Page 154 - 2019_01-Haematologica-web

R. Gressin et al.
years. With a median follow up of 52 months, the 2-year PFS of the 74 patients with analyzable data was 70%, thus reaching the primary objective of the study which was to improve median PFS by 6 months compared to the report- ed 18-month PFS for patients treated with R-CHOP.3,4 The 4-year PFS (57.6%) observed here for RiBVD-treated patients is in line with the median PFS reported for the BR regimen (35.4 months) and other bortezomib-containing regimens such as 24.7 months for VR-CAP (bortezomib, rituximab, cyclophosphamide, doxorubicin, and pred- nisone) and 26 months for the RiPAD+C regimen (ritux- imab, bortezomib, doxorubicin, dexamethasone, and chlorambucil).7-9 The favorable PFS with the RiBVD regi- men may be related to the marked depth of response (75.7% rate of confirmed and unconfirmed complete responses according to the Cheson 1999 criteria; 78% rate of complete responses according to the Cheson 2007 cri- teria). Although not strictly comparable outside of a ran- domized trial, it is worth noting that complete response rates with other regimens are lower: R-CHOP (34%), BR (40%), VR-CAP (53%) and RiPAD+C (51%).2,3,7,8 The high- er rate of complete responses (confirmed and uncon- firmed) with the RiBVD regimen translated into higher molecular response rates [76% of patients (41/54) were MRD-negative in blood and/or bone marrow at the end of treatment] compared to published data for R-CHOP in older MCL patients [67% MRD negativity (54/81 patients)].20 It is worth noting that 80% of patients in our study had high-risk MIPI scores and that 59% had ≥30% Ki67 positivity (range, 5% to 95% positivity) which appears high compared to the percentages in other studies in MCL patients over 65 years old (see Online Supplementary Table S4).
In keeping with results of high-dose cytarabine treat- ment in younger MCL patients,21-23 a recent phase II trial from an Italian group (FIL) confirmed the efficacy of the R- BAC500 regimen (which associates rituximab, bendamus- tine and cytarabine) for treatment of older MCL patients.12 In updated clinical results for the R-BAC500 regimen, complete response rates of 93% and molecular response rates of 78% (35/45) in the blood have been reported.24 The 2-year PFS and OS were estimated as 83% and 86%, respectively.24 However, it is worth noting that the base- line characteristics of the MCL cohort treated with R- BAC500, the patients in our study and those in other pub- lished series of MCL cases differ quite widely (Online Supplementary Table S4).
The rate of treatment discontinuation in our study was 20% (15/74 patients). This is broadly in line with rates reported for R-CHOP [17% (43/242)] and VR-CAP [18.8% (45/240)] in older MCL patients treated in first line,7 and is lower than that reported with the R-BAC500 regimen (33%)24 (Online Supplementary Table S4). The reported rates of premature therapy cessation for the BR and R-CHOP regimens are 8% and 5%, respectively.10
The 5.4% rate of toxic deaths reported here with the RiBVD regimen is in line with that observed for other first-line regimens used in older MCL patients. For instance, in a phase III study comparing the R-CHOP reg- imen to VR-CAP, Robak and colleagues reported 14/242 deaths (6%) and 11/240 deaths (5%), in the R-CHOP and VR-CAP arms, respectively.7 Of these deaths, a total of six were due to infection and three to cardiac failure.
A trial of lenalidomide, bendamustine and rituximab recently documented a complete response rate of 64%,
with molecular MRD negativity reached in 34% patients. Toxicity, however, was greater than with the RiBVD regi- men, with infectious grade 3-5 toxicities seen in 42% of the 51 recruited patients.13 Two large phase III trials of BR ± ibrutinib or BR ± ACP196 (acalabrutinib) are still ongo- ing and results are pending.
Grade 3-4 hematologic toxicities observed with the RiBVD regimen (51% neutropenia and 35% thrombocy- topenia) were in line with those seen in patients treated with other regimens such as R-CHOP (60% neutropenia and 18% thrombocytopenia), VR-CAP (85% and 57%, respectively) or R-BAC (rituximab, bendamustine and cytarabine; 49% and 52%, respectively).2,7,24
Lymphopenia (65% of grade 3-4), which is known to occur with the BR regimen, may contribute, with neu- tropenia, to the relatively high number of infectious episodes seen in this study.9,10 The rate of lymphopenia at 1 year was 32.5%, which is indicative of longer-term immunosuppression with the RiBVD regimen. Whether this is related to the use of dexamethasone or to the immunosuppressive effects of bendamustine remains to be established but indicates that precautionary measures to control infection are advisable.
Contrary to published results for subcutaneous adminis- tration of bortezomib, we noted a relatively high inci- dence of grade 3-4 neurotoxicity, which is a limiting factor for the RiBVD regimen.25 Although the incidence was comparable to that observed in our previous RiPAD+C trial (18% grade 3-4 toxicity, 7/39 patients), in which bortezomib was administered intravenously at the same dose,8 it is higher than that observed in other studies using comparable intravenous doses in which grade 3-4 toxicity was reported in 7% to 8% of patients treated with R-BV (rituximab, bendamustine, and bortezomib) and VR-CAP regimens.7,11 Further investigations will be required to understand the reason for this. Of note in this respect is the discovery of genetic risk loci for severe peripheral neu- ropathy in European patients with multiple myeloma treated with bortezomib.26
In this study, molecular response in peripheral blood at the end of treatment (after 6 RiBVD cycles) was identified as a major predictive factor for PFS and OS, thus further emphasizing the importance of the depth of response, beyond standard clinical complete response, in MCL.20 Indeed, there was no difference in OS between patients in complete or partial remission at the end of treatment, as defined by the IWG criteria with or without FDG-PET. This finding supports the notion that PET and molecular MRD provide different prognostic information in MCL, probably because they are measuring different types of disease activ- ity, in different disease compartments, with differing sensi- tivities. The maximum standardized uptake value (SUVmax) defined by FDG-PET, also described as an inde- pendent prognosis factor, was not analyzed in our cohort.27
Unexpectedly, neither the MIPI nor Ki67 scores (30% cut-off) had any impact on PFS or OS with the RiBVD reg- imen. This may reflect differences in treatment efficacy by RiBVD in patients with high-risk MIPI scores (70% OS at 36 months), compared to the efficacy of historical treat- ment controls in the original patient cohort that was used to define the high-risk MIPI score (40% OS at 36 months).28
Peripheral blood, but not bone marrow-based MRD status, was highly predictive of PFS and OS in this study (4-year OS of 86.6% for MRD-negative patients com-
144
haematologica | 2019; 104(1)