LAUSR

In their recent perspective, Cheah and Seymour [1] assessed the scarce evidence available on treatment criteria for relapsed follicular lymphoma (FL). We applaud their effort in broaching such a difficult subject because, in contrast to treatment-naïve patients, high-quality evidence is largely lacking in the relapsed setting. In the introduction, the authors claim that ‘...the natural history [of FL] is marked by successive lines of therapy resulting in progressively shorter periods of disease-free survival followed ultimately by the development of either chemo-refractoriness, large cell transformation, or death from treatment related toxicities.’ This claim, which we have made in the past ourselves [2], is supported by classical studies such as those by Johnson et al. [3], Weisdorf et al. [4], and Gallagher et al. [5], which included mostly patients treated with single alkylating agents or alkylator-based chemotherapy. At present, it has become clear that FL is a very heterogeneous disease, and that the different prognostic subgroups can largely be drawn based on response to first-line immunochemotherapy (ICT). Thus, patients with poor response to frontline ICT have a poorer response to subsequent lines of therapy as well as a shorter overall survival (OS) [6–8], and they also seem to have notably higher rates of histological transformation (HT). Although it is unclear whether there is a late plateau in the HT curves [9]; HT is more frequent [7,8] and aggressive [10,11] shortly after diagnosis. Conversely, patients with remission persisting 2 years after diagnosis (or perhaps as early as 1 year) can expect to live as long as the general population [12], which seems to indicate that their risk of death from either the disease or toxicity is lower. In a recent study analyzing the outcome of FL patients based on their response to front-line ICT [7], we found that most patients with ICT-refractory disease were dead at the time of analysis (31/53, 58%), mostly due to lymphoma (27/31, 87%) while only a minority of ICT-sensitive patients had died (35/290, 12%), 14 (40%) due to lymphoma and 10 (29%) due to toxicity (unpublished data). Importantly, however, a majority of patients remained alive. With this background, we believe it would be interesting to assess lymphoma-specific survival in studies with long-term follow-up. Lymphoma-specific survival is the time from diagnosis until death due to lymphoma (although it is also important to record deaths due to toxicity [13]). While progression-free survival (PFS; the time from diagnosis until progression or death by any cause) and OS (the time from diagnosis until death by any cause) remain the most relevant and informative endpoints, it is also essential to learnwhat patients with FL currently die from (as compared to the general population), in order to assess whether any specific early interventions or targeted follow-up are warranted. The patients in our study died mostly of lymphoma or treatment-related toxicities, but the median follow-up was relatively short (4 years in living patients), and it is conceivable that causes of death differ in patients with longer remissions. Indeed, in a recent study, El-Galaly et al. [11] estimated death due to lymphoma to be 13% at 10 years, although themedian follow-up of their studywas less than 6 years and the entire populationwas managed initially by watchful waiting. One of the concerns of lymphoma-specific survival is the problem of determining the cause of death [13], particularly in patients with long remissions and those lost to follow-up. Similarly, late infectious complications, secondary cancers, and cardiovascular deaths may be only partly due to treatment toxicity. However, other endpoints also involve a best medical guess. In particular, PFS puts a specific date on progression, when that is a continuous process. Furthermore, progression can be detected earlier or later depending on how close the follow-up visits are or how sensitive (PET vs. CT scans) or frequent image testing is carried out. Although flawed, lymphoma-specific survival might also provide important clues into a relevant problem.