Although the event rate did reach the expected level and longer follow-up is needed to adequately assess long-term toxicities, data from the phase 3 IELSG37 trial support the omission of radiotherapy in patients with primary mediastinal large B-cell lymphoma who achieve a complete metabolic response following chemoimmunotherapy.
Although the event rate did reach the expected level and longer follow-up is needed to adequately assess long-term toxicities, data from the phase 3 IELSG37 trial (NCT01599559) support the omission of radiotherapy in patients with primary mediastinal large B-cell lymphoma (PMBCL) who achieve a complete metabolic response (CMR) following chemoimmunotherapy.1
Data presented from the plenary session of the 2023 EHA Congress showed that at a median follow-up of 58.8 months (interquartile range, 41.9-63.7), the 30-month progression-free survival (PFS) rate with observation was 96.2% (95% CI, 91.1%-98.4%) vs 98.5% (95% CI, 94.2%-99.6%) with radiotherapy (P = .274). The stratified hazard ratio (HR) for PFS in terms of relative effect of radiotherapy vs observation was 0.68 (95% CI, 0.16-2.91), with an absolute difference of 1.2% (95% CI, –3.2% to 7.0%).
Additionally, the 30-month overall survival (OS) rate with observation was 99.2% (95% CI, 94.7%-99.9%) vs 99.3% (95% CI, 94.9%-99.9%) with radiotherapy (HR, 1.87; 95% CI, 0.17-20.7; P = .601).
“IELSG37 is the largest randomized trial ever conducted in PMBCL. [Based on our findings,] we can assume that mediastinal radiotherapy may be safely omitted in patients who obtain a CMR after initial chemoimmunotherapy,” said lead study author Maurizio Martelli, MD, of the Hematology Department of Translational and Precision Medicine at the University Sapienza in Rome, Italy, in a presentation of the data. “This is also in keeping with the results from single-institution retrospective studies conducted in Vancouver and Bologna [that used] a PET-driven approach in [this disease].”
PMBCL accounts for 6% of all diffuse large B-cell lymphoma (DBCL) and often presents in younger patients aged between 14 years and 30 years, according to Martelli. “It is considered a hematological entity, because it is a particular immune phenotype and gene expression profile that is closer to Hodgkin disease than to DLBCL,” he said.
PMBCL is an aggressive disease, but it can have a good prognosis if remission is rapidly achieved with dose-intensive chemoimmunotherapy. Although mediastinal radiotherapy could consolidate responses, it has been linked with increased risk of coronary or valvular heart disease and second malignancies. Thus, radiotherapy in PMBCL represents a “therapeutic dilemma” in the field, according to Martelli.
Previously, the IELSG-26 study (NCT00944567) set out to evaluate the role of [18F]fluorodeoxyglucose PET following treatment with rituximab (Rituxan) and anthracycline-containing chemoimmunotherapy in patients with PMBCL.2 Findings showed that of the 115 patients eligible for PET/CT scans, 47% had CMR, defined as a negative scan below the mediastinal blood pool uptake. CMR after chemoimmunotherapy was noted to be predictive of a higher 5-year PFS and OS rate.
In his presentation, Martelli noted that a Deauville score of 1 to 3 better discriminates patients who are at low or high risk of failure following induction treatment. Consolidation radiotherapy was administered to 102 of the 115 patients based on local policy, irrespectively of this score, at the end of first-line treatment.
With the IELSG37 study, investigators set out to evaluate whether mediastinal radiotherapy can be omitted in patients with PMBCL who achieve a CMR following conventional rituximab-containing chemotherapy regimens. To be eligible for enrollment, patients needed to have treatment-naïve PMBCL that was CD20 positive with no evidence of extranodal disease.3 They needed to be at least 18 years of age and eligible to receive curative-intent radiotherapy and chemotherapy.
After confirmed diagnosis of PMBCL at baseline PET scan, patients received standard chemoimmunotherapy with 1 of the following regimens: R-CHOP21, R-CHOP14, R-CHOEP, DA-EPOCH-R, or R-V/MACOP-B.1 After 5 or 6 weeks from the end of chemotherapy, an end-of-treatment (EOT) PET scan was sent for central review to a panel of 7 experts.
All patients determined to be PET negative, defined as having a Deauville score of 1 to 3, were randomly assigned 1:1 to receive mediastinal radiotherapy at 30 Gy within 8 weeks of chemoimmunotherapy or observation. Those who were PET positive, defined as having a Deauville score of 4 to 5, went off study but continued to be followed for objective response rate, PFS, and OS to local salvage therapy.
The primary end point of the study was PFS at 30 months from randomization in those who were PET negative at the end of induction chemoimmunotherapy. Secondary end points included OS at 5 years from registration and long-term toxicity.
The study was designed as a randomized, noninferiority trial that would compare radiotherapy with observation. The sample size of 376 patients was calculated assuming a 30-month PFS rate of 85% in both arms from randomization. The alternative hypothesis for the trial was a noninferiority margin corresponding with a PFS of at least 75% in the observation arm. To this end, 540 patients needed to be enrolled to the trial, with an expected PET-negative proportion of approximately 70%.
Two planned and 1 unplanned interim analysis revealed a number of events that was much lower than expected, Martelli said. When recalculating the sample size using the event rates, they would have needed to randomize a total of 1821 patients, which was not feasible.
As such, an independent data monitoring committee recommended not to increase the study size and duration to meet the planned noninferiority margins. They said to complete the planned accrual of 540 patients and to perform a data analysis for the primary end point after more than 80% of patients had a minimum follow-up of 30 months. In the final analysis, investigators needed to include the absolute difference between the 2 treatment arms.
A total of 545 patients were enrolled to 74 centers spanning 13 countries, which included Italy (n = 380), the United Kingdom (n = 44), Ukraine (n = 25), Switzerland (n = 17), Poland (n = 15), Czech Republic (n = 14), China (n = 12), Norway (n = 11), Canada (n = 10), Sweden (n = 7), Germany (n = 5), the United States (n = 3), and Portugal (n = 2).
Of the 545 patients enrolled, 536 were determined to be evaluable, and 530 had central PET review. Of the 530 patients, 239 were PET positive, and 268 patients were PET negative. Those who were PET negative, defined as having a Deauville score of 1 to 3, were randomly assigned 1:1 to observation (n = 132) or radiotherapy (n = 136). Patients were stratified by sex, regimen, country, and EOT PET score.
Demographic and clinical features at baseline were well balanced between the observation and radiotherapy arms, according to Martelli. The median age was 35.5 years (range, 29-46.5) and most patients were female (63% vs 65%), had an ECOG performance status of 0 (52% vs 54%), bulky disease (60% vs 65%), elevated lactate dehydrogenase (67% vs 67%), and a Diffuse Large B-cell Lymphoma Prognosis score of intermediate risk (73% vs 72%).
The most frequently received chemoimmunotherapy regimen in the observation and radiotherapy arms was R-MACOP-B/VACOP-B (28% vs 35%), followed by R-CHOP-14 (25% vs 25%), R-DA-EPOCH (18% vs 17%), R-CHOP-21 (21% vs 15%), and other (8% vs 9%). Most patients in the observation and radiotherapy arms had a EOT PET Deauville score of 2 (51% vs 52%) followed by a score of 3 (45% vs 45%), and a score of 1 (5% vs 3%).
Post-randomization follow-up was greater than 30 months in 97% (95% CI, 94%-98%) of the 268 patients. “The statistical power of the observed 30-month PFS [rate] of 98% in the radiotherapy arm with a sample size of 268 [patients] was more than 85%, with a difference of more than 6% at 30 months,” Martelli noted. “This confirms the power of our study.”
Regarding safety, 1 patient in the radiotherapy arm had grade 3 left ventricular systolic dysfunction and another patient in that arm had both grade 4 acute heart failure and grade 3 hypertension; all of these toxicities resolved.
Grade 4 metastatic melanoma was observed at 22 months from randomization in a patient in the radiotherapy arm; this patient died from concurrent sepsis. Grade 3 glioblastoma occurred greater than 55 months from randomization in a patient in the radiotherapy arm and this resulted in death. Lastly, grade 4 acute myeloid leukemia occurred at 15 months after randomization in a patient who received radiotherapy and this patient had a complete response after allogeneic transplant.
“Longer follow-up is needed to properly evaluate the long-term toxicity,” Martelli concluded.