The Hematologist

September-October 2019, Volume 16, Issue 5

Surrogate Endpoints for Cure in DLBCL: Are We There Yet?

Caron A. Jacobson, MD Instructor in Medicine
Dana-Farber Cancer Institute, Harvard Medical School, Boston, MA

Published on: July 25, 2019

Wang Y, Farooq U, Link BK, et al. Late relapses in patients with diffuse large b-cell lymphoma treated with immunochemotherapy. J Clin Oncol. 2019; doi: 10.1200/JCO.19.00014. [Epub ahead of print].

Diffuse large B-cell lymphoma (DLBCL) is regarded as a potentially curable lymphoma with chemoimmunotherapy. Studies have demonstrated that patients who have not relapsed by month 24 following diagnosis (EFS24) have a five-year overall survival (OS) similar to that of age-matched controls; therefore, this has been considered a marker of likely-cured disease.Most relapses occur in the first two years following diagnosis, and little is known about risks of relapse at later time points in the era of rituximab-based chemoimmunotherapy. Dr. Yucai Wang and colleagues performed a multicenter, prospective, observational trial of patients with DLBCL treated with rituximab-containing chemoimmunotherapy at the Mayo Clinic and the University of Iowa between 2002 and 2015. Double- and triple-hit lymphomas were not included in the study, but transformation from an indolent lymphoma was. Patients who were alive and disease-free at 24 months from diagnosis were analyzed for relapse rate.

Of 1,324 patients diagnosed and treated during this timeframe, 847 (64%) were alive and without evidence of disease relapse at EFS24, with a median follow-up of 83.2 months and 62.9 months from diagnosis and EFS24, respectively. Most of the initial cohort of patients had DLBCL alone (87%), whereas 13 percent had transformation from an indolent histology. More than half of the patients had advanced-stage disease (57.3%). Of the EFS24 patients, the cumulative risk of relapse at three, five, and eight years from EFS24 was 6.9 percent, 9.3 percent, and 10.3 percent, respectively. Among late relapses, a slight majority were men (56.4%), just under half had an elevated lactate dehydrogenase (LDH) at diagnosis (49.3%), and most had advanced-stage disease (75.6%), with 16.7 percent of patients having involvement of more than one extranodal site; the International Prognostic Index (IPI) was intermediate-high risk or high risk in 30.8 percent of patients. Approximately 30 percent of patients had a history of transformed lymphoma; the rest had DLBCL alone.

Among patients who had a late relapse following an initial diagnosis of DLBCL alone, most relapsed with DLBCL (73.5%), whereas patients with a history of transformed lymphoma were slightly more likely to relapse with an indolent lymphoma compared with DLBCL (55.0% vs. 45.0%). Overall, however, the cumulative risk of DLBCL relapse over time was greater than that of an indolent lymphoma relapse. While most patients who relapsed late had a germinal center B-cell (GCB) –like lymphoma, GCB lymphomas were associated with an increased risk of relapse of an indolent lymphoma, and not DLBCL, over the activated B-cell (ABC) –like subtype.

Salvage therapies and outcomes following salvage therapy were also analyzed among late-relapsing patients. Patients relapsing with DLBCL most commonly received a platinum or cytarabine-containing salvage regimen (44%), and 40 percent had a consolidation autologous stem cell transplantation (ASCT). In patients who relapsed with an indolent lymphoma, the most common second-line therapy was bendamustine-rituximab (37.5%), and approximately 20 percent of patients were consolidated with an ASCT. For all patients, the median postrelapse survival was 38.9 months; relapse with DLBCL was associated with a shorter survival than relapse with an indolent lymphoma (29.9 months vs. not reached). Following ASCT, median post-transplantation survival was 2.2 years.

These results suggest that although EFS24 is associated with an excellent outcome for most patients with DLBCL treated with chemoimmunotherapy in the frontline, a subset of patients is at risk for a delayed relapse. Survival in this group, especially those who relapse with DLBCL, is poor. It is not surprising that patients with transformed lymphomas would relapse late with an indolent lymphoma given the natural history of indolent lymphomas. However, the incidence of late relapses of indolent lymphoma in patients previously thought to have de novo DLBCL of the GCB subtype suggests either that these patients had a transformed indolent lymphoma at initial diagnosis or they have a genetic predisposition that led to the development of both lymphomas. These results have implications both for clinical investigation and clinical practice. Regarding clinical investigation for upfront therapy for DLBCL, endpoints focused on EFS24 may be inadequate, and longer follow-up may be necessary to draw meaningful conclusions. Furthermore, identifying patients at increased risk for late relapse, as well as acknowledging their poor OS, may help identify a group of patients who would benefit from alternative initial or salvage treatment strategies to be tested in future clinical trials. From a clinical practice standpoint, results such as these belie the importance of setting realistic expectations, both in terms of risk of late relapse of DLBCL, but also late relapse of an indolent lymphoma in patients with a history of GCB DLBCL. As a field, we should avoid using the “cure” word flippantly and focus instead on durability of response. The identification of this group of patients at risk for late relapse will hopefully spawn an investigation into clinical and pathologic biomarkers that can help us identify them at diagnosis.


  1. Maurer MJ, Habermann TM, Shi Q, et al. Progression-free survival at 24 months (PFS24) and subsequent outcome for patients with diffuse large B-cell lymphoma (DLBCL) enrolled on randomized clinical trials. Ann Oncol. 2018;29:1822-1827.

Conflict of Interests

Dr. Jacobson indicated no relevant conflicts of interest. back to top