as melphalan), is a major need. We previously showed that the combination of 2 nucleoside analogues [gemcitabine (GEM) + clofarabine (CLO)] followed by busulfan (BU) is highly synergistic against B- and T-cell lymphoma cell lines, through inhibition of DNA damage repair. Those observations led us to clinically study a novel high-dose regimen of GEM/CLO/BU with alloSCT for aggressive lymphomas.

METHODS:

Eligibility included age 12-65; refractory DLBCL, T-cell and Hodgkin lymphomas; acceptable end-organ function, and an 8/8 HLA-matched (A, B, C and DRB1) sibling (MSD) or unrelated donor (MUD). GEM was escalated from dose level 1 to 3; it was given on days -6 (d-6) and d-4 as a loading dose of 75 mg/m2 followed by infusion at 10 mg/m2/min over 40-90 min (total daily dose 475, 675 or 975 mg/m2), to be followed by CLO (40 mg/m2/day x 4, d-6 to d-3) and BU (pharmacokinetically targeted to an average daily AUC of 4,000 μMol•min x 4, d-6 to d-3). AlloSCT was on d0. Patients receiving a MUD transplant received rabbit ATG (0.5, 1.5 and 2 mg/Kg on d-3, d-2, and d-1, respectively). Patients with CD20+ tumors received rituximab (375 mg/m2) on d-14, d-7, d+1 and d+8. The GVHD prophylaxis included tacrolimus (tapered after d+100) and mycophenolate mofetil (from d0 to d+60). Dose limiting toxicities (DLT) were defined as any G4-5 nonhematological organ toxicity, or as G3 skin or G3 mucositis lasting >3 days at peak severity. Dose escalation of gemcitabine followed a Bayesian design targeting a maximal DLT probability of 25%. Non-relapse mortality (NRM), relapse incidence, G II-IV or III-IV acute GVHD or chronic GVHD were calculated with the cumulative incidence method with consideration of competing risks.

RESULTS:

Between 11/12 and 06/17, 48 patients were enrolled: 25 DLBCL (5 double hit), 13 T-NHL and 10 HL (Table); 27 MSD, 20 MUD.All patients received PBPC. The MTD of gemcitabine was level 3. DLT was observed at the following frequencies: level 1: 12.5% (1/8), level 2: 9% (2/9), and level 3: 25% (4/16). Regimen-related toxicities were manageable, including mucositis (47% G2, 42% G3), self-limited transaminitis (21% G2, 49% G3), hyperbilirubinemia (no cases of VOD) (30% G2, 19% G3), and dermatitis (13% G2). There were no cardiac, pulmonary, renal or CNS toxicities. Neutrophils dropped to <0.1 in all patients on median d0 (d-3 to +2), and engrafted on median d+10 (d+9 to +20). Platelets engrafted on median d+12 (d+8 to +38).

All patients engrafted; full donor chimerism was established early: a median of 100% (50-100%) on d+28, and 100% (98%-100%) on d+100. The rates of G II-IV and III-IV acute GVHD were 29.8% (35% MSD, 25% MUD) and 10.6% (19% MSD, 0% MUD), respectively. The rates of moderate-severe / severe chronic GVHD were 8.5% / 6.4%. 1/3 HL pts with prior exposure to anti-PD1 developed steroid-responsive hepatic chronic GVHD.

The γ-H2AX levels in PBMNC (N=19) increased throughout treatment, to a median of 8.5 fold from baseline (range, 0.4-46.3) (P=0.0001) and PARP-1 decreased to 0.45 fold (0.2-6.3) (P=0.4), indicative of activation of DNA damage response and apoptosis, respectively. The overall response rates on d+30 (in 34 patients with measurable disease) are 77% (DLBCL), 100% (T-NHL), and 75% (HL). Their respective CR rates are 69%, 100% and 75%. The d+100 / 1-year NRM rates were 6.25% / 14.5%. NRM causes were infection (6) and GVHD (1). At median follow-up of 18 months (1-53) the EFS / OS rates of the entire population are 44% / 56%. Outcomes appear especially promising in the DLBCL (EFS / OS, 54% / 54%) and the T-NHL (62% / 69%) subgroups.

CONCLUSIONS:

GEM/CLO/BU is a new, myeloablative, reduced-toxicity alloSCT regimen with high activity in this population of heavily pretreated and refractory aggressive lymphomas, which warrants further study in disease-specific trials

Disclosures

No relevant conflicts of interest to declare.

Author notes

*

Asterisk with author names denotes non-ASH members.