Systemic Front Line Therapy of Follicular Lymphoma: When, to Whom and How

Abstract

The natural history of follicular lymphoma is usually characterized by an indolent course with a high response rate to the first line therapy followed by recurrent relapses, with a time to next treatment becoming shorter after each subsequent treatment line. More than 80% of patients have advanced stage disease at diagnosis. The time of initiation and the nature of the treatment is mainly conditioned by symptoms, tumor burden, lymphoma grading, co-morbidities and patients preference. A number of clinical and biological factors have been determined to be prognostic in this disease, but the majority of them could not show to be predictive of response to treatment, and therefore can’t be used to guide the treatment choice. CD20 expression is the only predictive factor recognized in the treatment of FL and justifies the use of “naked” or “conjugated” anti-CD20 monoclonal antibodies as a single agent or in combination with chemo- or targeted therapy. Nevertheless, as this marker is almost universally found in FL, it has little role in the choice of treatment. The outcome of patients with FL improved significantly in the last years, mainly due to the widespread use of rituximab, autologous and allogeneic transplantation in young and fit relapsed patients, the introduction of new drugs and the improvement in diagnostic accuracy and management of side effects. Agents as new monoclonal antibodies, immuno-modulating drugs, and target therapy have recently been developed and approved for the relapsed setting, while studies to evaluate their role in first line treatment are still ongoing. Here we report our considerations on first line treatment approach and on the potential factors which could help in the choice of therapy.

Introduction

Follicular lymphoma (FL) represents 70% of all indolent non Hodgkin lymphomas with an increasing incidence in western countries over last two decades. More frequently FL is diagnosed in asymptomatic advanced stage patients. The clinical course is heterogeneous, characterized by a high response rate to first line therapy and by subsequent relapses that need subsequent therapy lines or eventually, a re-challenge with previous lines. The advent of biological agents as rituximab (R), the improvement of supportive care and availability of different chemotherapy (CT) regimens improved the prognosis of FL in the last decades,^1–5 with an expected median survival nowadays approaching or even exceeding 20 years, particularly in the younger patients.^6,7 Early disease progression and transformation into high grade lymphoma remain the major reasons for shorter survival.⁸ Classically, and according to most guidelines, the decision of when to treat and which treatment regimen to use is made considering stage (early versus advanced), tumor burden (low or high) and patients symptoms: in this review we also address others factors that could as well influence the decision making. Since patients with FL usually have a long life expectancy and the course of FL is characterized by recurrent relapses, the choice of the treatment regimen and its starting time should be made considering not only overall response rate (ORR) and time to progression, but also its impact on quality of life (QoL), future therapies (including transplantation), risk of transformation into aggressive lymphoma and risk of long-term side effects, like therapy-related secondary malignancies.⁹

Variables which could potentially influence the choice of treatment in FL include:

• Clinical prognostic scores: baseline FLIPI (Follicular Lymphoma International Prognostic Index) and FLIPI2 are validated as prognostic risk factors for early and advanced FL in clinical trial settings and outside of clinical trials, irrespectively to CT, R-based and immuno-chemotherapy-based treatment.¹⁰ More recently, the m7-FLIPI score, combining clinical parameters (FLIPI and Performance status according to ECOG score) with gene mutations (EZH2, ARID1A, MEF2B, EP300, FOXO1, CREBBP, and CARD11), demonstrated a prognostic value in FL patients receiving first-line immuno-chemotherapy (ICT).¹¹ Nevertheless, at present, no study has yet shown that patients of different clinical risk categories need different treatments. Since patients with higher FLIPI score have a higher risk of relapse,¹² it could be tempting to intensify the systemic treatment in these patients, but so far this has only shown to improve progression free survival (PFS) and not overall survival (OS).¹³

• Tumor burden: for patients with advanced stage, the British and the French study groups defined clinical criteria to identify patients with high tumor burden (HTB) (Table 1), for whom an immediate treatment should be recommended. For patients with HTB, the treatment goal is primarily the response rate, in order to obtain as soon as possible a relief of symptoms and a longer PFS. In this situation, an immediate local treatment as palliative radiotherapy (RT) and/or systemic therapy is highly recommended. Asymptomatic patients who do not fulfill all criteria can be defined as having low tumor burden (LTB) and do not generally require any immediate treatment.

• Bulky disease: despite several trials recognizing tumor bulk as a poor prognostic factor for response and survival,¹⁴ there has not yet been a worldwide consensus on how to define the “bulky disease”^13,15,16 and how to treat FL patients with a large tumor mass. The frequent practice of consolidation with involved-field radiotherapy (IFRT) after ICT is not supported by randomized studies. Only a retrospective analysis, performed in advanced stage bulky FL, failed to demonstrate significant differences in terms of ORR, PFS and OS between patients treated with consolidation IFRT and those did not.¹⁵ Given that the 2014 European Society for Medical Oncology (ESMO) guidelines advice the use of systemic treatment in early stage patients with adverse prognostic factors, patients with bulky disease may be treated with systemic approach even in early stages.¹⁷

• Age and sex: pediatric-type FL is a new separate disease in the 2016 revision of the World Health Organization (WHO) classification,¹⁸ and will not be considered in this discussion. In non pediatric patients, three large analysis were performed demonstrating that increasing age has a negative prognostic value: in one study patients younger than 40 years had a favorable prognosis compared to older ones;¹⁹ in the other two, follicular lymphoma-specific survival was similar for cases < 40 years and in those aged 40–60.^6,7 All studies concluded that younger patients have a similar or better prognosis than older patients, and there is, therefore, no need to adopt for them a more aggressive approach. In some prospective trials, gender was recognized as a prognostic factor in FL patients treated with R given either upfront²⁰ or as maintenance (RM).²¹ Females, particularly of older age, seem to have a better response quality (regarding remission quality and PFS) and this appears to be related to a lower clearance of the drug compared to men.²⁰ Based on the available data it is not yet justified to base treatment decisions on gender or age, although fertility preservation issues should be discussed with the patients in the decision making.

• Grading: most studies on the prognostic and predictive role of histological grade were conducted in the pre-Rituximab era, they enrolled a small number of patients and had a relatively short follow-up. Moreover, the lack of reproducibility among pathologists on the evaluation of grade makes the interpretation of international data even more difficult.^22,23 The 2008 WHO classification recognises three grades of FL according to the number and distribution of centroblasts (0–5, 6–15, and > 15 per high-power field, respectively). Grade 3 is further subdivided into 3A (centrocytes still present) and 3B (sheets of centroblasts). Differences in molecular, genetic and clinical behavior suggest that FL G1 and G2 have an indolent course while G3 has a more aggressive course. While FL G1 and G2 is applied a classical FL treatment algorithm (according to stage and tumor burden), FL G3B is usually treated following a Diffuse Large B-Cell Lymphoma (DLBCL) algorithm, due to many histological, biological and clinical similarities among the two entities. Some studies with a follow-up longer than five years demonstrated a potential curability of patients with FL G3B receiving an anthracycline–containing regimen.^24,25 For FL G3A there is more controversy, as some would treat them as FL, and others as DLBCL. Most studies, reporting the outcome of patients with FL G3A, were retrospective and not randomized, and their results and conclusions are not unanimous. The distinction between FL G3A and FL G3B is not always possible because the studies were performed before the publication of the 2008 WHO classification. Most studies failed to show a survival benefit^22,24 and potential curability in long-term FL G3A survivors after anthracycline-based CT^23,24,26 (Table 2) as reported for FL G3B. A retrospective Swedish study on a cohort of 505 patients with long follow-up showed that patients with advanced stage FL G3A have a similar outcome to FL G1–2 regardless if anthracycline were administered or not,²⁴ hence, they may be treated similarly.²⁷ No data are available on RT efficacy in early stage FL G3A, thus a systemic treatment is advised, particularly in bulky or high FLIPI score patients.¹⁷

• Genetic features, microenvironment, and serum factors: the t(14;18) with BCL2 rearrangement is present in up to 90% of the nodal FL G1–2 but is less frequent in FL G3B. High levels of circulating t(14;18) can be found in healthy individuals and can predict the onset of FL some years earlier.²⁸ A recent study, performed by German Low-grade Lymphoma Study Group, showed that FL with and without BCL2 breaks had no differences in survival outcome.²⁹ The last WHO classification also identifies a distinctive subtype of t(14;18)-negative nodal FL (diffuse-appearing FL) characterized by deletion 1p36. It is usually confined to the inguinal region with an indolent course and its standard treatment strategy is unknown.³⁰ Positivity of BCL6 rearrangements and Myc abnormalities are associated with negative outcome. The acquisition of Myc positivity in FL might identify a biological transformation into a more aggressive disease deserving a specific treatment approach. Others genetic factors (mutTP53, mutMLL2, mutEZH2 and delCDKN2A), microenvironment gene expression profile (genes related to T-cell and macrophage activation)³¹ and serum chemokines (IL-2R, IL-1RA, and CXCL9)³² are recognized to be associated with a poor prognosis in FL. However, all these factors need to be confirmed in prospective clinical trials before being applied for risk stratification, and risk adapted treatment in daily clinical practice.

• Primary extra-nodal localization: duodenal-type FL and primary cutaneous follicle centre lymphoma are molecularly distinct entities with specific characteristics, excellent prognosis and indolent course but, due to their rarity, there is not a worldwide consensus on their optimal treatment. Cutaneous follicle centre lymphomas can be treated with local RT.³³ Duodenal-type FL displays excellent prognosis with watchful waiting (WW) approach if asymptomatic or after complete resection, therefore, it seems justified to delay a systemic treatment until symptomatic disease progression.³⁴ Recently, a study demonstrated that also RT could be effective and safe in this setting.³⁵ In both cutaneous and duodenal localizations, systemic front-line treatment (R alone or ICT) is only required in symptomatic patients with extensive and multifocal disease.

• Risk of transformation: Histologic transformation (HT) to an aggressive lymphoma is a well-described event in the natural history of FL. However, it remains unclear if there is already a predisposition to HT and whether this can be detected at the time of diagnosis.³⁶ The risk of HT into aggressive lymphoma is approximately of 3% per year during the first ten years and seems correlated with advanced stage and high IPI and FLIPI scores.³⁷ These scores merely predict a poorer survival, reflect tumor load and clinical characteristics but are not markers of the individual susceptibility to transformation.³⁶ A more important point is whether early systemic treatment could influence the HT risk. This remains an open question as data on this issue are limited and controversial. Patients who were initially observed had an unexpectedly higher rate of transformation compared to those treated upfront with systemic therapy in the St Bartholomew’s series³⁷ but not in other retrospective studies. In our series, patients receiving CT have a higher risk of HT compared to patients in WW or treated with R alone;³⁸ similar results were also reported by an epidemiologic study from the Mayo Clinic.³⁹ The randomized studies comparing expectant management at diagnosis with immediate treatment do not help clarify this issue since data on HT are reported only by two studies conducted more than two decades ago: a GELA trial found no differences in the risk of HT,⁴⁰ while in a study of the National Cancer Institute, the patients randomly assigned to WW appeared to carry a higher risk of HT.⁴¹ This latter study, however, has never been fully published. The possible influence of the type of initial chemotherapy on the likelihood of transformation is also controversial: a retrospective trial concluded that initial treatment with anthracyclines may play a preventive role,⁴² but in another retrospective study the HT rate did not change according to whether patients had received doxorubicin or not.⁴³ The role of RM in preventing transformation remains also unsolved: in the PRIMA study there was no significant difference in the risk of transformation between patients who received RM versus those who did not. ⁴⁴ These data was later confirmed by Ardeshna.⁴⁵ On the other hand, in the National LymphoCare study patients on RM had a lower incidence of transformation.⁴⁶ In conclusion, now there is no clear evidence to suggest any strategy in order lower the risk of subsequent HT.

• Positron Emission Tomography (PET): FL of any grade is fluorodeoxyglucose (FDG) avid,^47–49 and its FDG avidity is heterogeneous within and across patients. The SUVmax is widely variable (range 3 to 40) and does not seem to correlate with the risk of progression.⁵⁰ However, there is increasing evidence of a correlation between the uptake intensity and the histological grade and ─ despite a certain degree of inconsistency in defining a SUVmax threshold─ FDG-PET/CT imaging can be an effective tool to detect histological transformation.^51–53 Indeed, large cell transformation is most often focal, and it is associated with higher FDG uptake. Hence, to increase the chances of detecting an underlying transformation, diagnostic biopsies should, whenever possible, be directed to the site of greatest FDG avidity.⁵⁴ It has been suggested that the emergence of a focal lymphoma site with SUVmax 3 times higher or more than the others on a single scan, or that has tripled its uptake on serial scans, raises suspicion of histological transformation and should be biopsied.⁵⁵ As data have become available, analogous to diffuse large B cell lymphoma and Hodgkin lymphoma, response assessment with PET/CT has been found to be an independent prognostic factor for FL progression and overall survival.^56–58 In a French prospective multicentre study of patients with HTB FL G1–3A treated with R-CHOP, without rituximab maintenance, the 2-years PFS was 51% and 87%, respectively, in patients PET-positive and PET-negative at the end of therapy.⁵⁸ Similar results were retrospectively found in other two cooperative studies, conducted on patients prospectively enrolled in the PRIMA⁵⁶ and the FOLL05⁵⁷ trials. A pooled analysis of central scan reviews of these three large studies confirmed that PET/CT performed at the end of ICT induction (with R-CHOP, R-CVP, or R-FM), is highly predictive of both PFS and OS, particularly when the PET-response status is defined using a positivity cut-off of 4 (uptake moderately increased above the liver at any site) on the now recommended Deauville 5-Point Scale (5PS).⁵⁹ Conventional CT-based response in this study was only weakly predictive of PFS and based on this evidence, the recent consensus guidelines of the ICML Imaging Working Group⁶⁰ and the Lugano Classification¹⁶ recommended that PET-CT rather than contrast-enhanced CT scanning should be considered as a new standard for initial staging and response assessment of FL. In the previously mentioned French cooperative study the response assessed by PET/CT after four courses of R-CHOP was also highly prognostic.⁵⁸ However, the role of PET/CT during treatment (interim-PET) is less well established and is not currently recommended. These data can be extremely important for the clinical practice since end-of-therapy PET can identify FL patients whose disease has no longer an indolent course with increased risk of progression or death.⁵⁹ Moreover, there is preliminary evidence suggesting that functional PET parameters such as the baseline metabolic tumor volume (MTV) can be used as biomarkers for the development of first-line PET-adapted approaches in FL.⁶¹ Unfortunately, the implication of PET results in the definition of treatment remains still poorly known and clinical studies to evaluate response-adapted strategies based on the results of PET/CT ─ possibly associated with assessment of minimal residual disease (MRD) ─ are highly warranted and represent a major challenge for clinical research in FL.⁶²

• Assessment of MRD (detection of BCL2/IGH rearrangement): it is already demonstrated that patients with MRD persistence after transplantation have a worse outcome.^63–66 However, the role of MRD assessment after standard ICT in FL remains very controversial. The available results of different studies are conflicting and have some limitations due to retrospective nature, small sample size, mixed tissue sources (peripheral blood versus bone marrow) and lack of prospective planning for MRD time points.^67–72 Two studies of the Fondazione Italiana Linfomi (FIL) showed that MRD is an independent outcome predictor in patients with FL receiving rituximab-intensive programs.^73,74 It has been suggested that monitoring MRD may be useful for starting preemptive therapies before a frank clinical relapse. On the other hand, maintenance may be unnecessary in patients with MRD negativity after induction treatment. To test this hypothesis, the FIL has launched the FOLL-12 phase 3 trial (NCT003170-60), which will evaluate in FL patients undergoing first-line therapy whether an FDG-PET and MRD response-based maintenance therapy can replace the standard RM therapy currently given to nearly all the patients responding to standard ICT.

Table 1.

Criteria for Low Tumor Burden definition

BNLI criteria

No rapid, generalized disease progression in the preceding three months

No life treating organ involvement

No evidence of renal or macroscopic liver infiltration

No bone lesions

No B symptoms or pruritus

Normal hematological function (Hemoglobin> 10g/dl, Platelets> 100×10^9/L, WBC count> 3×10^9/L)

GELF criteria

Normal serum concentration of Lactate dehydrogenase and beta2-microglobulin

Largest nodal or extra-nodal tumor lesion<7 cm

No more than 3 nodes in 3 distinct nodal areas with a diameter> 3 cm

No significant serous effusions

No risk of organ compression or compromise

No symptomatic spleen enlargement

No B Symptoms

Hemoglobin> 10g/dl, Platelets> 100×10^9/L, ANC>1.5×10^9/L

BNLI: British National Lymphoma Investigation; WBC: White Blood Cell; GELF: Group d’Etude des Lymphomes Folliculares. ANC: Absolute Neutrophil Count.

Table 2.

Outcome of patients with Follicular Lymphoma after first line anthracyclines-based chemotherapy according to histological grade

Study	Population studied	Treatments	Median follow-up	Conclusions of the authors
Miller et al., 1997[26]	FLCL (N=389)	Anthracycline-based regimen (100%)	17 years	No plateau of the survival curve was achieved after anthracyclines-based regimens
Rodriguez et al, 2000[25]	FLCL (N=62)	Anthracycline-based regimen ± RT (76%) Non-anthracycline-based ± RT (13%) RT only (11%)	15 years	One third of the patients never relapsed, suggesting a potential curability for FLCL after anthracyclines–based treatments
Chau et al., 2003[22]	FL G1 (N=92 ) FL G2 (N= 68) FL G3A (N=44) FL G3B (N=11)	Anthracycline-based chemotherapy (27%) Non-anthracycline-based (38%) RT (22%) Surveillance (13%)	55mo for FLG1 57mo for FLG2 80mo for FLG3	No difference in terms of OS and FFS among FL grades 1–3 or between G3A and G3B First-line anthracyclines did not influence OS or FFS in patients with G3.
Hsi et al., 2004[138]	FL G3A (N=35) FL G3B (N=10)	Anthracycline-based regimen (53%) Non-Anthracycline-based (31%) Not available (16%)	24mo	No difference in OS between FL G3A and G3B No survival improvement or potential curability after anthracyclines-based chemotherapy
Ganti et al., 2006[139]	FL G1 (N=59) FL G2 (N=135) FL G3 (N=136)	Anthracycline-based chemotherapy	9 years	FL G3A and 3B had a similar outcome, but those with a diffuse component of >50% had the worst outcome. OS and EFS curves showed a plateau for patients younger than 60 years of age at diagnosis
Shustik et al., 2011[23]	FL G3A (N=139) FL G3B (N=22)	Various initial therapies. Anthracycline-based regimen in 82% of G3B versus 36% of G3A One-third of the entire cohort received rituximab	45 mo	No difference in outcome between G3A and G3B and no plateau of the OS curves Analysis limited to FL G3A patients showed no evidence of curability with anthracycline-based therapy
Wahlin et al, 2012[24]	FL G1–2 (N=345) FL G3A (N=94) FL G3B (N=23)	Anthracyclines and Rituximab (7%) Anthracycline without rituximab (27%) Rituximab only (7%) Single alkylator (28%) Other regimens (3%) Local radiation (19%) Never treated (9%)	10 years	FL G3B patients reached a plateau OS curve beyond 5 years if treated with anthracyclines, while FL G1–2 and G3A patients continued to relapse beyond 5 years. FL G1–2 and G3A seemed equally indolent, with indistinguishable clinical courses, even in patients receiving anthracyclines.
Koch et al, 2016[140]	FL G3A (N=47) FL G3B (N=14)	Anthracycline and rituximab (52%) Anthracycline without rituximab (48%)	6.9 years	FL G3A and G3B had similar PFS and OS at 5 years, showing a survival curves plateau after 6 years
Mercadal et al., 2016[27]	FL G3A (N=88) G1–2 (N=369)	Anthracycline and rituximab (76%) Anthracycline without rituximab (24%)	5 years	no difference in terms of OS and PFS between the FL G3A and G1–2 cohorts in patients treated with anthracyclines

FL: Follicular lymphoma; G: grade; OS: Overall survival; FLCL: Follicular large cell lymphoma (characterized by more than 15 centroblastic cells per h.p.f. according to International Working Formulation); FFS: Failure Free Survival; RT: Radiotherapy; PFS: Progression Free Survival; mo: months.

All the above-mentioned prognostic factors are able to identify patients at high risk for early progression or death, but are not able to identify which patients need one or the other treatment strategy: currently the main parameters conditioning first line systemic treatment approach are still stage (limited versus advanced), tumor burden, grading of FL and presence of symptoms.

Treatment

Systemic front-line therapy for limited stage (stage I and II)

Stage I and II account for about 15–20% of FL and the median survival ranges up to 25 years from diagnosis. FL is highly radio-sensitive: RT achieves ORR of 90% if performed within one year from diagnosis and with a standard dose of 24 Gy in 12 fractions, without a higher risk of second malignancies.⁷⁵ When the involved sites are contiguous and can be safely encompassed within a radiation field, local RT can induce long-term remissions in about 50% of the patients and is therefore considered curative in the limited stage. The other 50% of the patients relapse within ten years outside of the radiation field.⁷⁶ Despite National Comprehensive Cancer Network (NCCN) and ESMO guidelines recommend IFRT as preferable first line strategy for non-bulky stage I and II FL, less than 40% of patients with limited stage are treated with RT alone in clinical practice.⁷⁷ Observation only after an excisional biopsy is also a feasible strategy: a retrospective analysis found a 10-years OS of 85% with this approach and, after a median follow-up of 7 years, more than 60% of the patients did not require any therapy.⁷⁸ Surgery alone could be considered in asymptomatic patients, who are not suitable for RT or systemic treatment due to short life expectancy or co-morbidities or in young women needing fertility preservation.

1. Systemic treatment VERSUS radiotherapy. No randomized trials comparing RT versus front-line systemic therapy are reported. The Lymphocare Study and other observational studies compared different approaches in early stage FLs: even if improved PFS was observed in patients treated with systemic treatment compared to those undergoing RT alone, no difference in OS was observed[79–82]. Since in limited disease both approaches (systemic and local treatment) seem not to have any survival prognostic value, the best approach should be the one with the best toxicity profile. Also favoring upfront RT, one study demonstrated a lower 10-years risk of transformation after RT in early stage disease (around 18%)⁸³ compared to the general risk of transformation (ranges from 15 to 31%).³⁷

2. Systemic treatment WITH radiotherapy: Since more than 50% of the patients treated with RT alone relapse outside the radiation field, the addition of systemic treatment to RT could reduce the risk of distant relapse. This combination does not seem to improve OS in older randomized trials and, moreover, it could result in overtreatment and be responsible for early and late side effects, unacceptable in this setting.^84,85 The only prospective trial (from the MD Anderson Cancer Center) exploring efficacy and tolerability of Radio-chemotherapy (RCT) concluded that this strategy led to a better Disease Free Survival at 10 years compared to RT alone, but without advantage in OS, probably because of a higher incidence of secondary tumors and myelodisplastic syndromes.⁸⁶ R may represent a less toxic systemic treatment alternative to CT and in vitro models provided evidence of significant synergism when it is associated with RT.⁸⁷ Furthermore, R may reduce MRD with the improvement of duration response and survival.⁶⁸ However, the available retrospective studies evaluating the outcome of R-RT combination showed improved disease control in terms of PFS for the combined strategy compared to RT alone, but no survival benefit was observed.^88–90

Systemic front-line therapy for advanced stage (stage III and IV)

About 80% of patients with FL are diagnosed with advanced stage. Despite the improved outcome, most patients still experience disease relapse at a median time of 1.5–5 years, depending on when the active therapy start⁹¹ and on the intensity of the treatment regimen used.^21,92 A standard first line therapy for advanced stage FL has not yet been defined. The main decisions to take at presentation are when to start therapy and which regimen to choose. This strategy depends on the histological grade (see above) and the tumor burden according to BNLI or GELF criteria.

1. Low tumor burden: As discussed above, since FL is considered an incurable disease and in most cases has an indolent and chronic course, patients with LTB and no symptoms can be followed until they fulfil the criteria for initiation of treatment. In the pre-rituximab era retrospective and prospective studies had excluded a potential survival disadvantage when CT is differed until symptoms or organ failure risk in patients with LTB.^93,94 WW delayed the exposure to CT and its potential side effects, with a better preservation of QoL. Spontaneous regressions (approximately 12% of patients) were observed in patients assigned to WW⁴⁵ and approximately 40% of patients aged over 70 years, never needed any treatment and died of non-lymphoma related causes.⁹¹ More recently, a chemotherapy-free approach with single agent R, given with the weekly × 4 schedule, achieved an ORR of more than 70% in untreated LTB FL, with a range of complete response (CR) of 26–36%[95, 96] and a median PFS of 23.5 months becoming longer if R is continued.^97,98 It was, therefore, tempting to evaluate if R alone could not substitute for WW in asymptomatic and LTB FL. A British randomized prospective study compared the outcome and QoL in the patients assigned to WW versus those to R. The patients treated with R upfront had a lower need of new treatment (46% versus 88%) and a better PFS (82% versus 36%) at 3 years, but this did not translate into a survival benefit (97% versus 94%). The QoL evaluation in this study showed that patients under the WW strategy were more worried about needing treatment and felt less control of their situation compared to patients treated with R upfront. Nevertheless, an initial surveillance strategy remains an acceptable approach for LTB FL patients, who do not want active therapy. How to perform WW approach remains object of discussion, clinical surveillance should be done every 3–4 months for the potential risk of rapid progression; radiological controls are recommended in case of new signs/symptoms that suggest progression disease.

2. High tumor burden: For patients with symptomatic advanced stage FL, various therapeutic options are available, ranging from single agent to multi-agents CT. The addiction of R to several CT regimens conferred an improvement in OS with a pooled HR for mortality of 0.63.^99–102 The optimal regimen of CT, given in combination with R, in advanced stage FL patients remains controversial; however age, co-morbidity, and patient preference should play an important role in the decision making. None of the ICT regimens demonstrated superior survival in FL G1–3A patients; nevertheless R-CHOP and purine analogues containing regimens (R-FM) obtained a higher ORR (91–93% versus 88%) and a longer time to treatment failure (62–59% versus 46% at 3 years) and higher 3 years-PFS (68-63% versus 52%) compared to R-CVP, but caused a higher toxicity.¹⁰³ An alternative ICT regimen is R plus bendamustine: its efficacy and toxicity in untreated advanced stage FL G1–2 were compared with those of R-CHOP or R-CVP in two randomised clinical trials.^92,104 R-bendamustine obtained a higher CR (40 versus 30%), longer median PFS (69.5 versus 31.2 months) and time to next treatment (TTNT) (not reached versus 42.3 months) compared to R-CVP or R-CHOP, with fewer side effects.⁹² R as a single agent has as well been studied in patients in need of treatment. Given in the standard 4 weekly infusions as induction, it obtained an ORR of 81% with 31% of CR and a median Event Free Survival (EFS) of 19 months in CT-naive FL G1–3A patients; data that can be extrapolated from the SAKK 35/98 trial, where responders to induction were randomised to RM versus observation. Tumor burden was not formally assessed according to GELF criteria when patients were enrolled, however almost all cases had a stage higher than I (96%) and half of them had bulky disease at the moment of starting the ICT (53%). A higher response rate and a longer EFS were obtained with prolonged treatment, with an improvement of ORR to 92% and of median EFS to 36 months.¹⁰⁵ If single agent R induction is to be used, the RESORT trial showed that RM fails to improve OS compared to observation or rituximab retreatment (RR): after a median follow-up of 3 years, the RR has not conferred any disadvantages regarding QoL and time to treatment failure. Nevertheless, RM strategy seems to confer a longer time to first CT, at the cost of an average use of R 3.5 times higher than with the RR strategy.¹⁰⁶ Recently, numerous trials have explored the safety and efficacy of the combination of R plus other biological agents. One of the most promising agents is lenalidomide that showed a synergism of action in combination with R (R2) and a good profile of tolerability in the relapsed FL setting.¹⁰⁷ The efficacy and tolerance of this combination were successively studied in untreated FL G1–3A patients.^108,109 Approximately half of the enrolled patients (45.4%) had an advanced stage FL and approximately half of them meet GELF criteria for HTB(54%). The R2 regimen achieved in the subgroup of FL patients an ORR range of 93–98% (CR/uCR range of 72–87%) with PFS of 89% and 78.5% at 2 and 3 years, respectively.^108,109 R2 in untreated FL G1–3a patients, strictly in need of treatment according to GELF criteria, was shown to achieved a higher CR/CRu rate than R alone (61% versus 36% in the independent radiological response review) in a randomised study at the expected cost of higher toxicity.^110,111 Moreover, a significant advantage of the combination arm was shown regarding TTNT.¹¹² Comparing the outcome after R2 regimen with those after ICT (as R-CVP, R-CHOP and R-FM),¹⁰³ where three years-PFS ranges from 52 to 83%, the R2 regimen could be considered a possible alternative first-line treatment in FL patients, because probably better tolerated. An on-going randomised study is now comparing the outcome of R2 versus ICT in HTB FL patients (RELEVANCE trail).

The phase III GALLIUM study compared the efficacy and safety of obinutuzumab (new ant-CD20) plus chemotherapy (CHOP, CVP or bendamustine) followed by obinutuzumab maintenance with those of rituximab plus chemotherapy followed by R maintenance in untreated indolent NHL patients (85% of them with FL). Results from a pre-planned interim analysis showed that obinutuzumab-based treatment significantly improved PFS compared to R-based treatment.¹¹³

Rituximab maintenance

A strategy of RM after first-line or after salvage therapy in relapsed patients was performed with the aim of improving PFS. A number of studies showed an improvement in disease control with RM after induction therapy with R alone,^98,105,114 with CT¹¹⁵ and with ICT²¹ with an HR of 0.54 in the meta-analysis.¹¹⁶ The price to pay for this effect is a higher rate of infection-related adverse events (HR of 1.67). All these studies failed to prove a statistically significant OS benefit with RM after first-line induction therapy,^116,117 although a favourable trend in OS was observed among patients receiving RM after salvage treatment.¹¹⁷ The schedules used for RM differ in various studies without any statistically significant effect on the outcome.¹¹⁶ In the studies that also enrolled patients in response or stable disease after first-line, RM was performed either with administrations every two months^21,105,114 or with four weekly doses every six months.^114,117 The different maintenance schedule is due to pharmacokinetic data, with detectable serum R ranging from 3 to 6 months after four infusions.^118,119 The optimal duration of maintenance with R in patients who maintain a status of remission or response is also not definitely established: an improvement of median EFS of approximately 20 months (16 versus 36 months) was obtained with 1 year of RM compared to induction only (trial SAKK 35/98) going up to 8 years-EFS of 45% in responding, previously untreated patients.^97,105 The SAKK 35/03 compared EFS after a short term RM (1 year) and after long term schedule with R every two months until a maximum of five years. After a median follow-up of 6.4 years, the difference in terms of EFS was ample (median EFS 3.4 versus 5.3 years), but insufficient to reach statistical significance and median OS was approximately of 8 years in both arms. Two years of RM is the currently preferred choice in the absence of a clear survival benefit with a longer term maintenance and because better tolerated (0% versus 11.6% of unacceptable toxicity).⁹⁸

Consolidation treatment

Since FL tends to relapse, consolidation with different approaches was studied for the advanced stage after response to induction therapy.

• - Autologous stem cell transplantation. Three randomized trials before the R era^120–122 and one after the advent of R¹²³ compared the outcome of High Dose Therapy followed by Autologous Stem Cell Transplantation (ASCT) versus conventional CT in first line: in all studies ASCT achieved a longer PFS, without any benefit in OS. This result was confirmed by a Cochrane meta-analysis,¹²⁴ which showed a trend toward increased risk of secondary malignancies in patients receiving ASCT. As a consequence, this approach is not justified as consolidation treatment after fist-line therapy outside of clinical trials.

• - Radioimmunotherapy (RIT). RIT was used as front-line therapy in FL^125–128 with interesting results, but its use as consolidation therapy is of potentially greater importance in clinical practice. Phase II trials demonstrated the feasibility, tolerability, and efficacy of this strategy after first line therapy with CT alone^129,130 or R-containing therapy.^131–133 In a randomized trial RIT with yttrium-90 (90Y) - ibritumomab tiuxetan (Zevalin) as consolidation after first line chemotherapy produced a longer PFS compared to observation alone, but without any survival benefit.¹³⁴ This benefit was not confirmed by a subsequent randomized study.¹³⁵ Nevertheless, both US Food and Drug Administration and European Medicines Agency approved 90Y-Ibritumomab tiuxetan as consolidation therapy in untreated FL patients in partial or complete response after first-line induction therapy. A randomized phase II trial, comparing RIT consolidation with RM, showed a higher PFS with RM than with RIT consolidation, without any difference in OS.¹³⁶ Moreover, both strategies (RM and RIT consolidation) showed comparable incremental quality-adjusted life-years before the first progression, but a higher incidence of haematogical toxicity was observed in the RIT arm.¹³⁷ Therefore consolidation with RIT after standard R-based induction therapy is not routinely performed.