CNS Relapse in Diffuse Large B-Cell Lymphoma Patients with Oral Cavity and Bone Marrow Involvement-Experience from an Inner City Hospital

Introduction:

Central nervous system (CNS) relapse in patients with Diffuse Large B-cell Lymphoma (DLBCL) is an uncommon yet serious complication with incidence reported between 2-10% depending on patient characteristics and risk factors. Recently N.Schmitz et al.(doi:10.1200/JCO.2015.65.6520) proposed and validated a robust CNS-IPI risk model that includes individual International Prognostic Index (IPI) risk factors and kidney/adrenal extra-nodal site involvement. Head & neck and bone marrow involvement were not associated with increased CNS relapse risk in the model. These two variables have long been associated with increased risk of CNS relapse. Aim of the present study was to evaluate CNS relapse in patients with DLBCL who had bone marrow and head & neck involvement at diagnosis.

Methods:

Retrospective chart review was carried out on patients treated for DLBCL at John H Stroger Jr.Hospital of Cook County, an inner city urban hospital in Chicago, between 2007- 2011. All patients had histologically confirmed diagnosis of DLBCL. Patients with cutaneous, CNS involvement at diagnosis and those with incomplete charts were excluded. Variables studied included age at diagnosis, stage at diagnosis, bone marrow infiltration, IPI-score at diagnosis, type of extra-nodal involvement, type of chemotherapy, initial systemic response, CNS relapse and timing of relapse. Decision regarding CNS prophylaxis was at the discretion of treating physician. CNS-IPI was calculated for each chart. Data was analyzed using descriptive statistics (frequency, mean, median), non-parametric Fischer's exact test and logistic regression analysis.

Results:

120 charts met the inclusion criteria. Median follow up duration was 56 months (IQR: 24 to 72 months). Average age at diagnosis was 51yrs. 61 (52.3%) patients were stage III/IV at diagnosis. 68 (57%) had IPI score >/= 2 at diagnosis. Extra nodal involvement was seen in 52(43%) patients. Extra nodal sites involved were head & neck 22(18%), followed by stomach 10(8%), visceral involvement 10(8%), visceral lining 6(5%) and bowel involvement 6(5%). Renal involvement was seen in 2 patients. Of the 22 patients with head and neck involvement, 18(82%) had oral cavity (tonsillar, pharyngeal) and 4(18%) had orbital involvement. 12(10%) patients had bone marrow involvement. 111 (93%) patients were treated with RCHOP and 9(7%) with DA-EPOCH R. Seven(5.8%) patients received CNS prophylaxis with intrathecal methotrexate in 3-4 doses, with oral cavity involvement being the indication in 5/7 (72%) patients. 87(73%) had complete treatment response to systemic chemotherapy, rest had progressive disease. 24(27%) of the 87 patients relapsed, median time to relapse was 18months (IQR 11-24months). Three (2.5%) patients had CNS relapse at a median of 38 months(IQR 6-48months). On logistic regression analysis, increasing CNS IPI-risk score was predictive of CNS relapse (p= 0.05; OR 2.9, CI: 1.001 to 8.558).Bone marrow involvement was not significantly associated with risk of CNS relapse (p=0.273). No statistically significant difference in CNS relapse was noted between patients with oral cavity involvement that did and did not receive CNS prophylactic treatment (p=1.00).

Conclusion:

Bone marrow and oral cavity (tonsillar and pharyngeal) involvement by DLBCL were not associated with increased risk of CNS relapse in the studied population, hence CNS directed prophylactic treatments in these group of patients may be omitted. Data from our cohort supports the use of CNS-IPI risk model to predict risk of CNS relapse.