Abstract 2793 Background: DLBCL is the most frequent aggressive non Hodgkin9s lymphoma in adults. The International Prognostic Index (IPI) is currently the most used tool to identify different risk patients. The role of an interim PET to early identify patients with bad response to chemotherapy is still controversial. Aims: The aim of this analysis is to evaluate the predictive value for event-free survival of an interim PET (after 2 chemotherapy cycles) and at the end of therapy (after 6 chemotherapy cycles) of dose-dense R-CHOP in patients with DLBCL. Methods: This is a prospective clinical trial for patients with DLBCL older than 65 with IPI 0–5 or younger than 65 with IPI 0–2. Treatment consists on 6 cycles of R-CHOP administered every 14 days followed by pegfilgrastim (6 mg per cycle) on day 2. In this sub-analysis we have included 69 patients who completed the 6 cycles of chemotherapy and who have PET evaluation at diagnosis, after 2 cycles of R-CHOP and at the end of treatment. All evaluations were made by combined PET/CT, except for 14 patients who were evaluated by PET alone. PET was determined to be positive or negative based on the local radiology reports. Interim PET results did not change the planned treatment. Results: 124 patients were included in the trial, 19 (15.3%) did not finish 6 cycles of treatment: 2 due to serious adverse events, 2 due to progression disease/stabilization, 7 due to investigators decision and 8 due to death. Over 105 patients who completed 6 cycles of treatment, 69 patients with complete PET evaluation (at diagnosis, after 2 cycles and at the end of treatment) were included in the analysis. Median age was 61.6 years old (range 18.2–82.8), 34 (49.3%) were older than 65 yo, 37 (53.6%) were male, 59 (85.5%) had ECOG 0–1. Characteristics of the disease at diagnosis were as follows: stage III-IV: 45 (65.2%), bulky disease: 17 (24.64%), >2 extra-nodal sites involvement: 4 (5.8%), B symptoms: 18 (26.1%), elevated LDH: 36/67 (53.7%), elevated beta-2-microglobulin: 23/62 (37.1%), IPI 3–5: 23 (33.3%). The median relative dose intensity (RDI) for cyclophosphamide and adriamycin calculated according to 3-week interval as a reference was 139.8% (≥130% in 75.2% of patients): 141.6% in younger patients and 138.6% in the elderly. Thirty-five (50.7%) patients achieved a negative PET evaluation after 2 R-CHOP cycles and 57 (82.6%) at the end of therapy. Patients with bulky disease had a positive interim PET more frequently (64.7% vs 44.2%), but most of them turned negative at the end of treatment (positive end-of-treatment PET for bulky disease: 11.7%). With a median follow-up of 28 months (limits 2.3–49), event-free survival (EFS) was 70.6% for patients with positive interim PET and 92.9% for patients with negative interim PET (p=0.14). The negative predictive value (NPV) of a negative interim PET was 94.3% and the positive predictive value (PPV) of a positive interim PET was 14.7%. EFS was 47.6% for patients with positive end-of-treatment PET and 95.6% for patients with negative end-of-treatment PET (p Conclusions: A negative interim PET is highly predictive of a good outcome. Patients with a positive interim PET have worse EFS (not statistically significant), but their PPV is 14.7%. Patients with bulky disease turn to a negative PET later. Although longer follow-up is needed, these results do not support early intensification of patients with DLBCL based only in a positive interim PET and, probably, other factors should be taken into account. Disclosures: No relevant conflicts of interest to declare.