PF309 IMPACT OF LATE RELAPSES ON OUTCOMES IN PATIENTS WITH DIFFUSE LARGE B‐CELL LYMPHOMA (DLBCL)

Background: The advent of rituximab has improved the prognosis of patients with DLBCL. Nevertheless, up to 30–40% of these patients will be refractory or will relapse after treatment. Late relapses (LR), arbitrarily defined as those occurring 24 months after diagnosis, are rare and have been linked to better prognosis in some studies. In addition, it is unclear whether these LR are eventually related to the original lymphoma or they represent a de novo lymphoma. Aims: To review clinical and biological characteristics of patients who received immunochemotherapy and then relapsed 24 months after diagnosis. Methods: All patients diagnosed with DLBCL in two Spanish hospitals from 2003 to 2016 were included. Amongst which, those who relapsed after having achieved complete remission with front‐line immunochemotherapy were identified and analyzed retrospectively for clinical characteristics. Furthermore, clonal relationship between the initial and relapsed lymphoma was analyzed by matching IGH gene rearrangements. Results: Ninety‐seven of 853 patients (11%) relapsed, out of which 42 (43%) were LR. The median age at diagnosis of LR patients was 66 years (range, 23–86). At diagnosis, patients with LR more frequently had a low or intermediate international prognosis index (n = 23; 58%) (p = 0.008), normal serum LDH (n = 14; 40%) (p = 0.01), and lack of bone marrow infiltration (n = 37; 88%) (p = 0.017) compared to early relapses. During the initial diagnostic work‐up, 44% (12/27) of LR patients had a germinal centre (GC) cell of origin phenotype determined by the Hans algorithm. When patients were reassessed at relapse, GC phenotype was found in 69% (n = 16) of LR patients and 30% (n = 7) of early relapse patients. In 8 out of 9 LR patients the clonal relationship was established between diagnosis and relapse. In the LR subgroup, median time from diagnosis to relapse was 4.4 years (range 2–11) and 17 patients (17%) relapsed 5 years from the diagnosis. Intensive salvage therapy was given to 20 (47%) LR patients, and 15 (35%) subsequently underwent ASCT. Of note, LR patients achieved a higher rate of complete remissions (n = 24; 62%) after salvage therapy than early relapse patients (n = 22; 41%) (p = 0.047). Median follow‐up was of 27 months (range, 6 months‐12 years) for the whole series. Event‐free survival at 2 years for LR patients and early relapse patients was 38% [95% CI, 23–54%] and 31% [95% CI, 19–44%] (p = 0.3) respectively, while 2‐year overall survival (OS) from relapse was 57% [95% CI, 39–71%] and 42% [95% CI, 29–55%] respectively (p = 0.36). However, better 2‐year OS from relapse was observed between the LR and early relapse subgroups in the cohort of patients younger than 70 at the time of relapse (p = 0.043) (figure 1). Summary/Conclusion: Patients with LR show favourable prognostic markers at diagnosis and respond better to salvage therapy. In our series, no differences in 2‐year OS from relapse, between patients with LR or early relapse were observed. Conversely, it appears that in the subgroup of young patients the overall survival is better in late relapsers.