Quantifying Population-level Mortality and Cure Rates in U.S. Non-Hodgkin Lymphoma Patients by Tumor Subtypes Open Access
Downloadable ContentDownload PDF
Background The non-Hodgkin lymphomas (NHL) are a heterogeneous group of hematologic malignancies arising from lymphoid tissue, with varied clinical and biological features. Many systems have been developed over time to classify NHL, reflecting increased understanding of features of different types of the disease. Some NHLs are potentially curable with chemotherapy and, less often, radiotherapy; however, treatment protocols vary by NHL subtype and continue to evolve. In addition, the clinical course and survival after NHL diagnosis vary among tumor subtypes, with the 5-year relative survival for the major NHL subtypes, ranging from 52% for the aggressive diffuse, large B-cell lymphoma (DLBCL) to approximately 75% for the more indolent follicular lymphoma (FL) and chronic lymphocytic leukemia (CLL)/small lymphocytic lymphoma (SLL). Better diagnostic and classification tools, as well as the epidemic of human immunodeficiency virus (HIV) infection can explain some increase in NHL incidence by tumor subtypes. In contrast, changes in population-level mortality patterns in NHL patients by tumor subtypes have not been systematically described. Evaluation of mortality changes due to NHL subtypes in the general population is important, since NHL encompasses a wide variety of disease subtypes for which incidence pattern varies, HIV epidemic increased incidence rates for some NHL subtypes, and treatment advances have improved survival for some, but not for all NHL subtypes. To date, little is known about how population-level mortality burden varies by tumor subtypes, the impact of HIV epidemic on the overall NHL mortality burden in the U.S. general population, or cure of disease in the modern treatment era for NHL patients. Objectives We conducted three studies to better quantify population-level mortality and cure rate in U.S. NHL patients by tumor subtypes. The study aims were to 1) to evaluate population-level mortality attributed to each of the 4 major NHL subtypes over time in the US general population, 2) to evaluate mortality burden in the U.S. adult population and its relationship with the HIV epidemic, and finally 3) to estimate proportion of DLBCL patients who may be considered cured in the modern treatment era. Results from the three studies will be a valuable resource for cancer control planners as well as treating clinicians. Methods We used incidence and survival data from population-based cancer registries from the National Cancer Institute's (NCI’s) Surveillance, Epidemiology, and End Results (SEER) program to realize these aims. Cause of death (COD) ascertained from death certificates was obtained from the National Center for Health Statistics (NCHS) to assess mortality burden from NHL cancer. We used incidence-based mortality (IBM) method to quantify mortality attributable to major NHL subtypes and to evaluate the impact of HIV on overall NHL mortality (aims 1 and 2). Joinpoint regression program was used to characterize changes in mortality rates by tumor subtypes over time. SEER HIV flag was used to identify HIV-infected individual at the time cancer diagnosis. Mixture cure model was used to quantify proportion of DLBCL patients who may be considered cured (known as ‘statistical’ cure) in the modern treatment era by important clinical and demographic characteristics (aim 3). We developed an epidemiologic risk profile model for DLBCL patients incorporating important clinical and demographic characteristics at the time of diagnosis. Finally as part of the 3rd aim, we evaluated risks of death from cancer and other causes compared with the general population by time since DLBCL diagnosis.Results Overall NHL mortality rates increased during 1975-1997, peaking at 10.9 per 100,000 person-years, then decreased subsequently in 1997-2011. Overall IBM rates mirror this trend during 1990-2011. For B-cell NHL subtypes, IBM rates decreased beginning in the mid-1990s, with yearly declines of -3.0% for diffuse large B-cell lymphoma (DLBCL), -2.7% for chronic lymphocytic leukemia/small lymphocytic lymphoma (CLL/SLL), and -5.3% for follicular lymphoma. Incidence rates for these subtypes did not decrease until after 2003. Corresponding five-year cancer-specific survival increased dramatically over time for DLBCL (from 40% to 67%), CLL/SLL (69%-80%), and follicular lymphoma (75%-86%). IBM for peripheral T-cell lymphoma was flat during 2006-2011, although incidence increased. Of 11,071 NHL deaths, 517 (4.6%) were in HIV infected persons during 2005-2012 covering about 8% of US general population. This proportion was higher in deaths mapped to DLBCL (7.3% with HIV), Burkitt lymphoma (33.3%), and CNS lymphoma (17.6%) cases, and among deaths from these subtypes, for people aged 20-49 years (46.6%), males (15.2%), and blacks (39.3%). IBM rates declined steeply during 2005-2012 for HIV-infected NHL cases (-7.6%/year, p-value=0.001). This trend reflects a steep decline in incident NHL among HIV-infected people after 1996, when highly active antiretroviral therapy was introduced. Five-year NHL-specific survival among HIV-infected cases increased dramatically from 9% to 54% during 1990-2008, whereas among HIV-uninfected cases survival improved but more modestly 62%-76%) during this period. Overall 8,274 deaths were recorded among 18,047 people with DLBCL; 76% of total deaths were attributed to DLBCL and 24% were attributed to non-cancer causes during 1992-2011 covering about 14% of US general population. Ten-year survival rates for low, medium, and high risk groups were, 80%, 60%, and 36%, respectively with predicted cure proportions of 63%, 44%, and 25%, respectively, although these cure estimates should be interpreted with caution as they rely on extrapolation beyond observed survival time. Mortality risks were elevated for a number of conditions including vascular disease (SMR 1.35); infectious disease (3.13); lung disease (SMR 1.29); gastrointestinal disease (2.50) and blood disease (4.64). Patients had a reduced risk of mortality due to neurological diseases (0.64). Conclusions Mortality due to three common B-cell NHL subtypes has fallen over time in the U.S. This decline reflects better survival after NHL diagnosis, likely from improved therapies, because the decline in NHL incidence occurred later. Furthermore, the HIV epidemic has strongly contributed to NHL deaths, especially for AIDS-defining NHL subtypes and groups with high HIV prevalence. Declining NHL mortality rates for HIV-infected cases reflect both declining incidence and improving survival. Finally, quantifying survival among DLBCL cases from the perspective of cure, we found that 63%, 44%, and 25% of DLBCL patients in low, medium, and high risk groups cured, although it was difficult to say with certainty the actual proportion of DLBCL patients who are cured. Additionally, we also noted newly diagnosed DLBCL patients treated in the immunochemotherapy era continue to experience excess risk of death from cancer and many other non-cancer causes. Clinicians should be aware of recently published strategies for the long-term medical care of DLBCL cancer survivors.