Continuing Medical Education Program in American Journal of Hematology
Follicular lymphoma: 2011 update on diagnosis and management†
Conflict of interest: Nothing to report
Follicular lymphoma (FL) is generally an indolent B-cell lymphoproliferative disorder of transformed follicular center B cells. FL is characterized by diffuse lymphoadenopathy, bone marrow involvement, splenomegaly, and less commonly other extranodal sites of involvement. In general, cytopenias can occur but constitutional symptoms of fever, nightsweats, and weight loss are uncommon.
Diagnosis is based on histology of preferably biopsy of a lymph node. Immunohistochemical staining is positive in virtually all cases for cell surface CD19, CD20, CD10 and monoclonal immunoglobulin, as well as cytoplasmic expression of bcl-2 protein. The overwhelming majority of cases have the characteristic t(14;18) translocation involving the IgH/bcl-2 genes.
The Follicular Lymphoma International Prognostic Index prognostic model for FL uses five independent predictors of inferior survival: age > 60 years, hemoglobin <12 g/dL, serum LDH > normal, Ann Arbor stage III/IV, number of involved nodal areas > 4. The presence of 0–1, 2, and ≥3 adverse factors defines low, intermediate, and high-risk disease with median 10 year survivals in the pre-rituximab era of approximately 71, 51, and 36 months, respectively. With the use of more modern therapies, specifically anti-CD20 monoclonal antibody, the outcome has improved.
Observation continues to be adequate for asymptomatic patients with low bulk disease and no cytopenias. For patients needing therapy, most patients are treated with chemotherapy plus rituximab, which has improved response rates, duration of response, and overall survival. Randomized studies have shown additional benefit for maintenance rituximab both following chemotherapy-rituximab and single agent rituximab. Autologous stem cell transplantation has not shown a survival benefit in first remission patients. Stem cell transplantation (SCT) including both autologous and allogeneic SCT or experimental agent therapy is considered for recurrent disease. Am. J. Hematol. 86:769-775, 2011. © 2011 Wiley-Liss, Inc.
Follicular lymphoma (FL) is the second most common lymphoma diagnosed in the United States and western Europe, approximately 20% of all non-Hodgkin lymphomas (NHLs), and 70% of indolent lymphomas . The median age at diagnosis of 60 years, and a slight female predominance [2, 3]. The incidence is slightly increased among relatives of persons with FL .
FLs are malignant counterparts of normal germinal center B-cells . Approximately 85% of patients with FL have t(14;18) that results in the overexpression of the BCL-2 protein, a member of a family of proteins that blocks apoptosis. However, it is likely that multiple genetic events are required for the development of FL since the t(14;18) translocation can be identified in the B cells from normal individuals and patients with diffuse large B-cell lymphoma.
Patients with FL present with asymptomatic lymphadenopathy, with waxing and waning present for years. Bone marrow involvement is present in 70% of patients, whereas involvement of other normal organs is uncommon. Less than 20% of patients present with B symptoms and also less than 20% of patients present with an increased serum lactate dehydrogenase (LDH).
FL recapitulates normal germinal centers (GC) of secondary lymphoid follicles . The neoplastic cells consist of a mixture of centrocytes (small- to medium-sized cells) and centroblast (large cells). The clinical aggressiveness of the tumor increases with increasing numbers of centroblasts. The WHO Classification  has adopted grading from 1 to 3 based on increased numbers of centroblasts counted per high power field: Grade I with 0 to 5 centroblasts/high power field (hpf) (follicular small cleaved), Grade II with 6 to 15 centroblasts/hpf (follicular mixed), Grade III with more than 15 centroblasts/hpf (follicular large cell). Grade III has been subdivided into Grade IIIa, in which centrocytes are present, and Grade IIIb, in which there are sheets of centroblasts .
Bone marrow involvement is very common with paratrabecular lymphoid aggregates . FL cells express monoclonal immunoglobulin light chain, CD19, CD20, CD10, and BCL-6 and are negative for CD5 and CD23. In virtually all cases FL cells overexpress BCL-2, due to t(14;18). Clonal Ig gene rearrangements are also present and most cases have extensive somatic mutations.
Histologic transformation of FL from an indolent disease to a diffuse large B-cell lymphoma occurs in 10–70% of patients over time, with a risk of 3% per year, and is associated with rapid progression of lymphadenopathy, extranodal disease (besides the marrow), B symptoms, and elevated serum LDH [7, 8].
The two best measures of outcome are the FL international prognostic index (FLIPI) and tumor grade . There is also evidence that the characteristics of the associated cells in the tumor microenvironment of FL influences disease behavior and prognosis [10–12].
The Follicular Lymphoma International Prognostic Index (FLIPI) includes five prognostic factors including patient age, stage, number of involved nodal areas, serum lactate dehydrogenase, and hemoglobin . The FLIPI was developed out of an international study of survival data in 4167 patients with FL diagnosed between 1985 and 1992. This is shown in Table I. A modified version, of this score, the FLIPI2, evaluates five parameters, with some overlap of the FLIPI . At present, the current utility of the FLIPI2 model remains uncertain. Since the incorporation of rituximab into the mainsteam therapy of FL, the FLIPI has continued to be a useful prognostic model .
Table I. Follicular Lymphoma International Prognostic Index (FLIPI) 
|High risk||3 or more||52%||36%|
FL tumors are graded from 1 to 3 and this grade has some prognostic utility. There has generally been suboptimal consensus of pathologists on grading FL. There is no evidence to support a different treatment approach between Grade 1 and Grade 2 FL. Differences in molecular genetics and clinical behavior suggest that FL grade IIIa may be an indolent disease, whereas IIIb is an aggressive disease .
FL grade 3 has been historically referred to as follicular large cell lymphoma. Since many studies likely include both Grade IIIa and IIIb, this heterogeneity may affect interpretation of the outcomes. Although the follicular architecture is intact, the clinical presentation, behavior, and outcome with treatment in many patients with FL grade IIIb more closely approximates that of diffuse large B-cell lymphoma (DLBCL) [16–18]. In contrast to DLBCL, the relapse rate of FL Grade 3b is in some series is higher, but survival is longer .
Investigation of the cellular microenvironment of FL has provided interesting insights into prognosis [10, 12, 20-27]. It has been suggested that FL is an immunologically functional disease in which an interaction between the tumor cells and the microenvironment determines overall clinical behavior. These studies which have observed an impact on prognosis of the normal infiltrate of macrophages, T cells, and T cell subsets will need additional study in larger data sets, and prospectively with uniformly treated patient populations.
Less than 10% of patients with FL have Stage I/II disease . Radiation therapy is the treatment of choice for limited stage FL and results in 10-year overall survival rates of 60–80%; with a median survival is approximately 19 years . However, most patients with Stage I disease treated in the United States do not receive radiation therapy . Adjuvant chemotherapy does not add additional benefit after local radiotherapy . If significant morbidity is possible from radiotherapy based on the location of the disease area or if the patient chooses to not receive radiation, observation may be a reasonable alternative, especiallly for Stage II patients .
In a large study of over 6000 patients with Stage I or Stage II FL diagnosed from 1973 to 2004, 34% of whom were initially treated with RT , patients who received initial RT had higher rates of disease-specific survival at 5 (90 versus 81%), 10 (79 vs. 66%), 15 (68 vs. 57%), and 20 (63 vs. 51%) years . Selected early stage patients can be observed without initial treatment with radiation. In one report, the median overall survival of selected untreated patients was 19 years. At a median follow-up of 7 years, 63% of patients had not required treatment .
The overwhelming majority of patients have advanced stage disease at diagnosis. However, patients with asymptomatic FL do not require immediate treatment unless they have symptomatic nodal disease, compromised end organ function B symptoms; symptomatic extranodal disease, or cytopenias. This approach is supported by randomized prospective trials of observation versus immediate treatment. One of the largest trials compared immediate treatment with chlorambucil to observation . At a median follow-up of 16 years, no difference in overall survival and cause-specific survival was seen between the two approaches. Similar results have been noted in other prospective trials of initial treatment versus observation .
A major question is whether rituximab might change this approach in early treatment in asymptomatic patients. A study compared observation to rituximab alone or rituximab followed by maintenance in previously untreated FL. The median time to next treatment was 34 months in the watch and wait patient but was not reached in the rituixmab treatment arm. The 3-year PFS was 33, 80, and 90% of the observed, rituximab, or R+maintenance patients, with 95% OS in all three groups. The important issues of time to second therapy, quality of life, impact on histologic transformation, cost, toxicity, and future responses to rituximab are not yet addressed .
Rituxmab has changed the paradigm of treating FL. Retrospective analyses have reported improvements in survival of patients with FL even prior to the rituximab era . More recent improvement is largely due to the inclusion of antibody based therapy . The benefit of adding rituximab to combination chemotherapy for the initial treatment has been demonstrated in multiple randomized trials of chemotherapy with or without rituximab (Table II) [38–42]. All of these trials have demonstrated improved response rates and time to progression in the rituximab plus chemotherapy arms, as well as improvement in overall survival.
Table II. Phase III Trials of Chemotherapy Versus Rituximab + Chemotherapy in Previously Untreated Patients with Follicular NHL
|Marcus ||CVP vs. CVP-R||Improved OS|
|Hiddemann ||CHOP vs. CHOP-R||Improved OS|
|Herold ||MCP vs. MCP-R||Improved OS|
|Salles ||CHVP vs. CHVP-R (+ interferon)||improved OS in high 3-5 FLIPI|
Other chemotherapy drugs plus rituximab have also been reported. Bendamustine plus rituximab (B-R) has been compared to CHOP-R in a randomized phase III trial with bendamustine (90 mg/m2 Days 1 and 2) plus rituximab (375 mg/m2 Day 1) in 513 patients with advanced follicular, indolent, and mantle cell lymphoma . In preliminary results, superior progression-free survival (54.8 vs. 34.8 months) and event-free survival (54 vs. 31 months) with less toxicity, including lower rates of Grade 3 and 4 neutropenia and leukopenia was observed in the bendamustine plus rituixmab arm. There was no difference in overall survival at a median follow-up of 32 months between the two arms. Fludarabine plus rituximab  and fludarabine, mintoxantrone, dexamethasone and rituximab  both showed response rates in excess of 90% in previously untreated patients but significant neutropenia, opportunistic infections have been observed.
Rituximab alone has been used as the first therapy in patients with indolent lymphoma, with overall response rates of around 70% and CR rates of over 30% reported [46–49]. The most impressive data of single agent rituximab is the recent update of the SAKK trial . Patients received four weekly doses, and then patients with stable disease or better were randomized to observation or four doses of maintenance with one dose every 2 months. A recent update of this study was reported in 2010. In this study, 202 patients with previously untreated or relapsed/refractory FL administered four weekly doses of single agent rituximab has been reported. The 151 patients with responding or stable disease at week 12 were randomized to no further treatment or prolonged rituximab maintenance every 2 months for 4 doses. At a median follow-up of 35 months, patients who received the prolonged rituximab maintenance had a two fold increase in event-free-survival (23 vs. 12 months). Now with longer follow-up, 35% of responders remain in remission at 8 years with 45% of newly diagnosed patients in this study in remission at 8 years with the additional maintenance rituximab.
Maintenance rituximab has been shown to benefit patients who did not receive chemotherapy and rituximab as part of initial treatment. A randomized trial of maintenance rituximab versus observation after cyclophosphamide, vincristine, and prednisone (CVP) with the majority having FL, reported that patients who received maintenance rituximab had improved rates of 3-year progression-free survival (68 vs. 33%). Survival rates were similar between the two groups . With the current paradigm of treating patients with chemotherapy plus rituximab, this study has less practical impact.
The use of maintenance rituximab after chemoimmunotherapy in patients with FL has been examined in a recently reported large randomized trial . While maintenance rituximab appears to improve progression-free survival rates, toxicities, albeit tolerable, are increased and the effect on overall survival is to date unclear. The Primary Rituximab and Maintenance (PRIMA) phase III intergroup trial in 1018 patients with previously untreated FL responded to chemoimmunotherapy (CVP-R, CHOP-R or FCM-R) randomly assigned maintenance with rituximab (375 mg/m2 every 8 weeks for 24 months) or placebo . At a median follow-up of 36 months from randomization, patients assigned to rituximab maintenance had a higher rates of progression-free survival (75 vs. 58%). A higher percentage of patients in complete response/CRu at 24 months (72 vs. 52%) 2 years post-randomization. There was a significantly higher percentage of patients with Grade III/IV adverse events and infections in the rituximab maintenance group. At this time, overall survival is the same in both groups.
Radioimmunotherapy alone has been used as the initial treatment in a limited number of patients with FL. 131I-tositumomab was given to 76 previously untreated patients with FL. The overall and complete response rates of 95 and 75%, respectively, and at 5-years, the OS and PFS was 89 and 59%, respectively .
Radioimmunotherapy has also been applied as consolidation following conventional chemotherapy induction in patients with FL. Both 90Yi-ibritumomab tiuxetan and 131I-tositumomab have been studied. This approach has been associated with very high response rates, conversions of PR to CR, and well maintained responses [54–58]. Probably the most significant study is a phase III trial comparing 90Yi-ibritumomab tiuxetan to observation following a CR or PR to induction chemotherapy for treatment naïve patients with FL . Of note, the majority of patients did not receive rituximab along with the induction chemotherapy. With a median follow-up of 3.5 years, both the PR and CR patients who received Yi90-ibritumomab tiuxetan had significantly longer median progression-free survival with improvement of about 24 months.
High-dose therapy and autologous stem cell transplantation has been used to consolidate first remission for patients with FL. These studies generally preceded the widespread use of rituximab. In general, about 50% of patients are disease free at 10 or more yrs following ASCT, but the risk of second malignancies both MDS, AML, and solid tumors has been observed with long follow up of these patients. Several randomized trials have examined the role of ASCT in previously untreated patients with FL following an induction therapy [60–64]. The majority of these studies have demonstrated a significant improvement in PFS, but no impact on overall survival. One reason for the lack of impact on overall survival has been the excess number of second malignancies. The findings in these studies are summarized in Table III.
Table III. Phase III Trials of AutoSCT in Previously Untreated Patients with Follicular NHL
|GLSG ||CHOP + ASCT (TBI) vs. CHOP + INF-α||Improved PFS, not OS|
|GOELAMS ||VCAP + ASCT (TBI) vs. CHVP + INF-α||improved EFS (FLIPI > 2), not OS|
|GELA ||CHOP + ASCT (TBI) vs. CHVP + INF-α||No diff EFS/OS|
|GITMO ||CHOP-R vs. R-high dose sequential therapy||Improved PFS, not OS|
Although allogeneic stem cell transplantation (HCT) can potentially lead to cure for patients with FL, due to the significant treatment related mortality, this is largely reserved for patients with relapsed and more refractory disease.
Part of the natural history of indolent NHL is progression to a higher grade histologic subtype, such as diffuse large B-cell lymphoma [7, 8]. A subgroup of patients with indolent NHL who transform to a more aggressive histology may attain complete remission following treatment with CHOP-like chemotherapy and some may be cured by high-dose chemotherapy followed by autologous hematopoietic cell transplantation.
Treatment of Relapsed FL
When patients with relapsed FL require treatment, there are a broad number of options, ranging from rituximab alone to combination chemotherapy plus rituximab, radioimmunotherapy and for selected patients stem cell transplantation.
A recent update of single agent rituximab therapy in patients with relapsed FL is from the SAKK trial. In that study patients with either newly diagnosed, relapsed or refractory FL were treated with four weekly doses of rituximab . Patients with either responding or stable disease at week 12 were randomized to observation or maintenance with one dose every 2 months for 4 doses. With long follow-up, 35% of responders remain in remission at 8 years. However, in the context of current induction therapy that includes chemotherapy and rituximab in the majority of patients, it is uncertain if the response data to single agent rituximab is as high or durable as in patients who did not receive rituximab plus chemotherapy induction. There is a evidence however that retreatment with rituximab in patients with relapsed, largely FL, who had previously responded to rituximab had a response rate of 40%, following retreatment, and the median time to progression of 18 months .
The combination of chemotherapy and rituximab has enhanced the efficacy of treatment of relapsed FL. Probably the largest study treated selected patients with relapsed FL who were previously not treated with an anthracycline or rituximab containing regimen [66, 67]. Patients were randomized to CHOP or CHOP-R and responding patients were randomized to 2 years of maintenance rituximab or observation. The overall and CR rates were significantly improved in the CHOP-R group, and the median PFS was improved by approximately 12 months. A recent update of this study now with a median follow-up of 6 years reported that maintenance rituximab also improved median progression-free survival by 2.4 years. The overall survival at 5 years following maintenance was 74% versus 64% with observation alone.
Another regimen in which a benefit for the addition of rituximab was seen for relapsed disease in a randomized trial employing fludarabine, cyclophosphamide, and mitoxantrone (FCM) . A number of phase II trials of other agents plus rituximab associated with quite high response rates included bendamustine plus rituximab with 90% RR and median PFS of 2 years [69–71]. Single agent bendamustine has an overall response rate of 77 with a median response duration of 6.7 months . The regimen FCR (fludarabine, cyclophosphamide, and rituximab) has a similarly high response rate but with significant myelosuppression . Phase II studies employing bortezomib and rituximab and bortezomib, rituximab and bendamustine have reported RR of approximately 50 and 93% respectively .
Two anti-CD20 radioimmunotherapy agents are FDA approved for treatment of patients with relapsed and refractory FL . The response rates in this patient population are similar with both agents, with 60–80% of patients responding. The median progression-free survival is about 12 months, although the approximately 20–37% of patients who achieve a CR have a median time to progression of approximately 4 years [75, 76]. A randomized trial compared single agent rituximab to 90Yi-ibritumomab in patients with relapsed indolent NHL (predominantly FL) . The overall and complete response rates were significantly higher with RIT, but no difference in time to progression or OS was observed. Retreatment with these agents remains controversial with uncertainty of delivery of full dose and concerns of second malignancies .
FL is extremely responsive to radiation therapy (RT); low dose RT (eg, total dose of 4 Gy) can be used for the palliation of patients who have symptoms related to a single disease site [79, 80].
The use of either autologous or allogeneic hematopoietic cell transplantation (HCT) in FL is controversial and the subject of numerous clinical trials. A large number of Phase II studies prior to the availability of rituximab, involving high-dose therapy and autologous HCT have shown that approximately 40% of patients with good performance status and chemosensitive relapsed disease may experience prolong progression-free and overall survival rates [81–85]. Prior to the widespread use of rituixmab for in vivo purging many strategies were taken to render the autologous stem cell collections free of lymphoma cells. Although single institution studies suggested that reinfusion of tumor free stem cells led to a decreased relapse rate, it remains controversial as to whether there is a benefit particularly now in the rituximab era. The only Phase III randomized trial (the CUP trial) comparing ASCT to conventional chemotherapy in relapsed patients with FL demonstrated a higher PFS and OS for ASCT, and no benefit for purging the stem cell graft . An retrospective analysis of patients undergoing ASCT following rituximab-based salvage therapy did not suggest a benefit of ASCT as compared to conventional therapy. Unfortunately as has been seen in ASCT in first remission, second malignancies, both solid tumors and MDS/AML are reported post ASCT.
Phase II studies have looked at the use of in vivo purging pre-ASCT and maintenance therapy with rituximab following ASCT in patients with relapsed FL. These suggest an improvement in PFS, similar to what has been seen following conventional chemotherapy and chemoimmunotherapy. A recent Phase III trial in patients with relapsed FL has investigated the inclusion of rituximab for in vivo purging pre-ASCT and 2 yrs of maintenance post-ASCT . An interim analysis demonstrated improvement in PFS for patients receiving rituximab for in vivo purging, maintenance and the combination of both as compared to no rituximab.
Allogeneic SCT has been investigated in patients with relapsed FL. Both myeloablative and reduced intensity conditioning (RIC) approaches have been used. Unfortunately myeloablative conditioning has a treatment-related mortality of up to 40%; however, the relapse rate is less than 20% . Enthusiasm for RIC allogeneic has lower treatment related mortality, [89–91], but some reports suggest that the relapse rate may be higher than conventional myeloablative conditioning. The role of allogeneic SCT for FL remains uncertain but certainly for younger patients with more resistant disease, it remains a potentially curative option for relapsed FL.
There are a multitude of new approaches that are being studied in patients with FL. This includes monoclonal antibodies, idiotype vaccines, immunomodulatory agents, and novel drugs such as kinase inhibitors.
Monoclonal antibodies directed against other B-cell-associated antigens and new anti-CD20 mAbs are being investigated in FL. The anti-CD80 mAb galiximab has an overall RR of 66% and complete response rate 19%, in patients with relapsed FL . A Phase III trial, which was then converted to a Phase II trial, reported that the median PFS was 12.0 months for R+G and 9.0 months for R+P. Overall response rate was 51% for R+G vs 48% for R+P (P = 0.455) and complete response was 20% for R+G and 15% for R+P (P = 0.251) . The notion of putting more than one antibody together has also been explored with epratuzumab, a humanized anti-CD22 mAb . The antibody alone has single agent activity and when combined with rituximab has RR of 54% in rituximab naïve patients and 64% in patients who have had prior rituximab. In treatment of naïve patients, the combination of rituximab and epratuzumab had an ORR of 84% . The anti-CD40 mAb HCD122 has a ORR of 27% in relapsed FL patients . This antibody blocks the CD40-CD40 ligand signal from T cells and mediates ADCC. Several new anti-CD20 monoclonal antibodies are being evaluated in patients with FL who are refractory to rituximab. These include several humanized antibodies that are designed to have less infusion toxicity and better ADCC effector function [97–99]. The other mAb that has been of interest is GA101, the first type II, glycoengineered and humanized monoclonal anti-CD20 antibody . In rituximab refractory patients in the high-dose cohort, the RR was 50% with median PFS or 11 months. Further studies are ongoing to better define activity of these agents in combination with chemotherapy, to better define the utility of these mAbs.
A number of immunostimulatory agents have been studied to enhance the activity of rituximab. These include cytokines such as IL-2, and immunostimulatory DNA sequences known as CpGs [101–103]. To date, although having immunomodulatory effects, the impact on enhancing the therapeutic effect of rituximab has been limited. A Phase II study of lenilidomide plus rituximab has reported high response rates, but Phase III study will be needed to demonstrate superiority over rituximab alone .
The other area of interest has been in active immunization, focusing largely on the idiotype protein as the antigen. To date, there have been three randomized studies using idiotype proteins coupled to KLH following induction of remission in patients with follicular NHL. The Favrille trial used rituximab for induction therapy. The median time to progression (TTP) was 9 months for the Idiotype-KLH vaccinated patients and 12.6 months in the control group (P = 0.019) . However, this difference was attributed to more patients with high-risk FLIPI scores in the Id-KLH arm. The Biovax study has preliminarily reported showing a slight improvement in PFS for the Id-KLH vaccinated patients as compared to control; however, the induction chemotherapy was intense and remissions had to be sustained for 12 months prior to initiation of vaccination . The final trial using the MyVax Id-KLH conjugate awaits final reporting. On the basis of these studies, it is unlikely at the present time that idiotype vaccination will be pursued in follicular NHL.
B-cell kinases are logical targets for therapy in FL. To date, three kinase inhibitors, CAL101, PCI-32765, and R788, that target PI3 kinase p110δ, btk, and syk, respectively. In patients with relapsed and refractory FL, the response rates to CAL101 , PCI-32765 , and R788  were 62% (including other indolent NHLs besides FL, 27%, and 10%, respectively). These agents are undergoing additional study, in combination with chemotherapy and as consolidation following remission induction to better define their role. They represent novel approaches to lymphoma therapy and are perhaps some of the more exciting agents currently being studied.