Response-guided first-line therapy and treatment of relapse in aggressive lymphoma: 10-year follow-up of the PETAL trial

Interim positron emission tomography (iPET) predicts outcome in aggressive lymphomas.^1,2 The PETAL (Positron Emission Tomography–Guided Therapy of Aggressive Non-Hodgkin Lymphomas) trial investigated whether treatment intensification after 2 cycles of cyclophosphamide, doxorubicin, vincristine, and prednisone (CHOP; plus rituximab [R] for CD20⁺ lymphomas) improved survival.³ After a median follow-up of 4 years, we reported that intensification failed to show a positive impact on outcome.³ 

The study protocol also made recommendations for the treatment of progression or relapse (collectively referred to as “relapse”) and asked for documentation of all lines of salvage therapy. Outcome after relapse was not reported in our original publication because of short follow-up.³ 

Here, we report the final results of the PETAL trial for both first-line and salvage therapy.

The PETAL trial was registered at ClinicalTrials.gov as #NCT00554164 and approved by the German Federal Institute for Drugs and Medical Devices and the ethics committees of the participating centers. The study was performed according to the principles of the Declaration of Helsinki. All patients gave written informed consent. The first patient was included in November 2007, recruitment was stopped in December 2012, und the database was locked in July 2023.

Patients aged 18 to 80 years with all types of aggressive non-Hodgkin lymphoma except Burkitt lymphoma and primary central nervous system lymphoma and a positive baseline PET scan were eligible. Interim PET was evaluated using the ΔSUV_max method (maximum standardized uptake value [SUV_max]) that has been shown to be superior to the commonly used Deauville scale for outcome prediction.^1,4 An SUV_max reduction of ≤66% compared with baseline defined a positive iPET.⁵ 

After 2 (R-)CHOP cycles, iPET-positive patients were randomly assigned to receive another 6 (R-)CHOP cycles or 6 blocks of an intensive methotrexate-based Burkitt lymphoma protocol.⁶ Interim PET-negative patients with CD20⁺ lymphomas were randomly assigned to receive another 4 cycles of R-CHOP with or without 2 additional doses of R. Randomization of iPET-positive patients was performed during the entire recruitment phase, whereas randomization of iPET-negative patients with CD20⁺ lymphomas was restricted to the period between February 2010 and October 2011, when the recruitment goal was reached. Before February 2010, all such patients received 6 cycles of R-CHOP, whereas, after October 2011, they received 6 cycles of R-CHOP with 2 additional doses of R. Patients with CD20⁻ lymphomas and a negative iPET uniformly received 6 cycles of CHOP. A CONSORT diagram showing recruitment and treatment allocation has been published previously.³ 

Recommended first-relapse treatments were 2 cycles of (R-)DHAP (dexamethasone, high-dose cytarabine, and cisplatin) followed by high-dose BEAM(-R) (carmustine, etoposide, cytarabine, and melphalan) and autologous blood stem cell transplantation (auto-SCT) for auto-SCT–eligible patients,⁷ and 6 cycles of (R-)ESHAP (etoposide, methylprednisolone, high-dose cytarabine, and cisplatin) for auto-SCT–ineligible patients.⁸ 

The primary end point of first-line randomization was event-free survival, defined as progression, relapse, treatment incompatible with the protocol (including radiotherapy), toxicity-related discontinuation, or death of any cause. It was tested using the log-rank test at a 2-sided significance level of .05. The sample size calculation for the primary end point has been published previously.³ Other end points of first-line and relapse therapy included time to progression, progression-free survival, and overall survival (OS). All authors had access to the primary data. The statistical analyses were performed by U.D., K.-H.J., and M.N. using IBM SPSS Statistics, version 29.0 (Armonk, NY).

Further details of treatments, definitions, and statistical methods are provided in supplemental Data, including a synopsis of the PETAL trial protocol.

Follow-up was updated for 621 of 677 patients (91.7%) who were alive at the time of the first analysis. The median follow-up time was 10.3 years (interquartile range, 7.7-11.9). Table 1, Figure 1, and supplemental Figure 1 detail baseline characteristics and outcomes for the entire study population and histologically defined subgroups.

Interim PET predicted outcome in the entire population and in all subgroups except anaplastic lymphoma kinase (ALK)-positive anaplastic large-cell lymphoma (ALCL; negative predictive value for relapse in the entire population, 75.7%; positive predictive value, 52.8%; and diffuse large B-cell lymphoma [DLBCL] subgroup, negative predictive value of 77.1% and positive predictive value of 46.9%; Figure 2). A total of 754 patients (87.5%) had a negative and 108 (12.5%) had a positive interim scan. Treatment intensification by 2 extra doses of R in the iPET-negative group or a switch from (R-)CHOP to the Burkitt protocol in the iPET-positive group had no statistically significant impact on outcome (Figure 3).

The median time to first relapse was 0.8 years (range, 0.0-14.1; Table 1). Patients relapsing within the first year had significantly shorter OS from first relapse than patients relapsing later. Patients with stage I relapse fared significantly better than patients in more advanced stages (Figure 4). The lowest postrelapse survival rates were seen in ALK-negative peripheral T-cell lymphoma (Table 1). By July 2023, 286 patients had died. The main causes of death (supplemental Table 1) were lymphoma progression (n = 97, 33.9%), infection (n = 64, 22.4%), cardiovascular events (n = 38, 13.3%), and second primary malignancies (n = 32, 11.2%).

One hundred patients (11.6%) developed a second primary malignancy (1 type, 84; 2 types, 16). The median interval between iPET evaluation and diagnosis of a second primary malignancy was 4.7 years (range, 0.3-13.3). The most frequent diagnoses were prostate (n = 14), colorectal (n = 10), and lung (n = 8) cancer (supplemental Table 2). There was no significant correlation between second cancer incidence and type of lymphoma or type of first-line therapy (data not shown).

Details of salvage therapy were obtained for 238 of 240 (99.2%) relapsing patients (supplemental Table 3). Overall, 224 (93.3%) patients received antineoplastic treatment. The median number of salvage therapy lines was 2 (range, 1-7). There were no statistically significant survival differences between patients receiving 1, 2, or 3 to 7 lines (supplemental Figure 2).

At first relapse, 94 of 133 auto-SCT–eligible patients (70.7%) were started on (R-)DHAP as recommended in the study protocol (median age, 53 years [range, 23-75]; median number of cycles, 2 [range, 1-6]), but only 50 (53.2%) proceeded to BEAM(-R) and auto-SCT. Logistic regression analysis revealed that the only factor to be statistically significantly associated with premature treatment termination was the interval between iPET assessment and relapse. With increasing interval, the risk of not reaching auto-SCT diminished (odds ratio, 0.968 per month; 95% confidence interval [CI], 0.931-0.994; P = .019). Numerically, the risk of premature treatment termination appeared higher in females, patients above age 60, patients with T-cell lymphoma, and patients with stage II to IV relapse, but none of these conditions reached statistical significance (supplemental Table 4). At the discretion of the investigators, 30 patients received other high-dose chemotherapy combinations in first relapse, and 7 underwent auto-SCT as second-, third- or 4th-line salvage therapy (4 after failing (R-)DHAP in first line). Sixteen of 107 auto-SCT-ineligible patients (15.0%) were treated with (R-)ESHAP as recommended in the protocol (median age, 72 years [range, 61-79]; median number of cycles, 4 [range, 1-6]), but, due to progression, only ten (62.5%) received more than 2 cycles.

Ineligibility for auto-SCT was associated with inferior survival in the DLBCL subgroup (progression-free survival: hazard ratio [HR], 1.538; 95% CI, 1.069-2.211; P = .019; OS: HR, 1.544; 95% CI, 1.059-2.251; P = .023; supplemental Figure 3), but not in the total population (supplemental Figure 4). There were no statistically significant survival differences between recommended first-relapse treatments and other types of treatment. This was true for both auto-SCT–eligible and auto-SCT–ineligible patients (supplemental Figures 3 and 4).

Considering all lines of salvage therapy together, 127 patients received chemotherapy, immunotherapy, and/or radiotherapy without transplantation, 69 received auto-SCT without allogeneic SCT (allo-SCT), and 28 received allo-SCT (preceded by auto-SCT in 18 cases). At allo-SCT, 8 patients were in complete remission, 10 were in partial remission, 2 had stable disease, none had progressive disease, and information was missing for 8. The proportion of primary refractory patients, the interval to, and the disease stage at first relapse, were similar among patients receiving chemotherapy, immunotherapy, and/or radiotherapy without transplantation, auto-SCT without allo-SCT, or allo-SCT (Table 2). Age, however, differed significantly (median age for chemotherapy/immunotherapy/radiotherapy vs auto-SCT vs allo-SCT, 68 [range, 23-86] vs 58 [range, 25-76] vs 46 [range, 21-67] years; P < .001). OS was best after allo-SCT, although the difference between autologous and allogeneic transplantation did not reach statistical significance (Figure 4). Six patients (21.4%) died of allo-SCT–related complications (fungal pneumonia [2 cases], hepato-intestinal graft-versus-host disease, pleuro-parenchymal fibroelastosis, pancreatitis, and myocardial infarction).

Multivariable Cox regression analysis revealed lymphoma subtype (bad, ALK-negative peripheral T-cell lymphoma; HR, 1.696; 95% CI 1.083-2.658; P = .021), time of relapse (good, >1 year after iPET; HR, 0.359; 95% CI, 0.244-0.530; P < .001), stage (bad, stage II-IV; HR, 4.343; 95% CI, 2.011-9.383; P < .001), and type of salvage therapy (good, autologous transplantation; HR, 0.505; 95% CI, 0.327-0.780; P = .002; good, allogeneic transplantation; HR, 0.310; 95% CI, 0.156-0.615; P < .001) to be independently associated with OS. Age failed to reach statistical significance (Figure 5).

In the PETAL trial, iPET predicted 10-year outcome of aggressive non-Hodgkin lymphoma. However, iPET-based treatment intensification did not improve survival. The positive predictive value of iPET was only 50%, that is, half of the iPET-positive patients did well on continued standard therapy and therefore did not need a treatment change. The positive predictive value of the interim assessment may be improved by combining iPET with other markers of treatment response.⁹ Among these, the most promising candidate is circulating tumor DNA whose prognostic value has been shown to be independent of iPET response.¹⁰ Patients with poor metabolic and molecular responses to first-line therapy may benefit from immunological approaches that have recently been established in the relapsed setting, such as chimeric antigen receptor T cells^11-13 or bispecific antibodies.^14,15 Prospective studies are required to assess the value of a switch from chemotherapy to immunotherapy in such patients. The negative predictive value of iPET was close to 80%, supporting the use of iPET alone in treatment de-escalation trials. Two such studies have shown recently that treatment can be shortened in limited-stage DLBCL if the interim scan is negative.^16,17 

Interim PET was predictive of outcome in all lymphoma entities except ALK-positive ALCL. Failure to predict outcome in this entity may have been related to small patient numbers and a paucity of events in a disease with excellent long-term outcome. However, similar observations were also made in another iPET study.² In contrast to other aggressive non-Hodgkin lymphomas, the prognostic value of iPET in ALK-positive ALCL appears insufficiently defined.

Adherence to salvage therapy recommendations was better for auto-SCT–eligible than for auto-SCT–ineligible patients, reflecting the lack of a therapeutic standard for the latter. Adherence to treatment recommendations had no impact on outcome. The study confirmed the prognostic importance of time⁷ and stage of relapse,¹⁸ and, for DLBCL, the superiority of auto-SCT over immunochemotherapy alone.¹⁹ Five-year survival rates were ∼20% for salvage therapy without transplantation, 45% for auto-SCT, and 60% for allo-SCT, which compares favorably with earlier reports on allo-SCT.^20-23 However, relapsed patients undergoing allo-SCT were highly selected in the PETAL trial. First, they were significantly younger than patients receiving other treatments. Second, 40% of patients with available information received allo-SCT to consolidate a previously achieved complete remission, half were in partial remission, and none progressed during the previous line of therapy. Patient selection precludes an unbiased comparison of allo-SCT with other treatments for relapsed aggressive lymphoma.

In conclusion, iPET predicted long-term outcome of aggressive non-Hodgkin lymphoma, but iPET-based intensification of chemotherapy failed to improve survival. Whether modern immunological treatment approaches are better suited to achieve this goal should be tested in prospective studies. The management of relapse has changed considerably since the initiation of the PETAL trial. Adherence to protocol-based recommendations had no impact on outcome. Whether this also holds true in an era of novel treatment options, requires further study.

The authors thank the patients, investigators, and PET centers from all parts of Germany for their participation. The authors thank the staff of all trial sites for documenting the patients’ follow-up over a 15-year study period. The authors acknowledge support by the Open Access Publication Fund of the University of Duisburg-Essen. Open Access funding enabled and organized by Projekt DEAL.

The study was supported by research grants from Deutsche Krebshilfe (nos. 107592 and 110515; to U.D.).

Contribution: U.D., A.B., K.-H.J., and A. Hüttmann designed the trial; U.D., B.H., I.E.K., F.K., M.H., A. Hochhaus, H.-G.H., D.B., G.P.-K., M.T., M.G., G.U.G., O.S., K.J., H.B., T.G., A.G., R.S., U.M.M., G.H., and A. Hüttmann recruited patients; U.D., A.B., B.H., I.E.K., F.K., M.H., A. Hochhaus, H.-G.H., D.B., G.P.-K., M.T., M.G., G.U.G., O.S., K.J., H.B., T.G., A.G., R.S., U.M.M., G.H., W.K., M.N., and A Hüttmann collected and assembled the data; U.D., A.B., W.K., K.-H.J., M.N., and A. Hüttmann analyzed and interpreted the data; U.D. wrote the first draft of the manuscript; and all authors read and approved the final version of the manuscript.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

A complete list of the members of the PETAL (Positron Emission Tomography–Guided Therapy of Aggressive Non-Hodgkin Lymphomas) trial investigators appears in the supplemental Appendix.

Correspondence: Ulrich Dührsen, Department of Hematology and Stem Cell Transplantation, University Hospital Essen, Hufelandstrasse 55, 45147 Essen, Germany; email: ulrich.duehrsen@uk-essen.de.