In this issue of Blood, Luminari et al demonstrate that patients with marginal zone lymphoma (MZL) who experience early progression (progression of disease at 24 months [POD24]) have poor survival.1 Overall survival after POD24 was 53% at 3 years, a stark contrast to the 95% 5-year survival rate from diagnosis in patients without early progression. These results extend the seminal study in follicular lymphoma (FL) from the National LymphoCare Study by Casulo et al2 to include indolent nonfollicular B-cell lymphomas (INFLs).
Outcomes in FL and INFLs have improved greatly in the rituximab era, and 10-year survival estimates range from 60% to 80%, depending on subtype.3,4 Although progression of a malignancy is never desirable, periods of remission, progression, and retreatment are expected in the disease course of indolent lymphomas. Despite a rapid increase in therapy options in the last decade, there remains a subset of patients who will transform to a more aggressive subtype or become refractory to available therapies; early identification of these patients is of great clinical interest. Prognostic indices at diagnosis are readily available for both MZL5 and FL6 ; however, the initial diagnostic variables become less relevant as patients move through their disease course. Dynamic assessment of prognosis based on updated information is critical to provide accurate risk assessment and optimal disease management.
The data from the NF10 registry presented by Luminari and colleagues provide important information for dynamic assessment of prognosis in MZL by identifying a high-risk population consisting of the 18% of patients who experienced early progression. Poor outcomes after POD24 were observed across marginal zone subtypes, with 3-year survival rates after early progression ranging from 33% for disseminated MZL to 71% for extranodal MZL, although the sample size limits firm subtype-specific conclusions. Just as vital, patients in the NF10 study who did not suffer early progression from initial systemic therapy had excellent outcomes, with 5-year survival rates from diagnosis ranging from 93% to 98% among the various marginal zone subtypes. These results confirm previous results in FL. Moreover, previous studies have shown that survival for patients without early progression in FL7 and INFL3 is similar to that of the background population over the next 5 to 10 years. These data provide powerful tools for the re-assessment of prognosis for clinicians as well as important reassurance for patients at their 2-year follow-up visits.
There are some important caveats that need to be considered. Indolent lymphomas are heterogeneous in their presentation and initial management. The POD24 analysis by Luminari et al was limited to patients who required immediate therapy, which made up only 24% of the NF10 INFL cohort. Similarly, the NLCS study in FL was based on the most aggressively treated patients. Using a 24-month cutoff may not be optimal for less aggressive management approaches.3,7 Statistically, use of a dichotomous end point like POD24 always involves a loss of information, and the prognostic impact of early events may vary based on the timing of the event.8 The role of early transformation versus early indolent progression or relapse events is also in need of further evaluation,9 as is the impact of maintenance therapy on early event-related prognosis. The utility of early events will require periodic reassessment as the management strategy and therapy options for patients with FL and INFL continue to evolve. However, from a research standpoint, the use of these end points can facilitate more rapid identification of high-risk patients in correlative studies of FL and INFL. Clinical and/or biologic features of these high-risk patients can then be used to generate predictive models at diagnosis to help aid clinical risk prediction and first-line therapy selection.
Finally, for the treating physician, it is important to remember that early progression end points such as POD24 are ways of identifying high-risk patients after they have been exposed to initial therapy. They are not meant to be therapy goals at the time of initial management for the individual patient. Therapy selection should still be made to maximize outcomes over the life of the patient, not to maximize achieving these surrogate outcomes.
Conflict-of-interest disclosure: M.J.M. has received research funding from Celgene and Nanostring and serves on an advisory board for Morphosys.