Introduction: The predictive value of 18-fluorodeoxyglucose PET performed after a few cycles of chemotherapy has been questioned in aggressive lymphomas. Inconsistent study results, however, may be due to procedural differences rather than an inability of the method to predict outcome. Whether changing treatment in pts. with an unfavorable interim PET (iPET) improves outcome, has not been determined in a randomized study. The PETAL trial (EudraCT 2006-001641-33, NCT00554164) was initiated to resolve these issues.

Methods: Pts. aged 18 to 80 yrs. with newly diagnosed aggressive lymphomas and a positive baseline PET received 2 cycles of rituximab (R), cyclophosphamide (C), doxorubicin (H), vincristine (O) and prednisone (CHOP) followed by iPET. The conditions of iPET were strictly defined: 3-week interval between the 2nd R-CHOP cycle and iPET to avoid inflammatory reactions (Eur J Nucl Med Mol Imaging 30:682, 2003), no G-CSF after the 2nd R-CHOP cycle to avoid altered glucose biodistribution (J Nucl Med 47:950, 2006), standardized uptake value (SUV)-based PET interpretation to improve objectivity of evaluation (favorable iPET response: reduction of maximum SUV by > 66 % compared to baseline; J Nucl Med 48:1626, 2007). Pts. with CD20-positive lymphomas and a favorable iPET were randomized to receive 4 additional cycles of R-CHOP or the same treatment plus 2 extra doses of R (Part A of the trial). Pts. with an unfavorable iPET were randomized to continue standard R-CHOP for 6 additional cycles or receive 6 blocks of a more complex and more intensive protocol yielding excellent results in Burkitt and other aggressive lymphomas (Part B). Its main components were hyperfractionated alkylating agents (C, ifosfamide) and high doses of methotrexate and cytarabine, with dose reductions in pts. > 60 yrs. Other constituents were R, H, O, vindesine, etoposide and dexamethasone (Blood 120: abstr 667, 2012). R was omitted in pts. with CD20-negative lymphomas. Sample size was based on the empirically derived assumption that treatment failure after 2 yrs. (TF: progression, relapse, treatment discontinuation due to toxicity, start of alternative therapy, death of any cause) could be improved from 80 % to 90 % in Part A and from 30 % to 45 % in Part B (alpha=0.05, power=0.8). Complete remission (CR), overall survival (OS) and toxicity were secondary endpoints.

Results: From 2007 to 2012 926 pts. were recruited by 57 participating oncological centers and analyzed by PET in 23 nuclear medicine institutions. With a median follow-up of 33 months 853 pts. are currently evaluable in the intent-to-treat population. 757 pts. had CD20-positive B cell lymphomas (80 % diffuse large B cell [DLBCL], 3 % primary mediastinal B cell, 8 % follicular lymphoma grade 3), 13 had CD20-negative B cell lymphomas and 83 had peripheral T cell lymphomas. Interim PET was favorable in 746 pts. (87 %) and unfavorable in 107 (13 %). It was highly predictive of outcome, time to TF being significantly higher in Part A than Part B (2-year probability: 79 % vs. 47 %; hazard ratio (HR) for B 3.4, 95 % confidence interval (CI) 2.6 – 4.6, p<0.0001; Figure). On multivariate analysis iPET response, International Prognostic Index and B vs. T cell lineage independently predicted TF. Interim PET was also predictive of OS (HR 3.9, CI 2.7 – 5.7, p<0.0001). In pts. with CD20-positive lymphomas and a favorable iPET, addition of 2 extra doses of R failed to improve TF (HR for 2 extra doses 1.2, CI 0.8 – 2.1) and all secondary endpoints. Likewise, in pts. with an unfavorable iPET response, a switch from R-CHOP to the Burkitt-type regimen showed no beneficial effect on TF (HR for Burkitt 1.6, CI 0.9 – 2.7), CR rate (50 % vs. 31 %, p=0.10) or OS (HR 1.0, CI 0.5 – 2.1). Similar results were obtained, when the analysis was restricted to DLBCL, and for covariate adjusted Cox regression of all survival endpoints. Although treatment related deaths (3 vs. 2 pts.) were comparable in both treatment arms, the Burkitt protocol was associated with more severe grade 3/4 leukopenia (84 % vs. 67 %, p=0.043), thrombocytopenia (63 % vs. 35 %, p=0.007) and mucositis (41 % vs. 12 %, p=0.002).

Conclusion: Applying strict rules to the procedure and its interpretation iPET proved highly predictive of outcome in pts. with aggressive lymphomas in this large multicenter trial. Because switching to a more aggressive protocol failed to improve outcome, our results do not support a change in cytotoxic regimen in poor iPET responders.


Duehrsen:Amgen: Honoraria, Research Funding; Roche: Honoraria, Research Funding. Klapper:Roche: Research Funding. Hoelzer:Amgen: Speakers Bureau; Medac: Membership on an entity's Board of Directors or advisory committees.

Author notes


Asterisk with author names denotes non-ASH members.