Dose-Dense and Dose-Intense Treatment Strategies
Although poor-risk DLBCL can be defined in a number of ways (as noted above), there clearly exists a population of patients who experience unfavorable outcomes with standard therapy. In an attempt to improve on the results seen, investigators have explored dose-dense and dose-intense strategies and consolidation with high-dose therapy (HDT) followed by rescue with ASCT to reduce relapse rates. Some early phase II studies in the US and randomized phase III studies in Europe suggested that better outcomes could be achieved with this approach compared with standard CHOP repeated every 21 days (CHOP-21).[34-38] However, nearly two decades ago, the US Intergroup study demonstrated that CHOP was associated with an OS similar to that of more intensive (and more complicated) regimens and was better tolerated, thus establishing CHOP as the standard of care. Nevertheless, outcomes remained disappointing for some patients, with 30% of patients being refractory or progressing shortly after induction, and patients experiencing 30% to 40% OS at 5 years.
In selected settings, the addition of etoposide(Drug information on etoposide) to CHOP, shortening the administration of CHOP to a biweekly interval (CHOP-14), and use of a more intensified anthracycline regimen,[36,41] did demonstrate advantages over standard CHOP, but these approaches were not broadly adopted.
A major development in the treatment of DLBCL occurred when rituximab(Drug information on rituximab) was assessed in phase II studies of aggressive B-cell lymphomas.[42,43] Several groups launched phase III studies in different DLBCL patient subsets to investigate the impact of rituximab.[2,4,44] One of the earliest reports of survival benefit was from the Groupe d’Étude des Lymphomas de l’Adulte (GELA), showing that DLBCL patients over 60 years of age who were treated with R-CHOP had superior outcomes compared with patients treated with CHOP. With more mature results, OS, EFS, and PFS remained superior with the addition of rituximab, and no significant increase in adverse events was reported. Given the improvement in outcomes with dose-dense and dose-intense strategies prior to the use of rituximab in this setting, studies followed that investigated whether the addition of rituximab could extend the benefits described.
One such trial, published by Economopoulos et al, began enrollment in 1999 and randomized patients to CEOP (cyclophosphamide, epirubicin(Drug information on epirubicin), vincristine, and prednisone(Drug information on prednisone)) every 2 weeks (CEOP-14) vs a standard 3-week cycle (CEOP-21). All patients enrolled from May 2002 onward received rituximab with each chemotherapy cycle, and patients achieving a complete response (CR) received rituximab consolidation. Response rates (CR rate, 51% CEOP-14 vs 53% CEOP-21, P = .786) and survival were similar between the two arms, demonstrating no difference with the dose-dense/intense approach. While the addition of rituximab to both the 14-day and the 21-day regimens improved time to progression and OS, no difference emerged between the arms. A dose-dense strategy also was explored in the RICOVER-60 trial. Based on prior success with CHOP-14, the German High Grade Non-Hodgkin Lymphoma Study Group (DSHNHL) randomized elderly patients (61 to 80 years) to receive 6 or 8 cycles of CHOP-14, with or without rituximab. Six cycles of R-CHOP-14 significantly improved 3-year EFS (66.5% vs 47.2%), PFS (73.4% vs 56.9%), and OS (78.1% vs 67.7%) compared to 6 cycles of CHOP-14, and there was no benefit from extending therapy to 8 cycles of R-CHOP-14 (compared to 6 cycles of R-CHOP-14). A significant challenge in interpreting the results of this study was the lack of comparison with R-CHOP-21, a standard approach in the US and many other countries.
Dr. David Cunningham, on behalf of the United Kingdom National Cancer Research Institute Lymphoma Clinical Study Group, presented results from a randomized phase III study comparing R-CHOP-14 to the standard R-CHOP-21 at the American Society of Clinical Oncology (ASCO) 2011 conference. Between March 2005 and November 2008, a total of 1080 treatment-naive patients were randomized to either 8 cycles of standard R-CHOP-21 or 6 cycles of R-CHOP-14 plus granulocyte colony-stimulating factor (G-CSF), with an additional 2 cycles of single-agent rituximab (as was given in RICOVER-60). This study was designed to detect an 8% improvement in 2-year OS for the superior regimen. The trial arms were balanced with respect to patient age, presence of B symptoms, bulky disease, stage, and IPI, and 52% of patients were 60 years of age or older. Overall response rates (ORRs) were similar between the two arms: CR or CRu (complete response, unconfirmed) was 63% with R-CHOP-21 vs 58% with R-CHOP-14 (P = .15). After a median follow-up of 37 months, PFS (P = .98) and OS (P = .78) were similar, and subgroup analyses failed to identify a subset of patients who benefited from the accelerated R-CHOP regimen. Grade 3/4 nonhematologic toxicities were comparable in the two arms. Among patients receiving the 21-day regimen, however, there was a significantly greater frequency of grade 3/4 neutropenia (77% vs 37%) and febrile neutropenia (11% vs 5%), respectively (P < .01 for both), compared with the 14-day regimen. The lower frequency of neutropenia in the R-CHOP-14 arm is likely a reflection of prophylaxis with G-CSF administered to all patients in this group. In addition, there were higher incidences of thrombocytopenia (5% vs 9%) and anemia (1% vs 3%) in the R-CHOP-14 arm, likely a consequence of treatment intensity. The authors concluded that based on these data, there is no evidence to support a shift from the standard R-CHOP-21 to an accelerated R-CHOP strategy.
Similar findings were presented in 2009 at the annual American Society of Hematology (ASH) convention by the GELA, which described the results of the planned interim analysis of the LNH03-6B, an open-label, randomized trial evaluating the efficacy of R-CHOP-14 compared to R-CHOP- 21 in elderly patients (median age, 72; range, 60–80 years). Patients were randomized to 8 cycles of R-CHOP-14 or R-CHOP-21, with a second randomization to prophylactic darbepoietin alfa vs conventional treatment for chemotherapy-induced anemia. A secondary endpoint of the study was to examine whether maintaining a hemoglobin level of 13 g/dL renders patients better able to tolerate the dose-dense regimen.
At baseline, more patients in the R-CHOP-14 arm had an age-adjusted IPI (aaIPI) score of 2 to 3 (67% vs 59%), and more patients in the R-CHOP-21 arm presented with B symptoms (43% vs 37%); otherwise, patient characteristics were similar in the two groups. At 2 years, there was no significant difference in ORR (81% with R-CHOP-14 vs 84% with R-CHOP-21), PFS (49% vs 63%), or OS (67% vs 70%). Grade 3/4 hematological toxicity was more frequent in the R-CHOP-14 group, similar to the United Kingdom trial. Unlike the United Kingdom trial, white blood cell growth factor support was not routinely required for patients who received R-CHOP-14, raising some questions about R-CHOP-14 administration in this trial. Taken together, these results indicate no particular benefit to administering R-CHOP-14 rather than R-CHOP-21 for older or younger patients with DLBCL.
An alternative dose-intensive strategy for treatment-naive patients with DLBCL is based on trials conducted prior to the rituximab era that demonstrated superiority of the intensive regimen ACVBP (doxorubicin [Adriamycin], cyclophosphamide(Drug information on cyclophosphamide), vindesine(Drug information on vindesine), bleomycin(Drug information on bleomycin), and prednisone) over standard CHOP.[36,41] In 2003, the GELA initiated a multicenter, phase III, open-label, randomized trial comparing the efficacy and safety of R-ACVBP vs R-CHOP in younger patients with an aaIPI of 1. Compared to R-CHOP-21, the R-ACVBP regimen produced superior 3-year EFS (81% vs 67%, P = .0035) and OS (92% vs 84%, P = .0071). However, R-ACVBP was associated with substantially more hematologic and nonhematologic toxicity, and this regimen incorporates vindesine, an agent that is not widely available in the US and therefore has not been broadly adopted. Another intensive regimen developed by investigators at the National Cancer Institute (NCI) focused on pharmacokinetics to overcome drug resistance in DLBCL, using dose-adjusted (DA) etoposide, prednisone, vincristine, cyclophosphamide, doxorubicin(Drug information on doxorubicin) (EPOCH), and rituximab (DA-EPOCH-R). The doxorubicin, etoposide, and vincristine in this regimen are infused over a period of 96 hours and are pharmacodynamically dose-adjusted based on the neutrophil level nadir. In a phase II trial of 72 previously untreated patients with DLBCL, this approach demonstrated high 5-year PFS and OS. The Cancer and Leukemia Group B (CALGB) reproduced the NCI results in a multi-institutional phase II study; 5-year OS rates were 84%, and the regimen was tolerated without significant grade 4 nonhematologic toxicities. An ongoing phase III trial (CALGB 50303) comparing R-CHOP vs DA-EPOCH-R in patients with newly diagnosed DLBCL (examining outcomes in ABC and GCB subtypes) will provide a systematic comparison of these two regimens.
Another approach to intensifying the initial treatment for DLBCL involves consolidation following induction with HDT and ASCT. The role of ASCT in frontline therapy for DLBCL has been unclear, particularly in the era of immunochemotherapy. While consolidation with HDT and ASCT has been shown to reduce relapse rate in DLBCL in some settings,[38,51] first-line ASCT has been challenged by a meta-analysis published in 2008, in which a total of 3079 patients from 15 randomized trials were included. Despite better rates of CR, the EFS and OS were the same regardless of whether patients were treated with conventional chemotherapy or HDT followed by ASCT.
Additional data from the results of the SWOG-led Intergroup phase III trial S9704 were presented at the 2011 ASCO conference. This trial investigated the benefits of autologous transplant in first remission in patients with advanced-stage diffuse aggressive NHL with high-intermediate/high IPI score, following 5 cycles of CHOP with or without rituximab (CHOP±R). Between September 1997 and December 2007, a total of 370 patients were enrolled (median age, 51 years); 59% were male, 62% had stage IV disease, 85% had an elevated LDH level, and 32% had a high IPI score after adjusting for age. Of those eligible, 253 patients were randomized following standard therapy to CHOP±R × 3 additional cycles (n = 128) or transplant (n = 125) involving CHOP±R × 1 followed by ASCT using TBI- or BCNU-based conditioning regimens. Initial results demonstrated that the addition of ASCT resulted in a significantly higher PFS at 2 years (69% vs 56% [hazard ratio (HR) 1.72, 1.18–2.51]; P = .005), but no difference in 2-year OS was seen (74% vs 71%, [HR 1.24, 0.81–1.91]; P = .32). Possible confounders to the OS data included 18% of patients in the standard arm who underwent salvage therapy with ASCT at relapse and who were alive and disease-free, and the prolonged enrollment to the trial, during which time the type and number of supportive care medications for ASCT changed considerably.
Exploratory subset analyses showed that the benefit of ASCT was observed primarily in patients with high-IPI disease, in whom 2-year OS rates were 82% (ASCT) vs 64%. Although the study was not powered to detect differences in treatment after stratifying for IPI risk, based on the results of this trial, future studies should examine the benefits of consolidation with autotransplant for young patients with high-IPI DLBCL following treatment with R-CHOP-21. A DSHNHL study of HDT with rituximab for primary treatment of younger patients with high-risk DLBCL also demonstrated high 3-year OS (78.7%) and EFS (72.7%), suggesting that further investigation is needed in this population. 
Table 1 compares the outcomes of various up-front intensified regimens for DLBCL.
Over the past 20 years, outcomes for patients with DLBCL have significantly improved as a result of the addition of rituximab to combination chemotherapy and improvement in salvage therapy and supportive care. However, the existing applications of dose-dense strategies and dose-intensification have limited evidence to support their use in the frontline setting in the rituximab era. As discussed, head-to-head comparisons of R-CHOP-14 with standard R-CHOP-21 failed to demonstrate benefits for the accelerated approach when rituximab was included with chemotherapy, nor could the investigators identify a particular subgroup of patients for whom R-CHOP-14 was of benefit. Moreover, strategies involving additional rituximab, such as maintenance therapy, have not been shown to improve outcomes in DLBCL and are not currently recommended.[2,56] Although other intensification approaches such as DA-EPOCH-R hold promise, R-CHOP-21 remains the current frontline regimen in the US. Unfortunately, with this approach, approximately half of patients treated will progress or relapse. For young patients with high-risk IPI, consolidation with HDT and ASCT should be considered an approach worthy of investigation, given the results of the SWOG trial subset analysis that indicated that the benefit of consolidation with ASCT in first remission was observed primarily in this subgroup. The NCCN guidelines panel recommends initial treatment with 6 cycles of R-CHOP-21 for patients with advanced-stage disease; however, other comparable anthracycline-based regimens may also be acceptable in certain circumstances (see Table 2). Alternate options suggested by the panel include dose-dense R-CHOP-14 or DA-EPOCH-R, in addition to participation in a clinical trial with new regimens when available. The recently presented data from the trials discussed above suggest that there is no specific clinical scenario in which R-CHOP-14 would be preferred, and that this regimen has additional toxicities. We recommend that eligible patients be enrolled in CALGB 50303 to assess the merits of DA-EPOCH-R compared to R-CHOP.
Since modifying traditional approaches has not provided the hoped-for therapeutic benefit, what can be said about novel approaches? First, can we be more effective at risk stratification? Risk stratification based on clinical features, such as IPI, is useful particularly in standardizing patients enrolled in clinical trials, but it may be less promising than biologic risk stratification when making treatment decisions. To ensure that patients are informed, treatment decision-making should include discussions about the significance of the cell of origin and projected outcomes with standard induction therapy, based on molecular subtype. Given the rather dismal outlook for DLBCL patients with the ABC subtype, enrollment in a clinical trial should be strongly encouraged. Applying the knowledge we have gained from GEP, therapeutic targeting of molecular pathways may ultimately lead to improvement in outcomes in DLBCL. Several novel agents exploiting this strategy are undergoing evaluation, both as single agents in the relapsed-disease setting and in combination with R-CHOP. Some examples include immunomodulatory drugs (IMiDs), spleen tyrosine kinse (Syk) inhibitors, Bruton's tyrosine kinase inhibitors, protein kinase C inhibitors, histone deacetylase inhibitors, proteasome inhibitors, anti-survivin agents, mTOR (mammalian target of rapamycin) inhibitors, and alternative antibody therapies. With the expanding information on the pathophysiology of and biomarkers for DLCBL, future studies should build upon these biologic underpinnings with rational study design and novel therapeutics. Such approaches will provide key opportunities for further advances in the treatment of DLBCL, given that chemotherapy intensification appears to provide limited additional benefits over the current standard of care.
Financial Disclosure: Dr. Flowers has served as a consultant for Spectrum, Celgene, OptumRx, Seattle Genetics, Allos, Genentech/Roche (unpaid), and Millennium/Takeda (unpaid). Dr. Flowers leads research studies that are supported by Spectrum, Novartis, Millennium/Takeda, and Gilead.