Treatment patterns and survival outcomes in patients with R/R follicular lymphoma: a multicenter cohort study

Relapsed or refractory (R/R) follicular lymphoma (FL) is associated with poor outcomes which show consistent decline with increasing lines of therapy. Although patients with FL have shown a high response rate to novel treatments, such as anti-CD20 antibodies combined with chemotherapy and lenalidomide, there is considerable variation in treatment patterns and outcomes in patients with R/R FL. Patients who relapse within 2 years of immunochemotherapy or with histological transformation, as well as those who relapse multiple times, are at particularly high risk of early mortality and are therefore in need of novel treatments. Several targeted therapies including lenalidomide, enhancer of zeste homologue 2 (EZH2) inhibitors, inhibitors of phosphoinositide 3-kinase (PI3K), anti-CD20/CD3 bispecific antibodies, and anti-CD19 chimeric antigen receptor (CAR) T-cells are being evaluated in patients with FL who need therapy in the third line or later. However, the development of these therapies is limited by lack of consistent treatment endpoints and outcomes.

Casulo, et al.,¹ recently published a study in Lancet Haematol describing the treatment patterns, survival outcomes, and duration of response in patients with R/R FL using data from the Lymphoma Epidemiology of Outcomes (LEO) cohort study (NCT02736357) and the LEO Consortium. The key findings are summarized below.

Study design

This was a multicenter cohort study of adult patients aged ≥18 years, diagnosed with Grade 1–3a R/R FL between March 2002 and July 2018, who were receiving systemic therapy in the third-line setting or later after previous therapy with an anti-CD20 antibody and an alkylating agent.

The index therapy was defined as the first line of systemic therapy after the patient had received at least two previous therapies, comprising of an alkylating agent and an anti-CD20 therapy.

The primary endpoints included:

• overall response rate (ORR)
• progression-free survival (PFS)
• overall survival (OS)

Secondary endpoints comprised of:

• complete response rate
• event-free survival
• time to next treatment
• duration of response
• risk of transformation

Outcomes were also assessed in subsets for therapy class, age, gender, FL grade, Group d’Etude des Lymphomes Folliculaires (GELF) criteria, FL International Prognostic Index (FLIPI) status, lactate dehydrogenase concentration, time from diagnosis to index therapy, progression of disease within 24 months, treatment history, best response assessment, and treatment refractory status. The impact of including subsequent lines of therapy after the index therapy was assessed in a prespecified sensitivity analysis using a resampling approach.   

Results

Patient characteristics

A total of 441 patients were included, with a median time from diagnosis to index therapy of 52 months and a median age of 60 years (range, 52–68 years) at initiation of index therapy. Selected patient characteristics are shown in Table 1.   

Table 1. Patient characteristics at index therapy*

BM, bone marrow; ECOG, Eastern Co-operative Oncology Group; FLIPI, Follicular Lymphoma International Prognostic Index; GELF, Group d’Etude des Lymphomes Folliculaires; HSCT, hematopoietic stem cell transplantation; LEO, Lymphoma Epidemiology of Outcomes; PI3K, phosphoinositide 3-kinase. *Adapted from Casulo, et al.1 †Population for which data was available. ‡Any of several multiagent, dose-dense, highly myelosuppressive regimens generally delivered with the intent of consolidation with a stem-cell procedure in the event of response. In all but three patients, the salvage regimen contained ifosfamide or a platinum-based agent. §Included Bruton’s tyrosine kinase inhibitors, serine/threonine-protein kinase mTOR inhibitors, proteosome inhibitors, bispecific antibodies, histone deacetylase inhibitors, checkpoint inhibitors, polatuzumab, antibodies (anti-CD19, anti-CD22, anti-CD47, and anti-CD137), Bcl2-associated agonist of cell death inhibitors, a toll-like receptor 4 agonist, vaccines, immunomodulatory drugs, IL-15 superagonists, and other novel or investigational agents.
Characteristics, % (n/N†) (unless otherwise stated)	N = 441
Age
Mean age, years	60
18–59 years	52
≥60 years	48
Gender, Female/Male	41/59
Disease grade
1–2	86 (349/408)
3a	14 (59/408)
Disease stage
I or II	13 (47/357)
III or IV	87 (310/357)
ECOG Performance Status
0–1	95 (304/320)
2–4	5 (16/320)
FLIPI score
0–1	28 (82/294)
2	37 (109/294)
3–5	35 (103/294)
GELF criteria
≥1 GELF criteria	34
No GELF criteria reported	66
Treatment location
LEO center	64 (280/437)
Other	36 (157/437)
Index therapy class
Immunochemotherapy	30
Salvage therapy, cellular therapy, or both‡	21
Other	14
Anti-CD20 monotherapy	12
Novel therapy with or without anti-CD20§	9
Lenalidomide with or without anti-CD20	8
PI3K inhibitor with or without anti-CD20	6
HSCT
No HSCT	87
Autologous HSCT	10
Allogeneic HSCT	3
Rituximab maintenance
No	86
Yes	14
Clinical trial
Treated off clinical trial	77 (323/421)
Treated on clinical trial	23 (98/421)

Treatment pattern and outcomes

Most patients (94%) received index therapy in the third line and there was considerable variation in treatment patterns and sequencing before index therapy. Further details on the treatment patterns can be explored using the Lymphoma Epidemiology of Outcomes interactive Sankey plot.

Most patients who received bendamustine plus rituximab as their index therapy had previous R-CHOP (rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisolone) therapy. Treatment patterns over time also demonstrate the addition of bendamustine-rituximab as first line management; with 45%, 40%, and 15% of 143 patients, diagnosed between 2010 and 2018, treated with bendamustine-rituximab, R-CHOP, and R-CVP (cyclophosphamide, vincristine, and prednisolone), respectively.

Patients in the prespecified sensitivity analysis were more heavily pretreated with index therapies in fourth line or later (35% vs 6%) and had a lower 5-year OS (71% vs 75%) compared with those in the primary analysis. Factors such as presence of GELF criteria, FLIPI score at index therapy, disease progression within 24 months, and treatment refractory status retained their association to poor outcomes in the resampled datasets.

At a median follow-up of 71 months, overall and complete response rates to index therapy were 70% and 47%, respectively. Two-year PFS was 40% and 5-year OS was 75% (Table 2); 26% of patients died. Patients received a median of one subsequent line of therapy after their index therapy. 

Table 2. Primary and secondary outcomes at index therapy*

CI, confidence interval; DoR, duration of response; EFS, event-free survival; NA, not applicable; NE, not estimable; OS, overall survival; PFS, progression-free survival. *Adapted from Casulo, et al.1
Outcomes	2-year, % (95% CI)	5-year, % (95% CI)	Median, months (95%CI)
PFS	40 (36–45)	24 (20–29)	17 (15–19)
OS	90 (87–93)	75 (70–79)	167 (14–NE)
EFS	40 (35–45)	24 (20–29)	17 (15–19)
Time to next treatment	48 (43–54)	29 (24–34)	23 (19–27)
DoR (280 responders)	47 (41–54)	33 (28–40)	19 (16–29)
Transformation (cumulative incidence)	1 (1–3)	7 (5–10)	NA
Cause of death
Lymphoma	7 (5–10)	17 (14–22)	NA
Other causes	1 (1–3)	3 (1–5)	NA

Outcomes by subsets of clinical interest

High response rates were observed by therapy class with an ORR of 38% for PI3K inhibitor with or without anti-CD20; 49% for novel therapy with or without anti-CD20; 55% for lenalidomide with or without anti-CD20; 66% for anti-CD20 monotherapy; 67% for others; 76% for salvage therapy, cellular therapy or both, and 84% for immunochemotherapy. However, the median PFS was <2 years for all therapy classes, with 2-year PFS ranging from 25% (hazard ratio [HR], 1.94) for a PI3K inhibitor with or without anti-CD20 to 48% (HR, 0.79) for salvage therapy, cellular therapy, or both. The 5-year OS in patients ranged from 58% (HR, 2.29) for lenalidomide with or without anti-CD20 as their index therapy to 71% (HR, 1.38) for PI3K inhibitor with or without anti-CD20 and 81% (HR, 0.99) for other therapy classes. Lower ORR was observed in patients treated on clinical trials compared with those treated off clinical trials (65% vs 72%).

In patients aged ≥60 years, 5-year OS was significantly lower compared with those aged 18–59 years (66% vs 82%; HR, 2.17). Lower ORR (61% vs 74%), 2-year PFS (33% vs 44%; HR, 1.27) and 5-year OS (68% vs 78%; HR, 1.46) was associated with the presence of GELF criteria compared with its absence. Similarly, inferior ORR (62% vs 80%), 2-year PFS (28% vs 54%, HR 1.77), and 5-year OS (64% vs 78%; HR, 2.09) was observed in patients with FLIPI score of 3–5 compared with those who had a score of 0–1 at index therapy.

Outcomes were inferior for patients initiating index therapy within 36 months of their initial diagnosis, compared with those initiating the index therapy after more than 36 months (2-year PFS, 30% vs 45%; HR, 1.33; 5-year OS, 63% vs 81%; HR, 1.89). The inferior 5-year OS remained significant after adjusting for age and gender (HR, 1.92; 95% confidence interval [CI], 1.33–2.76). Lower ORR (68% vs 77%) and significantly lower 5-year OS (72% vs 81%; HR, 1.60; 95% CI, 1.04–2.46) was characteristic of patients refractory to therapy with an alkylating agent compared with those who were not refractory.   

Conclusion

This large cohort study showed that patients with R/R FL received heterogenous treatments in the third-line setting or later, with differences in therapy-specific overall responses. Although there were high response rates to novel therapies, these were short-lived. The 5-year OS was higher than expected and similar across most therapy classes; however, the median PFS was <2 years across all therapy classes and patients who were refractory to therapy with an alkylating agent had lower ORR and significantly lower 5-year OS. Strengths of the study were inclusion of patients from eight academic centers, independent review and classification of treatment, efficacy, cause of death, and resampled datasets for sensitivity analysis. The cohort included was also representative of patients who were likely candidates for clinical trials, considering their management at academic centers. Although the data analyzed was mostly retrospective and therefore subject to inherent bias, the findings highlight the unmet needs of patients with R/R FL and provide an evidence-based standard against which findings from clinical trials can be compared.