CHICAGO—After first-line therapy for follicular lymphoma, positron emission tomography (PET) status is strongly predictive of survival, and use of PET-computed tomography (CT) rather than contrast-enhanced CT scanning should be considered the new gold standard for response assessment in clinical practice. That was the conclusion of researchers presenting a pooled analysis of centrally reviewed scans in three multicenter studies, reported here at the American Society of Clinical Oncology Annual Meeting (Abstract 8502). Other experts, however, still question the overall value of imaging for such patients due to the limited benefits and potential risks.
Although follicular lymphoma mostly has an indolent natural history, there are still about 15 percent of patients who die within five years, said Judith Trotman, MD, Associate Professor of Medicine at the University of Sydney, who reported the new data. “High-risk FLIPI/FLIPI-2 scores alone fail to identify these patients. I share your frustration in the limits of CT research trying to crystal-ball predict overall survival. Despite the recommendations against routine use of PET-CT for follicular lymphoma in the 2007 International Harmonization Project criteria [JCO 2007;25:579-586], though, it is commonly used in response assessment.”
She noted that the predictive value of 18F-FDG PET-CT in response assessment after induction rituximab-chemotherapy for advanced-stage, symptomatic follicular lymphoma was recently reported in three trials. To provide more precise survival estimates from a larger patient cohort with longer follow-up, she and her colleagues conducted a pooled analysis of centrally reviewed scans of those three studies to identify the best cut-off points for survival when applying the increasingly adopted 5-Point Scale for response assessment of FDG-avid lymphoma.
Patient data and conventional CT-based response assessment were recorded for all patients undergoing central PET review in the prospective multicenter GELA (PRIMA and PET Folliculaire) and FOLL05 studies. PRIMA was a retrospective analysis of local PET interpretation within a prospective study with independent CT assessment of 122 patients. Results were confirmed by independent scan review of 61 patients.
FOLL05 was a retrospective analysis of local PET reports within a prospective study with local CT assessment of 202 patients; and PET Folliculaire was a prospective standardized PET acquisition/assessment in accordance to the 5-Point Scale with local CT assessment of 106 patients.
Scans were assessed independently by three reviewers applying the standardized 5-Point Scale. PET status with a score of 3 or 4 or more was compared with patient characteristics, CT-based assessment and survival endpoints of progression-free survival (PFS) and overall survival (OS). Only scans of sufficient quality for central review were accepted.
A total of 246 patients, median age of 56, were reviewed for post-induction PET-CT. Three-quarters of the patients received R-CHOP, and 15 percent had rituximab maintenance therapy. PET was performed a median of 30 days after the last chemotherapy.
Trotman reported that of the 246 scans performed at the end of the induction immunochemotherapy, about one-quarter were positive, with a cut-off of 3 or more (FDG uptake was more in the mediastinum), and about 17 percent with a cut-off of 4 or more (moderately greater in the liver). Patient and baseline disease characteristics did not differ significantly between PET-positive and PET-negative patients.
With a median follow-up of 55 months, the CT and bone marrow-based overall response rate was 96 percent, including about half that were complete responses (CR). “Both PET cut-offs were highly predictive of PFS and OS, with a cut-off of 4 or more points being most reproducible and discriminatory,” Trotman said.
There was no difference in the individual baseline characteristics of PET-negative or PET-positive patients with either cut-off. A FLIPI (Follicular Lymphoma International Prognostic Index) score of 3 to 5 was associated with PET-positive status.
“There was excellent concordance between post-induction CT and PET status,” she said. Using these cut-offs, the hazard ratios for PFS and OS of PET-positive versus PET-negative patients were found to be 3.9 and 6.7, respectively. For PET-positive patients, the four-year PFS rate was 23.2 percent versus 63.4 percent in those who became PET-negative. The four-year OS rate was 87.2 percent versus 97.1 percent.
“Conventional CT-based response—CR/CRu versus PR—was weakly predictive of PFS, but not OS,” she said.
“In conclusion, this independent review of 246 scans and a median follow-up of 4.6 years after first-line rituximab-chemotherapy for follicular lymphoma confirms that post-induction PET-CT status is strongly predictive of PFS/OS. When performing PET-CT, conventional CT assessment provides limited additional value. PET-CT applying the 5-Point Scale should be considered the new gold standard for therapeutic response assessment in clinical practice.”
She stated that post-induction PET-CT is a platform to study response-adapted therapy: “Achieving PET-negative status can better assure patients, especially those otherwise in CRu or PR. The inferior survival of patients remaining PET-positive compels us to study PET-response-adapted approaches.”
Concerns from Discussant
The Discussant for the study, Christopher R. Flowers, MD, Associate Professor of Hematology and Medical Oncology at Emory University, said it is important to note that only scans of sufficient quality for central review were used in the study. “The response rate is quite high. Those who had a poor response had progressive disease. The response rate was high using PET as well. PET-positive scans with a score of 4 or more independently predicted PFS and OS in multivariate analysis.”
Similar results had also been seen in the U.S. National LymphoCare Study of 394 patients, with PET after rituximab induction therapy being predictive for PFS and OS, while CT could predict only PFS, Flowers continued. He questioned whether this PET population was unique: “The behavioral characteristics are different from the broader population of follicular lymphoma patients. Those with stable disease fared poorly.”
He also wondered whether the results could be generalized to other regimens, including those that are commonly used, as well as rituximab and bendamustine plus other novel agents that will appear in years to come. “It is still unclear how PET clearance behaves in the broad population that receives maintenance therapy. How can PET be used to tailor therapy for follicular lymphoma?”
When Is the Right Time?
Commenting on the study for this article, Mikkael Sekeres, MD, Director of the Leukemia Program at the Cleveland Clinic Taussig Cancer Center, said: “When is the right time to get a PET scan in follicular lymphoma? This is a slow-growing cancer. It is interesting to correlate PET-positive and PET-negative scans following first-line therapy for OS. But is the test finding what it is supposed to find? There should be some demonstrable effect on management. How much are we spending on PET scans for cancer when there is no impact on outcome or alteration of surgical approaches?
“It is provocative to show that PET-positive scans can predict survival, but would you send a PET-positive patient for transplant? If not, then the patient should not get a PET scan at the end of first-line therapy. If we still follow the same paradigm, why do PET scans?”
In answer to a similar question from the audience after her presentation, Trotman responded: “Early follow-up of follicular lymphoma is important. R-CHOP patients have a median PFS of only 10 months. If I was treating a patient who was PET-positive, I would see that patient more frequently than a patient who is PET-negative.”
PET Scans and Transplantation
Another reason to do PET scans may be to avoid sending patients for autologous stem cell transplantation (ASCT), said René-Olivier Casasnovas, MD, of Hospital Le Bocage in Dijon, France, who presented a study at the meeting about using a PET-driven consolidation strategy in patients with high-risk diffuse large B-cell lymphoma (DLBCL) (Abstract 8503).
He noted that the GELA (Groupe d'Etude des Lymphomes de l'Adulte— i.e., Study Group of Adult Lymphoma) standard for young patients with high-risk DLBCL (an age-adjusted International Prognostic Index [aaIPI] score of 2-3) is rituximab, doxorubicin, cyclophosphamide, vindesine, bleomycin, and prednisone (R-ACVBP) induction plus consolidative BEAM and ASCT.
“R-CHOP induction might be as efficient as, and possibly less toxic than, R-ACVBP,” he said. “Also, patients who have a negative interim FDG-PET after two or four cycles of induction chemotherapy have a better prognosis. Early PET-negative patients with high-risk DLBCL may not need first-line ASCT.”
In untreated DLBCL, about half the patients who only receive rituximab chemotherapy will survive, versus about three-quarters of those who go on to have a transplant.
The study he reported was a Phase II randomized trial designed in 2007 to compare the dose-dense intensive regimen of R-CHOP14 with the standard dose-dense regimen of R-ACVBP14. Eligible patients were ages 18 to 59 who had previously untreated CD20+ DLBCL and an aaIPI score of 2 to 3.
Patients were randomly assigned to four cycles of either R-ACVBP14 or R-CHOP14 induction. Consolidation treatment was driven by centrally reviewed PET assessment after two (PET2) and four (PET4) induction cycles.
Patients classified as PET2-/PET4- received a sequential immunochemotherapy consolidation. Those who were PET2+/PET4- underwent ASCT. PET4+ patients were considered as not responding to induction treatment and were eligible for salvage therapy.
The primary endpoint was to evaluate the complete response rate after four induction cycles. The 222 patients (median age of 46) were randomized to receive either R-ACVBP (114 patients) or R-CHOP (108 patients). Virtually all had stage III/IV disease and elevated lactate dehydrogenase (LDH) levels, and one-quarter had an ECOG status of 2 or more.
After induction treatment, 47 percent of the R-ACVBP arm and 39 percent of the R-CHOP arm patients had a complete response. PET2 and PET4 were negative in 30 and 53 percent of patients in the R-ACVBP arm, and 25 and 40 percent of patients in the R-CHOP arm, respectively. There were about one-quarter PET2-/PET4- patients in both arms.
Patients allocated to ASCT received the planned treatment in 83 percent of cases. Virtually all of those allocated to rituximab-chemotherapy received the planned treatment. Due to more frequent PET4+, patients in the R-CHOP arm more often received salvage therapy as post-induction treatment (39%) than did patients in the R-ACVBP arm (27%).
With a median follow-up of 45 months, the event-free survival (EFS) rate was slightly higher in the R-ACVBP group (43%) than in the R-CHOP group (31%). Both four-year PFS and OS were similar for both groups of patients (75% and 83%, respectively).
In conclusion, Casasnovas said, “The primary objective was not reached due to missing bone marrow reassessment in the R-ACVBP arm. The better EFS in the R-ACVBP arm is related to a better metabolic response compared with R-CHOP. PFS and OS were similar in both arms, but after a higher frequency of salvage therapy in the R-CHOP arm. So, the PET-guided strategy may equalize the efficacy observed between the two arms. PET2 negativity allowed sparing ASCT in one-quarter of these high-risk patients with no risk of the disease control.”
The trial shows that a PET-driven treatment of high-risk DLBCL patients is feasible in a multicenter trial setting, he continued, and based on PET visual criteria at four cycles, the CR rate was higher in the R-ACVBP arm and salvage was more frequently used after R-CHOP, possibly explaining similar PFS and OS in the two induction arms.
Flowers, who was also the Discussant for this study, said: “PET-positive regions should be confirmed by biopsy. This is a cautionary tale. Surveillance imaging in NHL is commonly recommended by guidelines, but the use of imaging is being questioned due to limited benefits and potential risks. In general, studies show that scan-only relapse rates in survival are relatively infrequently detected by PET only, with one to eight percent survival seen detected in isolated relapse.”
In addition, Flowers said, lifetime cancer mortality is increased for patients who undergo CT scans. “The five-year cumulative probability of lymphoma is important to consider even though the risk is quite low.”
There are alternatives to surveillance imaging, he stressed: “For example, use of absolute lymphocyte count [ALC] after completing therapy and in the follow-up period has predicted relapse. ALC/absolute monocyte count may also predict relapse,” and elevated LDH in DBLCL patients in remission has be shown to have a positive predictive value of only 14 percent.