Not as BRIGHT as hoped: Discordant adverse events in BRIGHT versus StiL causes stir at ASH 2012
| Dec 11, 2012
Co-authored by Mara Jeffress, PhD, Associate Consultant
The combination of CHOP plus Rituxan (rituximab, Genentech/Roche; R-CHOP) has been the traditional standard of care in the United States and Europe for follicular lymphoma (FL) and mantle cell lymphoma (MCL). That regimen is associated with a range of adverse effects, including cardiac, hematologic and neurologic toxicity, as well as infectious complications. Many physicians feel that CHOP has a role in treating more aggressive diseases such as diffuse large B-cell lymphoma but debate whether such an aggressive regimen is warranted for treatment of patients with indolent lymphomas. Very encouraging data for the two-drug combination of Treanda (bendamustine, Teva Pharmaceuticals) plus Rituxan was first reported at ASH 2009, suggesting that it may be a more effective and less toxic regimen than the standard R-CHOP.
Treanda is approved for use as a monotherapy in relapsed and refractory indolent non-Hodgkin’s lymphoma (iNHL). This combination has been studied in front-line FL in two Phase III trials: the German Study Group Indolent Lymphomas (StiL) NHL 1 study (first reported at ASH 2009) and the U.S. BRIGHT study. An ASCO 2012 update (median follow-up 45 months) of StiL (Rummel, Abstract 3) showed a doubling of the median progression-free survival (PFS) in the R-Treanda arm (69.5 months versus 31.2 months for the R-CHOP group, p<0.0001; HR=0.58). Overall survival did not differ between the two groups, partly because nearly half of the R-CHOP patients whose disease continued to progress were then permitted to cross over to the experimental arm. Both drug combinations had similar overall response rates (92.7% for R-Treanda and 91.3% for R-CHOP), but the R-Treanda combination had significantly more complete responses (39.8% vs. 30.0%, p=0.021) than the R-CHOP combination. In addition, the R-Treanda regimen reported lower rates of alopecia (15% vs. 62%), Grade 3/4 neutropenia (29% vs. 69%), leucopenia (37% vs. 73%), peripheral neuropathy (18% vs. 73%), stomatitis (16% vs. 47%), and infections (31% vs. 41%) and led to less growth factor support overall (4% vs. 20%). Skin rash (42% vs. 23%) and Grade 3/4 lymphocytopenia (74% versus 43%) were the only toxicities that were increased in the R-Treanda arm. In sum, the StiL study showed similar response rates, improved PFS and less toxicity versus R-CHOP and was hailed with much excitement at ASCO. The initial presentation of results from STiL also supported a Category 1 recommendation for its use in newly diagnosed follicular NHL by the National Comprehensive Cancer Network and led to off-label utilization of this regimen in first-line therapy in nearly one-fifth of U.S. patients (Kantar Health, CancerMPact Treatment Architecture, 2012).
The one downside of the StiL trial was that it only enrolled patients from Germany. Thus, Cephalon/Teva initiated the BRIGHT study in March 2009 to confirm the results in U.S. patients and to support regulatory approval by the FDA. The results of this much anticipated follow-on study, comparing efficacy and safety of R-Treanda with the standard treatment regimens of R-CHOP or R-CVP as first-line treatment for indolent FL or MCL, were presented Tuesday morning at ASH (Flinn, Abstract 902). The primary objective of BRIGHT was to demonstrate non-inferiority of complete response (CR) rate of R-Treanda versus standard treatment. Secondary measures included overall response rate (ORR), PFS, and overall survival. While BRIGHT did prove non-inferiority, the results were disappointing compared with the StiL study. The data cutoff for this presentation was June 27, 2012.
Of 447 randomized patients, 436 received treatment (R-Treanda n=221; R-CHOP/R-CVP n=215 (R-CHOP n=99; R-CVP n=116)) and were evaluable for safety. Of these, 419 patients (R-Treanda n=213; R-CHOP/R-CVP n=206 (R-CHOP 97; R-CVP n=109)) were evaluable for efficacy. The arms were well balanced. The CR rate was numerically higher (31%) for R-Treanda than R-CHOP/R-CVP (25%) and statistically non-inferior (p=0.0225); interestingly, R-Treanda was statistically superior (Rate ratio= 1.95; p=0.018) in the subpopulation of MCL patients. The ORR was 94% for R-Treanda and 84% for R-CHOP/R-CVP (90%/83%). Disappointingly, PFS and OS data are immature and were not presented. The speaker, Dr. Flinn, did present the immature PFS data when pressed by the audience. The 30-month curves were overlapping and crossing, a trend that does not bode well for final results.
The adverse event profile in BRIGHT favored Treanda, but the benefit was surprisingly less than was observed in StiL. Neuropathy (all grades: 6% R-Treanda vs. 38% R-CHOP/48% R-CVP), neutropenia (Grade 3/4: 45% vs. 86%/56%) and alopecia (all grades: 4% vs. 41%/21%) were higher in the R-CHOP/R-CVP arms. Nausea (all grades: 63% R-Treanda vs. 58% R-CHOP/39% R-CVP), vomiting (all grades: 27% vs. 13%/13%), lymphopenia (Grade 3/4: 62% vs. 33%/28%), opportunistic infections (all grades: 11% vs. 7%/9%), rash (all grades: 15% vs. 7%/9%), and respiratory disorders (Grade 3/4: 7% vs. 2%/2%) were higher in the R-Treanda arm. Growth factor support was unusually high (compared with the StiL study), with 40% of patients in the R-Treanda and R-CVP arms and 61% of patients in the R-CHOP arm receiving growth factor support. Dose delays were more common for R-Treanda-treated patients (35% vs. 19%), and dose reductions were less common (22% vs. 29%), resulting in a similar dose intensity in all arms. Fatal adverse events occurred in six R-Treanda patients (pneumonia, respiratory failure and sepsis; acute respiratory failure; cardiac arrest; pneumonia; chronic obstructive pulmonary disease; lung cancer) and one R-CHOP/R-CVP patient (sepsis). This concerning result was questioned by the audience and the answer that “deaths were equal across all arms” seemed to side-step this important issue.
Many investigators rushed to the microphone to question Dr. Flinn about the discrepancy in adverse events reported in StiL and BRIGHT. Most concerning are the high rates of infection, nausea and vomiting not previously seen with Treanda. Dr. Flinn did offer some explanations. For example, the definition of infection was broad, including shingles and oral herpes, which might have increased these rates. He had no explanation for the increased nausea and vomiting, saying only that anti-emetic use was equivalent in all arms. But the audience was not satisfied and was left stirred up and buzzing. A similar sentiment followed a separate presentation given by Dr. John Burke on Sunday afternoon (Abstract 155), which showed improved quality of life (QoL) in the R-Treanda arm, but physicians in the audience seemed skeptical, criticizing the lack of quantification provided.
Until more follow-up (and mature PFS and OS) data is provided we will have to remain unsatisfied. For now, in patients with advanced indolent NHL and MCL, R-Treanda produces a CR rate that is non-inferior to that of R-CHOP/R-CVP. Additionally, the adverse event profile of R-Treanda was distinct from that of R-CHOP/R-CVP. While the lower incidence of neuropathy, neutropenia and alopecia are certainly benefits for R-Treanda that both physicians and patients will appreciate, they are countered by the higher rates of other toxicities equally troublesome. The one bright spot is the significantly higher CR rate (51% vs. 24%) in the subgroup of patients with MCL.
So what will happen to this combination going forward? Initially, Cephalon planned to submit a regulatory application for the first-line setting in the U.S. in the first-half of 2011, according to company guidance in late 2010 (Cephalon Q3 earnings call, October 28, 2010). However, it appears that Teva’s acquisition of Cephalon has pushed back the timeline. In February 2012, Teva filed Treanda for approval in the front-line indication and was waiting for FDA response and a Prescription Drug User Fee Act (PDUFA) date (Teva conference presentation, February 27, 2012). The FDA’s answer came on October 23, 2012, asking for more data. The NCCN seems similarly nonplussed, having demoted first-line R-Treanda for FL from Category 1 to Category 2A in its latest guidelines update for NHL (v3.2012). With the results of the StiL study demonstrating clear efficacy and the BRIGHT trial showing such different results with a potential lack of PFS benefit, Treanda approval in the front-line setting is now questionable.