Rebirth of radiotherapy for elderly patients with diffuse large B-cell lymphoma in the rituximab era

In this issue, Dr. Odejide and colleagues present a clinically relevant analysis comparing standard rituximab, cyclophosphamide, doxorubicin, vincristine and prednisone (R-CHOP) with abbreviated R-CHOP followed by radiation therapy (RT) for elderly patients (≥ 66 years) with stage I/II diffuse large B-cell lymphoma (DLBCL) using the linked Surveillance, Epidemiology and End Results (SEER) and Medicare claims database [1]. Four trials [2–5] have compared chemotherapy followed by RT versus chemotherapy alone in the pre-R era. However, the varying conclusions from these trials allows for considerable debate about the role of radiation in the modern R era.

The authors used the linked SEER–Medicare database to address this issue. This is an appropriate population for analysis given the age distribution of DLBCL [6] and that these linked data provide sufficient clinical information to evaluate the administration of chemotherapy, R and radiation. The authors report the largest dataset examined with a systematic approach to address this question for elderly patients. The major threats to the validity of these findings are the availability of claims data to serve as a surrogate for clinical care and outcomes, the retrospective dataset used to assemble the cohort and the non-randomized nature of observational data. The authors reasonably address each of these threats in their careful use of claims data, the multiple variable regression models and the propensity score analyses to confirm results. Nevertheless, only data from randomized controlled trials can decisively address the optimal approach for patients with limited stage DLBCL treated with R-CHOP.

For the overall cohort where treatment was ascertained from the claims data (n = 4322), concordance between the SEER radiation therapy variable and the authors’ claims-defined radiation therapy variable was 87%. This is consistent with prior reports of concordance between these two variables in SEER–Medicare [7]. The authors also describe the baseline characteristics of patients treated with these two approaches and examine predictors of abbreviated R-CHOP + RT, providing insight into practice patterns throughout the United States.

One challenge for this data source is an incomplete handling of bulky disease. The authors conducted sensitivity analyses comparing outcomes of interest (overall survival [OS], time to second-line therapy and adverse events) using a cohort of patients with stage I disease, thus eliminating patients with bulky stage II disease. The findings of these sensitivity analyses were consistent with the results of the entire cohort, affirming the validity of these results. Another challenge of this data source is that the definition of relapse was based upon encounters that had Medicare claims for subsequent chemotherapy. This approach has previously been applied in SEER–Medicare [8], and the authors also utilized an institutional lymphoma dataset to verify receipt of second-line therapy as a reasonable surrogate for relapsed disease.

To compare the options for managing patients with stage I/II DLBCL, the authors identified patients who received 3–4 cycles of R-CHOP + RT and patients who received 6– 8 cycles of R-CHOP. This provides pure cohorts for comparison, similar to per protocol analyses of clinical trials data. Other patients not included in these analyses may have received fewer or more cycles of R-CHOP than was initially intended due to patient preference, toxicity, progression or other reasons, but were excluded from analyses. Obviously, this retrospective analysis could not reconstruct the treatment intent of the provider, but as the authors note, this limitation should be considered in applying these results to clinical decision-making. Despite these limitations, this study provides the most comprehensive comparison to date in elderly patients of abbreviated R-CHOP with RT and standard R-CHOP.

In addition to the authors’ work, other limited data in the R era evaluating the role of abbreviated chemotherapy followed by RT have also been published [9–12]. In Southwest Oncology Group (SWOG) 0014, 60 patients with limited-stage aggressive non-Hodgkin lymphoma (NHL) and at least one adverse risk factor were treated with four doses of R and three cycles of CHOP, followed by 40–46 Gy of involved-field radiotherapy (IFRT) [9]. With a median follow-up of 5.3 years, treatment resulted in a progression-free survival (PFS) of 88% and an overall survival (OS) of 92% at 4 years, compared to a PFS of 78% and OS of 88% at 4 years in SWOG 8736. While these results suggest that abbreviated R-CHOP + RT is at least as effective as a full course of chemotherapy and associated with lower rates of early relapse, the follow-up results of SWOG 8736, showing that there were no longer differences in OS between the two treatment arms largely due to late relapses and lymphoma deaths in patients who received abbreviated CHOP + RT, need to be considered before discarding the potential role for more cycles of therapy [4].

The RICOVER-60 RT study evaluated R-CHOP × 6 followed by observation or the addition of RT in patients who were 61–80 years old with DLBCL of any stage and bulky (≥ 7.5 cm) or extranodal disease [13]. When analyzed per protocol, statistically significant improvements in event-free survival (EFS; 80% vs. 54%), PFS (88% vs. 62%) and OS (90% vs. 65%) were noted in favor of the radiation arm, but when examined in a multivariable intent-to-treat analysis only EFS differences were observed. The UNFOLDER trial involving patients 18–60 years old with bulky/extranodal aggressive NHL of all stages also demonstrated efficacy of RT with R-CHOP in an early interim analysis [14]. A pooled subset analysis from multiple randomized trials examining the role of RT in skeletal involvement also showed an improvement in EFS with radiation [15]. While these studies lend support to the notion that abbreviated R-CHOP and radiation may be an acceptable alternative for patients with limited stage DLBCL, none of these definitively address the optimal therapy for this population.

Despite the limitations of this retrospective analysis, the authors demonstrate that aggressive therapy for elderly patients with limited stage DLBCL results in excellent OS, and that an abbreviated course of chemotherapy followed by RT is associated with decreased risk of relapse and lower toxicity. The authors’ work adds value to the existing debate, and is in line with the recently published data. Fit elderly patients should be considered for aggressive, curative therapy, and many with limited stage disease can defer a full course of chemotherapy in exchange for abbreviated therapy followed by RT. Additional studies are needed to delineate the population best suited for this approach.

Potential conflict of interest

Disclosure forms provided by the authors are available with the full text of this article at www.informahealthcare.com/lal.