Central nervous system prophylaxis in patients with diffuse large B cell lymphoma, are we treating ourselves? A response to the recent BCSH Guideline.

Abstract

We were interested to read the recent British Committee for Standards in Haematology (BSCH) guidelines on the role of central nervous system (CNS) prophylaxis in the management of lymphoma. The authors concluded that patients with diffuse large B cell lymphoma (DLBCL), International Prognostic Index (IPI) stage 3–5 with either a raised lactate dehydrogenase (LDH) and more than one extra nodal site, or involvement of certain anatomical sites (breast, testicular, epidural) should be treated with 3–6 doses of intrathecal methotrexate. The paucity of evidence for all aspects of this guidance was acknowledged (McMillan et al, 2013). The North Trent Cancer Network covers a population of approximately 1·9 million in the cities/towns of Sheffield, Doncaster, Rotherham, Barnsley and Chesterfield. After a review of the available evidence several years ago, the clinicians in the network made a unanimous decision not to administer CNS prophylaxis to patients with DLBCL receiving R-CHOP (rituximab, cyclophosphamide, doxorubicin, vincristine, prednisone) chemotherapy. More recently, it was agreed that patients with testicular lymphoma should receive high dose systemic methotrexate as part of their treatment regimen with R-CHOP. The diagnoses of all new cases of DLBCL in the region are made by the Sheffield Integrated Haemato-Oncology Diagnostic Service (HODS). Since 2006, all cases of DLBCL have been discussed at a single weekly teleconferenced region-wide multidisciplinary team (MDT). Therefore, centralized diagnostic and clinical data are readily available. We believe we are one of only a few regions who have adopted this conservative approach to CNS prophylaxis and have therefore reviewed all cases of CNS relapse over a 6-year period and assessed each case with reference to the BCSH guidelines (McMillan et al, 2013), identifying those patients who would have been eligible to receive CNS prophylaxis. Those patients managed palliatively have not been included. Review of MDT and HODS records identified 620 de novo cases of DLBCL treated with curative intent over 6 years. 12 patients were diagnosed with isolated CNS relapse and one had peripheral and CNS relapse, giving a prevalence of 2·1% (1·9% with isolated CNS relapse). Median age was 53 (range 21–79) years and the cohort included 10 male patients, three of whom had testicular disease. Nine patients had IPI < 3; three patients did not have an LDH available, resulting in an IPI of two; one other patient had an IPI of four. The three patients with testicular disease and CNS relapse were all treated with R-CHOP alone. According to current BCSH guidelines (McMillan et al, 2013) 4/13 patients would have received CNS directed treatment (three testicular disease). Eight patients died of CNS disease shortly after presentation, two are alive with progressive lymphoma, three patients are in remission (2–48 months post relapse), one of whom has significant cognitive impairment following brain radiotherapy, reiterating the poor prognosis associated with CNS relapse. In the rituximab era, the CNS relapse rate in the North Trent region is 2·1% in de novo DLBCL. Although every attempt has been made to identify patients who developed CNS relapse though database searches it is possible that patients may not have been discussed at the network MDT and therefore we acknowledge that, as a retrospective review, there maybe a degree of under-ascertainment. Although the CNS relapse rates seen here are lower than published data from the larger phase three trials, such as RICOVER-60 (Boehme et al, 2009), they are still consistent with published figures (Kumar et al, 2012). This is despite our conservative approach to CNS prophylaxis. Furthermore, less than a third of these cases would have received prophylactic intrathecal methotrexate under the new guidance. This is lower than the figures described by van Besien et al (1998), who used a similar strategy to identify those at risk of CNS relapse before the widespread use of rituximab and reported that 44% of the 605 patients meeting the criteria to receive CNS prophylaxis developed CNS relapse. The majority of the published data on CNS relapse are retrospective and of variable quality therefore reaching definitive conclusions is difficult. Evidence collected in the rituximab era suggests a reduction in the frequency of CNS relapse, perhaps because of better initial disease control (Boehme et al, 2009; Guirguis et al, 2012). The authors of the BCSH guidelines state their ‘recommendation is principally based on the low rates of CNS relapse reported by some studies using the strategy of IT chemotherapy, even though doubt still remains as to the efficacy of this practice’ (McMillan et al, 2013). We agree with the opinions expressed by Kridel and Dietrich (2011) in that the evidence for CNS prophylaxis in patients with high-risk DLBCL is limited. An alternative interpretation of the available evidence is that, with the possible exception of testicular disease, there is no reliable strategy for the identification of an individual at risk of CNS relapse and, similarly, the benefit of intrathecal methotrexate prophylaxis alone is unproven. Thus, the current data do not support the application of this guidance. We believe that the BCSH should reconsider its adoption. However, in the meantime we feel our experience may provide reassurance to others wishing to adopt alternative strategies.