Watchful Waiting As Initial Management of Advanced-Stage Follicular Lymphoma in the Rituximab Era: Analysis of the National Cancer Data Base

Introduction: National guidelines recommend observation without active therapy ("watchful waiting", WW) for follicular lymphoma (FL) presenting in advanced stage with low burden of disease and without symptoms. The national LymphoCare registry study reported that 17.7% of all FL patients in the United States (US) are initially managed with observation (Friedberg et al., J Clin Oncol 2009). Some recent trials suggest that early therapy with rituximab in this setting may have benefits, but this strategy remains highly controversial (Ardeshna et al., Lancet Oncol 2014). It is unknown whether socioeconomic or health care related-factors affect the decision how to manage FL. The aim of our study was to evaluate trends in the use of WW as initial management strategy for FL in the US, and associated factors.

Methods: We selected stage III/IV FL patients diagnosed in 2004-2012 from the National Cancer Data Base (NCDB)-an oncology outcomes data base capturing >70% of incident cancers in the US. We excluded cases with B-symptoms, medical contraindications to chemotherapy or radiation, unknown treatment status, or <1 month of survival. WW was defined as no recommendation for chemotherapy or radiation therapy within 100 days from FL diagnosis. Patients for whom treatment was recommended but not administered because of refusal or other reasons were not designated as WW. We analyzed trends by joinpoint regression, and the association between WW strategy and available factors by a mixed-effects logistic regression model clustered on hospital, reporting odds ratios (OR) and intra-class correlation (ICC) with 95% confidence intervals (CI).

Results: The analytic cohort of 18,783 advanced-stage FL patients without B symptoms or medical contraindications to treatment had median age at diagnosis of 63 years (range, 18 to 90), with 52% women and predominantly white non-Hispanic ethnicity (82%). In this population, 31.4% were initially managed without active treatment. This proportion remained unchanged between 2004 and 2012 (annual percent change, -0.3% per year, P = 0.23). Among patients who received chemotherapy (which in 75% was multi-agent), 89% started it within 100 days of diagnosis, and 94% within 6 months.

Patient-related factors significantly associated with WW included: age >65 years (OR, 1.23 versus <50 years, CI, 1.07-1.41) or ≥75 years (OR, 1.25, CI, 1.06-1.42), female sex (OR, 1.07, CI, 1.0-1.15), and a history of previous cancer (OR, 1.34, CI, 1.22-1.48). Lymphoma-related factors included stage IV (versus III, OR, 0.61, CI, 0.57-0.65), grade 2 (versus 1, OR 0.70, CI, 0.63-0.76) or grade 3 (OR, 0.28, CI, 0.25-0.32). There was a significant variation by geographic region and treating hospital. The use of WW ranged from 39% in New England and on the West Coast, to 28% in the Midwest and 22% in the South (joint P <0.0001). The use also significantly varied by hospital (ICC, 6.8%, CI, 5.5-8.3%), and was significantly higher in academic/research cancer programs than in community programs (OR, 1.31, CI, 1.12-1.53). There was no significant association between WW and race, number of comorbidities, median income, type of health insurance, or distance from treatment facility. Unadjusted overall survival at 5 years for patients who did not require upfront treatment, as anticipated, was better (76.9%, CI, 75.5-78.2%) than for those who needed therapy at diagnosis (74.3%, CI, 73.4-75.2%), but the difference was not significant after adjustment for confounders and excluding early deaths within 3 months of diagnosis (hazard ratio, 0.93, CI, 0.86-1.01).

Conclusions: Despite guidelines,only one third of patients with advanced-stage FL in the US are managed by WW. Patient selection appears to primarily occur on clinical grounds, as indicated by the association with grade and stage (factors correlating with disease burden). However, significant variation by region and hospital, and higher rate of WW in academic hospitals suggests that decision-making may be influenced by physician expertise or payment models in oncology practices, offering an opportunity for education or improved health care policy. However, considering lack of direct clinical assessments of tumor burden, cytopenias and symptoms in the NCDB, evaluating this requires further research. Clinical trials are unlikely to detect a survival difference between WW and active treatment in FL within 5 years of follow-up.