Diffuse large B cell lymphoma (DLBCL) is the commonest sub type of non-Hodgkin's lymphoma (NHL) accounting for 30–50 % of NHL cases. Around 2% to 10% of patients with diffuse large B-cell lymphoma (DLBCL) experience central nervous system (CNS) relapse after initial therapy which is associated with a poor prognosis and most often a fatal outcome. The incidence of CNS relapse can vary from <1% in younger, good-risk patients, to around 30% in patients with multiple risk factors, however, the relapse risk was reported to be lower in the rituximab era in some studies. Moreover, optimal modality of CNS prophylaxis remains to be defined, with both systemic and intrathecal (IT) chemotherapy being widely used. As the incidence of CNS relapse and type of prophylaxis used varies in different reports, it is important to study this risk in different populations to implement optimal prophylaxis strategies. The Objectives of this study was to evaluate the incidence of CNS relapse in DLBCL patients at our institution and to study risk factors and the type and role of CNS prophylaxis.

We retrospectively analyzed patients diagnosed with DLBCL at King Khalid University Hospital, Riyadh, from January 2011 to June 2019. Data were collected from computerized hospital information system and from the files of the patients. Variables studied included age at diagnosis, stage at diagnosis, international prognostic index (IPI) and CNS-IPI score, site(s) of extra-nodal involvement, type of chemotherapy received, CNS prophylaxis and CNS relapse. CNS prophylaxis was administered on the basis of presence of high-risk features like presence of ≥2 extranodal sites, involvement of bone marrow, bone, testes, nasopharynx and paranasal sinuses. Patients with presence of CNS involvement at diagnosis and primary CNS lymphoma were excluded.

A total of 101 patients were diagnosed with DLBCL during the study period. There were 58 males and 43 females with a median age of 56 (range: 16-87) years. Ann Arbor stage of I-IV was assigned in 9, 21, 17 and 50 patients, respectively. The lung was the most common extranodal site involved in 27 (26.7%) patients, and liver and bone marrow involved in 20 (19.8%) patients each. Gastrointestinal tract was involved in 9 (8.9%) patients, kidneys in 5 (4.95%), breast in 4 (4%), and testis and adrenal in 2 (2%) patients each. Twenty-five (24.75%) patients had high risk CNS-IPI score, 44 (43.5%) had intermediate risk score and 32 (31.7%) had low risk score. Ninety-four (93%) patients received R-CHOP chemotherapy while rest of the patients received other types of chemotherapy, mostly a milder regimen (R-CVP), because of comorbidities and poor performance status. Sixteen patients received CNS prophylaxis, which was IT methotrexate (MTX) ± cytarabine/hydrocortisone in all patients. Nine of 25 (36%) patients with high-risk CNS-IPI score did not receive CNS prophylaxis. After a median follow up of 36 months (range 4-114), 2 (2%) patients developed CNS relapse and died shortly after this diagnosis. Both the patients with CNS relapse had high risk CNS-IPI score and did not receive CNS prophylaxis.

CNS relapse of DLBCL was uncommon in this patient population despite limited use of IT CNS prophylaxis in high-risk patients. Low incidence of CNS relapse in many high-risk patients despite limited use of IT prophylaxis may be related to rituximab use and/or other factors. Our data indicate that IT CNS prophylaxis may be adequate for DLBCL patients at high risk of CNS relapse.