Six-month rituximab-lenalidomide regimen in advanced untreated follicular lymphoma: SAKK 35/10 trial 10-year update

The outcome of patients with follicular lymphoma (FL) has improved continuously during the last 2 decades with long-expected survival,^1-4 mainly due to the development and introduction of anti-CD20 antibodies, often in combination with chemotherapy.^5-8 The disease is clinically and biologically very heterogeneous. Some patients have no need for treatment even after a long period of wait and watch, whereas others have a symptomatic, high tumor-burden disease already at diagnosis.

For asymptomatic patients with low tumor burden, a watchful waiting policy has long remained a widely accepted approach,⁹ but rituximab (R) monotherapy, sometimes with an extended R schedule, has resulted in delayed time to need chemotherapy, although without survival or cost benefit.¹⁰ Re-treatment with R instead of maintenance is also an option for low tumor-burden FL.¹¹ 

For patients with FL with generalized symptomatic disease, the combination of R and chemotherapy, mostly followed by R maintenance, has become the standard first-line treatment in many countries.^5-8 Chemotherapy-free treatments are also effective, even in patients with high tumor burden, as shown by the Swiss Group for Clinical Cancer Research (SAKK) and the Nordic Lymphoma Group. These groups have developed and conducted trials with R monotherapy, showing long-term survival comparable with immunochemotherapy but with less toxicity, because immunochemotherapy was only administered in subsequent lines of therapy to those who needed it.^12-18 This has provided a rationale for further development of chemotherapy-free treatment strategies.^19-22 With today's expected long survival for patients with FL, long-term treatment safety is becoming increasingly important.

Several nonchemotherapy drugs have shown clinical efficacy, with a favorable toxicity profile. Lenalidomide (L) is an immunomodulating drug with boosting effect on natural killer cells, T-cell costimulation, and monocyte-mediated antibody-dependent cellular cytotoxicity, thereby enhancing the activity of R against CD20⁺ tumor cells in vivo.^20-23 This synergism was supported by very promising results in early clinical trials^23-28 and confirmed in our previously published SAKK 35/10 trial.²⁷ The long-term follow-up data of this trial in previously untreated patients with FL are presented here and confirm the durable benefit of rituximab-lenalidomide (RL).

Details of the SAKK 35/10 study design and eligibility criteria have been published previously.²⁷ In short, the inclusion criteria were as follows: untreated, histologically confirmed grade 1 to 3A FL with CD20 expression on immunohistochemistry and without signs of histological transformation (confirmed by central pathology review); and stage II, III, and IV in need of systemic therapy (defined by the presence of at least 1 of the following conditions: symptomatic enlarged lymph nodes, spleen, or other lymphoma manifestations, bulky disease [longest diameter ≥6 cm], clinically significant progression over at least 6 months of any tumor lesion, B symptoms, hemoglobin <100 g/L or platelets <100 × 10⁹/L, or any clinically significant progressive decrease in hemoglobin or platelet count due to lymphoma). The presence of measurable disease was mandatory. Patients in need of urgent response, for example, because of existing or imminent organ compression, were excluded. Detailed information on patient characteristics is provided in Table 1.

Patients were randomly assigned 1:1 to receive RL or R monotherapy without any maintenance. The R schedule was based on the results of previous FL trials of the Nordic Lymphoma Group^13,15 and the dose of L as previously described²⁷; in short, R monotherapy (375 mg/m² IV at day 1 for weeks 1-4; and in responding patients, repeated at day 1 of weeks 12-15) or R (given at the same schedule) in combination with L (15 mg daily orally, starting 14 days before the first R administration and continuously given until 14 days after the last, up to a total 18 weeks). The continuous L schedule was chosen to have a continuous immunomodulatory activity.

Random assignment to treatment arms was stratified based on histological grade (1-2 vs 3A), bulky disease (<6 vs ≥6 cm), FL International Prognostic Index (FLIPI)²⁸ score (1-2 vs ≥3), and center, using the minimization method.

No adaptation of R dose was allowed, whereas the L dose was adapted to renal function and adverse events (supplemental Table 1). If either L or R had to be permanently suspended, the treatment continued as single-agent therapy.

Trial treatment was discontinued in patients with less than minor response (MR; defined as ≥ 25% reduction in the sum of the product of tumor diameters) at the first restaging of week 10 (±1 week), and these patients were followed until progress or, if still symptomatic, treated at the discretion of the local investigator but were still part of the intention-to-treat cohort. Patients with a response (greater than MR at week 10) were planned for a total of 18 weeks of treatment. Staging and restaging procedures were described previously.²⁷ Response evaluation always included computed tomography (CT) scan and bone marrow biopsy.

Follow-up assessments included routine blood counts, beta-2 microglobulin, and lactate dehydrogenase evaluation every 3 months, and physical examination every 6 months. Chest X-ray and abdominal ultrasound (or CT scan/magnetic resonance imaging, if deemed necessary) were repeated every 12 months. CT scans were mandatory at 30 months and 5 years after randomization for all patients, except those with earlier documented progression/relapse.

The primary end point was complete response (CR)/CR unconfirmed (CRu) at 6 months. Response was assessed according to the National Cancer Institute International Workshop criteria.²⁹ Secondary end points were progression-free survival (PFS), time to next antilymphoma treatment (TTNT), duration of response, and overall survival (OS).³⁰ Furthermore, CR/CRu at 30 months was used as a secondary end point.³¹ 

The sample size was calculated to detect a 20% increase in the CR/CRu rate with 90% power, using an 1-sided z test with alpha set at 10%. Survival functions were estimated by the Kaplan-Meier method, and treatment arms were compared by the log-rank test. Cox proportional hazards models were also used for the estimation of hazard ratios (HRs). Quantitative variables were summarized by median, minimum, and maximum values. The summary statistics presented for categorical data were the count and percentage of patients in each category. Associations between categorical variables were assessed using Pearson χ² tests.

Post hoc analyses on survival from a risk-defining event, according to progression of disease within 2 years from start of first-line therapy (POD24),³² were performed. Survival from the risk-defining event was from the time of POD24 or from 2 years after random assignment for the non-POD24 reference group.

All analyses were performed on an intention-to-treat basis.

The trial was planned and conducted in accordance with the Declaration of Helsinki and was approved by the respective ethics committees of participating centers. All patients provided written consent.

Between April 2011 and October 2013, a total of 154 patients were randomized. The long-term results are based on a data cutoff on 26 May 2023, with a median follow-up for the intention-to treat (ITT) population of 9.5 years (interquartile range, 8.6-10.3).

Seventy-seven patients (median age, 63 years; 52% with stage IV and 47% with poor-risk FLIPI score) were assigned to the single-agent R, and 77 patients (median age, 61 years; 48% with stage IV and 47% with poor-risk FLIPI score) were assigned to the combination arm.

The CONSORT diagram (supplemental Material, supplemental Figure 1) shows the patients’ flow through the trial. All patients were in need of therapy (indications for treatment are provided in supplemental Table 2).

Eight cycles of R were completed, as planned, by 71% of patients in the R arm and 91% in the RL arm, in which 58% of patients also were given ≥90% of the planned dose of L, whereas 29% had at least 1 dose-level reduction of L. In 7 of 10 patients, who began treatment at a low L dose level, according to their baseline creatinine clearance, a subsequent dose escalation was possible.

Treatment was discontinued according to the protocol due to insufficient response at week 10 in 16 of 76 patients (21%) treated in the R arm and in 3 patients of 77 (4%) in the RL arm. In total, 14 patients stopped due to toxicity, 1 in the R arm and 13 (of whom 11 patients stopped L only) in the RL arm (supplemental Figure 1).

The primary end point analysis was previously published, showing a significantly higher CR/CRu rate in patients treated with RL at both 6 and 30 months. The impact of the higher activity of the RL combination remained so in the 31 patients aged >70 years.²⁷ 

The late outcome data show a significant and sustained improvement for patients in the RL arm compared with the R monotherapy arm across all time-to-event end points, except for OS, which did not differ between the arms. At a median follow-up of 9.5 years, patients in the RL arm had a longer PFS than those in the monotherapy arm (median, 9.3 vs 2.3 years; HR, 0.58; 95% confidence interval [CI], 0.37-0.89; P = .013) and a longer TTNT (median, not reached vs 2.1 years; HR, 0.43; 95% CI, 0.27-0.67; P < .001; Figure 1A-B). The rates of early progression (POD24) were 29% in the RL arm vs 34% in the R arm (P = .483). Duration of response was longer with RL (median, not reached vs 3.2 years with R only; HR, 0.42; 95% CI, 0.21-0.86; P = .014) and with >60% of responders still in first remission at 10 years, compared with <30% in the R-only arm (Figure 1C). The OS was similar in both arms (77% vs 78% at 10 years; P = .881; Figure 1D).

In total, 14 and 15 deaths have occurred in the R and RL arms, respectively. The main cause of death was related to the progress of lymphoma in 4 patients in the R arm and 6 in the RL arm. “Other” reasons were reported for 3 patients in the R arm (pulmonary embolism, cardiac failure, and probable cardiac event) and 1 patient in the RL group (accident). Two deaths were recorded early, within 30 days after trial treatment discontinuation, both due to lymphoma. One death was reported as serious adverse event in the RL arm, whereas another patient, also in the RL arm, had started a new therapy before death.

According to multivariable analysis performed to explore the effects of the stratification factors, the response (CR/CRu) did not depend on FLIPI²⁸ or PRIMA–prognostic index (PRIMA-PI)³³ scores (low or intermediate vs high), tumor bulk (yes vs no), or grades (1 or 2 vs 3A) (supplemental Table 4). However, the PRIMA-PI score discriminated patients with significantly different risks of relapse/progression in both arms (supplemental Figure 2A-B), whereas FLIPI was not useful for patients treated with R monotherapy (supplemental Figure 2C-D).

The overall safety in both treatment arms was consistent with the first published analysis,²⁷ and no further safety concerns were noted during long-term follow-up. The improved outcomes with RL were not associated with any unexpected toxicity.

Two cases of biopsy-proven transformation to diffuse large B-cell lymphoma were reported in the RL arm and 2 cases of Hodgkin lymphoma in the R arm.

Four skin cancers (3 squamous cell cancer and 1 basal cell cancer) were observed in the RL arm. The number of second solid cancers was significantly higher in the RL arm, but their causal relationship with treatment is not clear. Two cases of prostate cancer, 2 lung adenocarcinomas (1 in situ), 1 (fatal) small cell lung carcinoma, and 1 rectal cancer occurred in the RL arm, and only 1 solid tumor (prostate cancer) occurred in the R arm (P = .09).

Details on the second-line treatments after trial therapy and their subsequent outcomes are given in supplemental Table 3; a trend toward higher CR/CRu rate was apparent for the R arm (P = .067), but when limiting the analysis to the subset of patients treated at relapse with either R-CHOP (rituximab, cyclophosphamide, doxorubicin, Oncovin [vincristine], and prednisone) or R-bendamustine, overall remission rates and CR/CRu rates were not significantly different.

Survival after second-line therapy does not show any difference between the arms (median, not reached; HR [R vs RL], 0.79; 95% CI, 0.34-1.84; P = .59).

R was first approved as monotherapy for patients with advanced and symptomatic chemotherapy-refractory FL.³⁴ Furthermore, first-line R monotherapy has shown efficacy in randomized trials,^12-16,18 suggesting that initial chemotherapy might be deferred in a significant proportion of patients with FL without compromising the outcome. However, no trial has compared single-agent R with chemotherapy or R-chemotherapy. The SAKK 35/10 trial is, to our knowledge, the first and only randomized trial demonstrating that the addition of L to R in first-line treatment is associated with significantly better response rates, PFS, and TTNT than R monotherapy. Although the RL 18-week schedule we used was significantly shorter than that which is currently considered standard, our long-term follow-up of the trial confirms and extends the initial results. OS was also excellent but with no differences between the treatment arms.

RL was approved in the United States as second-line therapy for FL in 2019, based on the results from the AUGMENT (www.clinicaltrials.gov identifier: NCT01938001), which compared RL with R + placebo in pretreated patients who were relapsed/refractory to at least 1 prior systemic therapy, showing a superior efficacy of the combination over R monotherapy, with an acceptable safety profile.³⁵ 

In our trial, all patients were previously untreated and in need of therapy. The administration of L 15 mg daily was initiated 2 weeks before R with the intention to modulate the immune system, and a continuous dosing of 15 mg daily was maintained until 2 weeks after the last R infusion. Although the continuous schedule may have resulted in higher rates of severe neutropenia and skin rash, it appeared manageable in our short-duration schedule, and the cumulative dose administered per month is equivalent to the intermittent regimen of 20 mg daily for 3 weeks, followed by 1 week without L, used in most other RL trials.

Response (CR, CRu, partial remission, or MR at week 10) after a first cycle of 4-weekly R administrations was required to continue with a second cycle, whereas most nonresponding patients got an alternative treatment outside the trial. Responses might have been slower in the R monotherapy arm, making the early difference between the 2 arms biased; but after the long-term follow-up, the median PFS in the ITT population was 2 times longer in patients in the RL arm than those in the R monotherapy arm, and TTNT was also prolonged.

Our early evaluation of response, with the primary end point being CR/CRu rate at week 23 (ie, 7-8 weeks after the last R infusion), affected the CR rate, because responses continued to improve over time. However, at the end of induction with 2 cycles RL, the observed CR rate in our trial is in the same range as in the GALLIUM trial, a first-line R-chemotherapy trial.⁸ 

Due to the initially asymptomatic disease, a few of our patients had experienced a period of watchful waiting, but at inclusion in the trial, defined treatment criteria were required, and most showed advanced stage and poor-risk FLIPI, similar to large registration frontline immunochemotherapy phase 3 trials.^8,36 Our patients had higher median age (62 years vs 56 years) and more poor-risk FLIPI (47% vs 28%) than the MD Anderson single-arm RL pivotal trial (ClinicalTrials.gov identifier: NCT00695786), which enrolled 110 untreated patients with indolent lymphomas, including 50 patients with FL²⁴; and all fulfilled the required criteria for treatment initiation, which was not the case in the pivotal trial. Importantly, our treatment schedule is shorter, only 18 weeks in the combination arm (vs 18 months). The inferior CR rates with our RL regimen (61% vs 87%) suggest that the length of the treatment schedule and patient characteristics, as well as the presence of symptomatic disease, are affecting the outcome. However, after an extended follow-up for both trials, the difference in long-term outcomes was less evident, with a median PFS of 9.3 years in our RL-treated patients but not reached in the updated pivotal trial.³⁴ 

The RELEVANCE trial (www.clinicaltrials.gov identifier: NCT01650701) is, to our knowledge, the first randomized trial comparing RL with immunochemotherapy in previously untreated patients with FL. Maintenance was given in both arms, with RL for 1 year followed by R for a second year in the experimental arm and with R alone for 2 years in the immunochemotherapy arm. CR/CRu at week 120 (CR30) and PFS were coprimary end points. Neither end point was significantly different between the RL and R-chemotherapy arms, and the estimated OS was not different.³⁷ 

CR30 has been identified as a reliable surrogate of PFS,³¹ and the CR/CRu30 rate of 48% (95% CI, 44-53) in the RL arm in the RELEVANCE trial was slightly higher than the 42% (95% CI, 30-53) obtained with our much shorter schedule. Positron emission tomography/CT was not used in any of the trials, whereas CR30 also included CRu. The 6-year PFS of 60% (95% CI, 55-64) in the RL arm in the RELEVANCE trial seemed slightly better than our result of 50% (95% CI, 38-61). The maintenance schedule affected the RELEVANCE results, and the lack of maintenance in our trial prevents direct comparisons of PFS, also with most other studies.

The long-term follow-up of the RELEVANCE trial with the prolonged RL schedule resulted in an estimated survival of 89% at 6 years,³⁷ and in the pivotal MD Anderson trial, the OS was 98% at 8 years.^24,37 In our trial, with longer follow-up, OS was similar in both arms: 77% vs 78% at 10 years, which is very promising, especially because our regimen is much shorter and therefore more feasible and less expensive. Moreover, our survival is comparable with the estimated 10-year OS of 80% in both arms of the phase 3 PRIMA trial, including patients with high tumor-burden FL, randomized to R maintenance after response to first-line R-chemotherapy.³⁸ 

Our RL regimen was also effective in patients with advanced, bulky disease. The PRIMA-PI score was relevant to use in both treatment arms. High-risk patients, according to PRIMA-PI, had inferior PFS in both arms than those with low and intermediate risk, but FLIPI (which was a stratification criterion) was useful only in the RL arm. In a planned subgroup analysis of the RELEVANCE trial, both PFS and CR/CRu30 were independent of FLIPI in the RL arm, whereas FLIPI was useful in the R-chemotherapy arm. No data on PRIMA-PI were, however, presented in the RELEVANCE trial.

An important surrogate of long-term outcome is POD24. The POD24 rates in the RL arm (29%) were higher than those reported in other recent trials (12% in RELEVANCE and 17% in the maintenance arm of PRIMA); however, the excellent median OS in our study mirrors the outcomes observed in the maintenance arm of the PRIMA study. The shorter RL treatment duration can be 1 explanation for the higher rate of POD24 in this study, whereas the efficacy of second-line therapies allowed for reaching a long-term OS. We have previously reported that in patients initially treated with R alone, the rates of POD24 are higher than those reported after immunochemotherapy.^17,39 The high percentage of POD24 had less effects on survival, because after initial biologic therapy, such as R only or RL, a second-line R-chemotherapy will often be effective, whereas POD24, after immuno-chemotherapy, is likely to reflect chemo-refractoriness and disease aggressiveness. Thus, outcomes in patients experiencing POD24 is dependent both on the type of first-line therapy and also second-line and later therapy.

Age is often a risk factor for worse outcome. In our trial, patients aged >70 years attained good durable responses and tolerated the treatment well, indicating that the older patients may represent a most suitable population for this nonchemotherapy of short duration.²⁷ A potential limitation of our study is that quality of life and patient-reported outcomes could not be assessed.

In the RELEVANCE trial, the annual transformation rates in the RL and R-chemotherapy groups were 0.68% and 0.45%, respectively. These rates appear higher than the 0.28% annual rate observed in our trial. The cases of second primary malignancies increased from 38 (7%) in 2017 to 57 (11%) in 2020 in the RL group and from 48 (10%) to 67 (13%) in the R-chemotherapy group in the RELEVANCE study. These rates seem comparable with the rate of 13% observed in our RL arm. Due to the small numbers, it is challenging to evaluate the clinical relevance of the significantly lower rate (1%) observed in the R-alone arm.

With RL regimens, both hematological and nonhematological toxicity is higher than with R monotherapy. The rates of the most common grade 3 and 4 events observed with our short combination was neutropenia in 23% and skin rash in 5% of the patients, which appeared to be in the same range as in other RL trials in patients who were not previously treated.^24-26,40 

One limitation of our trial is that the RL schedule is not exactly the one used in the US Food and Drug Administration–approved AUGMENT trial, but due to the positive long-term results of our 18 weeks short regimen, this is now used in current practice in the Nordic countries.

In conclusion, the results of our study support the growing evidence that RL regimens offer an alternative to standard immunochemotherapy, also in frontline. However, more data on the optimal duration and schedule are required. Nevertheless, our 18-week schedule might be more patient friendly in terms of toxicity (that appears acceptable and manageable), and the immune suppression is also less prolonged, making it a potentially attractive backbone for novel combinations with small molecules and bispecific antibodies. The excellent OS with RL and also with R only confirms that chemotherapy-free strategies may be a relevant treatment option. The goal for future patients is to identify response predictors for an optimized treatment while maintaining maximal quality of life without the risk of toxicity and worse outcome, in terms of both long-term side effects and OS.

The authors thank patients, coinvestigators, research nurses, and data managers at each of the participating centers for their invaluable contribution, as well as the central study team at the Swiss Group for Clinical Cancer Research coordinating center for their administrative support and their assistance in data collection and study conduction. Furthermore, the authors thank Verena Voelter for help in the initial study development.

The Swiss State Secretariat for Education, Research, and Innovation (SERI) provided support for the SAKK activity. Financial support was provided by Celgene (contract number IIT2976).

Contribution: E.K. and E.Z. designed the study, performed research, analyzed the data, and wrote the manuscript; S.S. performed statistical analysis and contributed to manuscript writing; S.H. contributed to statistical analysis; and all authors contributed to patient management and data collection, reviewed and approved the manuscript, and shared final responsibility for the decision to submit it.

Conflict-of-interest disclosure: E.K. received advisory board fees from AbbVie, Pfizer, and Pierre Fabre, and provided educational lectures for Janssen, AbbVie, and AstraZeneca. E.Z. received research funding from Celgene, Roche, and Janssen; has served on the advisory board of Celgene, Roche, Mei Pharma, AstraZeneca, and Celltrion Healthcare; received travel grants for meetings from AbbVie and Gilead; and provided expert statements to Gilead, Bristol Myers Squibb, and Merck Sharp & Dohme (MSD). B.E.W. received research funding from Incyte. U.M.M. has served on the advisory board of Roche, Janssen, Bristol Myers Squibb, Amgen, AbbVie, and Gilead; and has received travel grants for meetings from AbbVie, Janssen, Amgen, Gilead, and Bristol Myers Squibb. M.H. has served on the advisory board of MSD, Bristol Myers Squibb, Amgen, Incyte, Sanofi, and Novartis; has given lectures to MSD, Novartis, and Bristol Myers Squibb; and has received travel grants for meetings from Amgen, MSD, Sanofi, Roche, and Novartis. A.J.M.F. has received research funding from Celgene. T.Z. reports consultancy fees from BeiGene Switzerland. P.d.N.B. has served on the advisory board for Roche, AbbVie, and SERB. F.K. received the registration fee for a congress from Roche. The remaining authors declare no competing financial interests.

The current affiliation for D.R. is Inselspital University Cancer Center UCI, Das Tumorzentrum Bern, Bern, Switzerland.

Correspondence: Eva Kimby, Karolinska Institutet, Department of Medicine Huddinge, Center for Hematology and Regenerative Medicine (HERM), Medicinaren 25/Neo, HERM 7th floor, 141 83 Stockholm, Sweden; email: eva@kimby.se.