Is watch and wait still acceptable for patients with low-grade follicular lymphoma?

In their original description of follicular lymphoma (FL), or nodular lymphoma, in the 1950s, Rappaport recognized that patients with these illnesses could be subdivided based on the proportion of large cells in the tumor.¹  Although the names have changed over time, the World Health Organization (WHO) classification continues this distinction by dividing FL into grade 1, grade 2, and grade 3.²  It has been recognized for some time that a subset of patients with what is today called grade 3 FL has a more aggressive illness and one that might have durable responses to therapy used for diffuse large B-cell lymphoma. For example, in the Working Formulation, follicular large cell was included in the intermediate-grade lymphomas, not low-grade lymphomas.³  Grade 3 FL seems to be biologically different based on a higher proliferative rate,⁴  a higher median standardized uptake value on positron emission tomography scans (Andrew Zelenetz, Memorial Sloan Kettering Cancer Center, oral communication, March 30, 2016), and different immunological characteristics.^5,6  Some reports,^7-10  but not all,¹¹  have found that a subset of these patients has a plateau on their survival curve when treated with anthracycline containing combination chemotherapy regimens.

The WHO has tried to recognize the fact that some patients with grade 3 FL have an illness that behaves more like diffuse large B-cell lymphoma by subdividing grade 3 into grade 3A and 3B (ie, sheets of centroblasts), with 3B representing those patients who should receive aggressive therapy.²  The difficulty here is that it is not clear that pathologists can make this distinction consistently. For example, a study measuring pathologists’ ability to consistently divide patients into grade 1, grade 2, and grade 3 showed a success rate of only 60% to 70%.¹²  This does not yield much confidence that they will be able to subdivide grade 3 consistently. For this reason, some have elected to treat patients with grade 3 FL in a manner similar to patients with diffuse large B-cell lymphoma. Certainly, this will mean some patients are overtreated, but it will not lead to withholding aggressive therapy from a patient who might have the possibility of being cured. In the remainder of this paper, low-grade FL will represent patients with grade 1, grade 2, and, if it is reported, grade 3A FL.

It has also become apparent that FL cells have a complex interrelationship with infiltrating immune cells. Initial reports showed that tumors with a high proportion of infiltrating T cells had a superior survival to those with higher proportion of monocyte/macrophages in the tumor.^13,14  Initial reports of gene profiling in FL confirmed this finding and found a strikingly better survival in patients whose lymphomas had a profile suggesting a higher proportion of infiltrating T cells.¹⁵  More recently, it has been shown that immune checkpoint inhibitors cause responses in a proportion of patients with advanced/refractory FL.¹⁶ 

For many years, it appeared that the median survival of patients with low-grade FL was approximately 10 years and that it might not be greatly impacted by therapy.^17-20  However, more recently, numerous reports have suggested a significant improvement in the survival of patients with low-grade FL.^21-23  The survival of patients with low-grade FL treated by physicians in the Nebraska Lymphoma Study Group diagnosed before or after the year 2000, and those who never received rituximab and those who did, suggests that survival is significantly changing and that the availability of rituximab might be an important component in this improvement (Figure 1).

The increasing evidence that survival is improving for patients with low-grade FL, and that new therapies, particularly monoclonal antibodies, might be responsible for this improvement has led to suggestions that it is no longer appropriate to observe these patients without therapy.

In the 1970s, it was first noted that some patients who underwent biopsies of distinct sites of disease at diagnosis often had different histologic features in different sites: follicular growth pattern in one and diffuse large B-cell lymphoma in another.²⁴  It was also noted that patients with FL could undergo histologic transformation to diffuse large B-cell lymphoma through the course of their disease.^25,26  This histologic transformation was heralded by rapid change in the size of one or more lymph nodes and accompanied by a dramatic acceleration in the rate of disease progression. The progression occurred independently of how the patient was treated even being noted in patients who had received no initial therapy.²⁷  Indeed, an autopsy study of patients with FL who died with lymphoma showed that 90% of the patients had transformed to diffuse lymphoma.²⁸  A number of groups have more recently reported on the rate of histologic transformation, and a consensus emerged that the rate was not clearly affected by type of treatment.²⁹  Recent estimates suggest that the rate of histologic transformation may be 2% to 3% per year.^29-31  However, the explanation for the lower incidence remains unclear.

The clinical acceleration of the disease course after transformation was a universal finding, but results of treatment after transformation were not uniform in different research groups. At the National Cancer Institute, complete response rates were comparable to de novo diffuse large B-cell lymphoma, and some patients were long-term disease-free survivors.²⁶  By contrast, none of the patients reported from the University of Iowa obtained a complete response, and survival was only a few months.²⁵  More recent reports of outcome after transformation suggest that both the response rate and survival have improved considerably.^29-31 

Despite the considerable efforts to define transformation in molecular terms, no unified picture has yet emerged. The treating physician has no readily available information to predict transformation or prevent it.

In his 1984 Karnofsky Memorial Lecture, Saul Rosenberg reviewed the results of patients with low-grade non-Hodgkin lymphoma treated at Stanford University between 1961 and 1982.¹⁷  One important observation was the occurrence of spontaneous regression in some patients who were followed without therapy. This occurred in 30% of patients with follicular small cleaved cell (ie, analogous to grade 1 FL) non-Hodgkin lymphoma. Eighty-three patients with low-grade lymphoma were managed without therapy until symptomatic progression of disease. The median survival of these patients was 11 years (ie, better for follicular small cleaved lymphoma with 80% surviving to 17 years). The median time to requiring therapy was 16.5 months for follicular mixed (ie, presumably grade 2 FL) and 48 months for follicular small cleaved. This same group subsequently reported the outcome of no initial therapy in patients with stage I and II low-grade FL. They described 43 patients of whom 63% still had not been treated at a median follow-up of 86 months.³²  The estimated survival at 10 years was 85%, not different from patients who received immediate treatment.

One of the authors of this paper was involved in a randomized trial of patients with low-grade non-Hodgkin lymphoma (ie, predominantly patients with low-grade FL) where patients were randomly assigned to watch and wait or to a very aggressive combination chemotherapy regimen (ie, ProMACE-MOPP [methotrexate, doxorubicin, cyclophosphamide, etoposide, prednisone, mechlorethamine, vincristine, procarbazine]) followed by total nodal radiotherapy. Patients in the watch-and-wait group could receive involved field radiotherapy for local symptoms. The updated results of this study with a median follow-up of nearly 20 years show that there is no significant difference in overall survival between those randomly assigned to watch and wait vs those receiving aggressive combined modality at diagnosis (D.L.L., unpublished data). Overall survival in both arms at 25 years is about 28%. Three patients assigned to watch and wait have never been treated, whereas the others received ProMACE-MOPP at progression. Some differences in the 2 groups have emerged. For example, the rate of undergoing histologic progression is slightly higher (26% vs 12%, P = .06) for those initially assigned to watch and wait. Nevertheless, there has been no overall survival difference.

In 2003, the British National Lymphoma Investigation reported a trial where patients with asymptomatic, advanced-stage, low-grade lymphoma were randomized to receive immediate treatment with chlorambucil 10 mg daily continuously or observation until symptomatic disease progression. In both groups, radiotherapy to symptomatic lymph node sites was allowed.³³  At a median follow-up of 16 years, overall survival or death from lymphoma did not differ between the 2 groups, with a median survival of 5.9 years for patients treated with chlorambucil and 6.7 years for patients initially managed with observation. Of the patients assigned to observation without therapy, 19% had not received systemic treatment after 10 years of follow-up.

In 2012, a group of investigators from Europe and the United States described 120 patients with low-tumor-burden FL who were initially managed with watch and wait.³⁴  The decision to observe the patients initially was made by individual physicians and their patients. The median time to initiating therapy was 64 months. The only observation at presentation that predicted being more likely to be treated were 4 or more nodal areas involved with disease. The 4-year freedom from treatment failure for the watch-and-wait patients (ie, 79%) was not inferior to a subgroup of patients with similar prognostic characteristics who were initially treated with a rituximab-based regimen (69%), The 5-year overall survival of patients treated with initial observation was 87% and did not differ from similar patients who were treated initially (ie, 88%).

In 2014, investigators from the United Kingdom reported a randomized trial of patients with advanced-stage, asymptomatic, nonbulky FL where patients were initially assigned to watch and wait, 4 weekly doses of rituximab, or 4 weekly doses of rituximab followed by rituximab maintenance for 2 years.³⁵  The rituximab induction without maintenance arm was closed early. The study found a significant difference in the time to start a new treatment with 54% of the patients in the watch-and-wait group being treated by 3 years in contrast to 12% of the patients in the group that received rituximab followed by maintenance therapy. Patients in the watch-and-wait group had poorer results in the Mental Adjustment to Cancer Scale score and the Illness Coping Style score between baseline and month 7. However, there was no significant difference between the 2 groups in chance of death, with 9% of the patients in the watchful waiting group and 6% of the patients in the maintenance rituximab group having died. The 3-year overall survival was 94% for watchful waiting and 97% for the maintenance rituximab group.

In 2015, investigators from Denmark described 286 patients with advanced-stage FL initially seen between 2000 and 2011 who were managed with watch and wait.³⁶  The 5-year progression-free survival was 35% and the 10-year overall survival was 65%. The cumulative risk of dying from lymphoma within 10 years was 13%. The 10-year risk of histologic transformation was 22%. The authors felt that abandoning watch and wait as a management strategy would lead to overtreatment of selected patients.

The National Comprehensive Cancer Network (NCCN) guidelines for the therapy of patients with FL with extensive stage II and stage III/IV include watch and wait as an acceptable initial management approach for patients with low-grade FL who are asymptomatic, have no threatened end-organ function, do not have cytopenias secondary to the lymphoma, or lack bulky disease, and in whom the disease is not steadily progressing.³⁷  For patients with localized low-grade FL, the NCCN guidelines also include watch and wait as being acceptable management in selected cases in whom the potential toxicity of involved field radiotherapy outweighs the hoped-for clinical benefit.

The British Society of Hematology has approved management guidelines for patients with FL.³⁸  The recommendations include watch and wait as an appropriate approach for patients with asymptomatic, advanced-stage FL, particularly for patients over 70 years of age. The authors also included observation as a possible option for selected patients with localized FL in whom total excision of all macroscopic disease had been done.

The Lymphoma Canadian Scientific Advisory Committee issued consensus findings on the management of patients with FL.³⁹  The authors felt that watch and wait was appropriate for patients with asymptomatic, advanced-stage FL who did not have large masses, bulky disease, symptomatic splenomegaly, impending organ compromise, cytopenias secondary to bone marrow infiltration, or anxiety without treatment. The authors felt that watch and wait in patients with localized disease was controversial, but might be reasonable in selected cases.

The Italian Society of Hematology, the Societa Italiana di Ematologia Spermentale, and the Gruppo Italiano Trapianto Midollo Osseo have published joint guidelines for the management of patients with FL.⁴⁰  The authors felt that asymptomatic patients with stage II-IV, nonbulky disease should be offered watchful waiting as the initial treatment approach. However, they felt that, for patients with localized disease, watchful waiting was not recommended except for patients with a short life expectancy or contraindications to radiotherapy.

The American College of Radiology made recommendations for the treatment of patients with localized FL.⁴¹  After reviewing the published results with watch and wait and radiotherapy, they concluded that there is a high death rate among patients who progress following watch and wait. They recommended that watch and wait be reserved for selected patients who are ineligible for definitive radiotherapy.

The European Society of Medical Oncology has made treatment guidelines for patients with FL. They recommended radiotherapy for patients with limited nonbulky stage I and II. For early-stage patients with high tumor burden, they recommended the same treatment as for patients with stage III and IV. For advanced-disease patients, they recommended initiating therapy when symptoms develop and observing patients who are asymptomatic.⁴² 

Given the lack of proof from a prospective trial that there is a significant prolongation of survival for patients with asymptomatic, non–life-threatening low-grade FL with immediate therapy, we believe that a reasonable physician can make the recommendation for watch and wait—and we sometimes do. However, there are caveats in making this recommendation (Table 1). A physician’s enthusiasm for watch and wait in this setting should not override the patient’s preference for therapy based on anxiety about not treating a known cancer. The fact that there is no proof that delaying therapy does not decrease survival does not make it impossible that it is still true. Studies in FL take extended periods of time, and most of the studies suggesting no impact on survival in delaying therapy were done in the time before the availability of monoclonal antibodies.

It would be better if an objective method using a biomarker could be used to choose patients for the watch-and-wait approach. It is possible that identification of patients most likely to benefit, and least likely to be harmed, by observation without therapy for newly diagnosed FL will be able to be identified more definitively in the future. In a recent report in Lancet Oncology,⁴³  the authors tried to improve on the FL international prognostic index^44,45  by incorporating gene mutations into the prognostic model. They examined patients who received first-line chemoimmunotherapy rather than all patients. The use of biological markers improved the ability to predict treatment outcome. It was particularly good at identifying those patients with a bad outlook. It is reasonable to hope that utilizing similar approaches from all patients at the time of diagnosis might be able to identify those unlikely to progress rapidly and for whom watch and wait might be a favored initial approach. However, until that time, the lack of symptoms, the absence of disease involvement in sites that might be imminently dangerous, and a patient’s desire to avoid therapy as long as possible represent the most practical way to identify those patients for whom watch and wait is a reasonable treatment option.

Contribution: J.O.A. and D.L.L. wrote the paper.

Conflict-of-interest disclosure: J.O.A. is a consultant for Celgene, Roche, Spectrum, Ziopharm, Conatus Independent Data Monitoring Committee, GlaxoSmith Kline Independent Data Monitoring Committee, and is a member of the board of directors for Tesaro bio, Inc. D.L.L. declares no competing financial interests.

Correspondence: James O. Armitage, University of Nebraska Medical Center, 987680 Nebraska Medical Center, 8722 LTC, Omaha, NE 68198-7680; e-mail: joarmita@unmc.edu.