Diffuse Large B-cell Lymphoma of the Uterus: A Propensity Score Matched Analysis of the Texas Cancer Registry

Abstract

Introduction: Primary malignant lymphoma of the uterus is rare. Among such cases, extra-nodal diffuse large B-cell lymphoma (DLBCL) is the most common. Most of the epidemiology and prognosis for uterine DLBCL is derived from case reports and series. The main objective of this research was to compare survival between females with primary uterine DLBCL to females with primary malignant DLBCL of other sites. Methods: We retrospectively queried incidence files of the Texas Cancer Registry from 1997-2017 for women with DLBCL. To address factors that can confound survival analysis, propensity score matching (PSM) was used to balance the two cohorts of interest with respect to stage, treatment, race and age at diagnosis. The survival difference between the two cohorts was investigated using a Cox proportional hazards model based on the robust sandwich estimate for the covariance matrix to account for the dependent nature of the matching process. Results: We identified 12,444 female primary DLBCL cases of which 32 were of the uterus. Post 40:1 PSM, we compared 26 uterine to 1,040 other cases. The distributions of stage (p=0.0008), race (p=0.03), and age (p<0.0001) were significantly different prior to matching, but more similar post matching with stage (p=0.97), race (p=0.75), and age (p=0.27). Uterine cases exhibit a survival benefit prior to matching (p=0.001) which could be attributed to the aforementioned younger age and early stage at diagnosis as well as racial differences. However, the uterine cases observed significant survival benefit persisted post matching (p=0.045). Conclusion: Uterine DLBCL is a rare presentation of primary malignant lymphoma. Most women affected by DLBCL of the uterus are aged 64 years or younger (78.1%) and the disease tends to be localized at presentation (63.3%) facts that are not true for female DLBCL in general with localization (31.5%) and age (0-64: 40.1%). Survival in patients with uterine DLBCL is superior in comparison to patients with non-uterine DLBCL and does not appear to be an artifact of possible confounders.