Project Summary/Abstract
Non-Hodgkin lymphoma (NHL) and multiple myeloma (MM) are cancers arising from mature lymphocytes that
collectively pose a substantial public health, personal and financial burden. Present knowledge of their
etiology is inadequate to inform prevention strategies; data from prospective studies, and for specific histologic
subtypes, are particularly limited. Inflammation, heightened immune activation and growth factor dysregulation
contribute to their pathogenesis, as indicated, for example, by well-known associations of NHL risk with severe
immune compromise and with pre-diagnosis plasma immune marker profiles that indicate inflammation and B-
cell activation, as well as by consistent findings of an increased risk of MM and some NHL subtypes in obese
individuals and in persons with autoimmune disease. Relevant to the present application, evidence from both
model organism and human occupational studies support the hypothesis that environmental pollutants with
carcinogenic and/or immunotoxic effects—such as dioxins and hazardous air pollutants (HAPs)—may be risk
factors for MM and/or NHL. However, those studies have many limitations, and most do not assess general
population risk from low-dose, passive and possibly chronic exposure to those pollutants or to combinations of
multiple pollutants. Geographic information system (GIS)-based modeling of these exposures offers a
valid and well accepted tool for examining environmental risk factors for MM, NHL and common
histologic subtypes of NHL and could yield novel insights into their etiopathogenesis. The prospective
studies proposed in this application will leverage strong investigator expertise and an active collaboration and
will include eight large and diverse cohorts, including geocoded residential histories spanning many decades
and a projected sample size of ~5,386 incident MM and ~20,354 incident NHL cases. Cox proportional
hazard models will be used to examine GIS-modeled passive exposure to dioxins and HAPs (individual and
mixtures of pollutants, the latter using quantile G-computation and Bayesian approaches) with known or
plausible carcinogenic or immunotoxic properties in relation to risk of MM, NHL, major histologic subtypes and
clinically relevant molecular subtypes of diffuse large B-cell lymphoma (DLBCL), an aggressive histologic type
of NHL with unknown etiologies. The analyses will incorporate already-harmonized individual-level time-varying
risk factor data to assess and control confounding by lifestyle, occupation, demographic and medical history-
related risk factors. Additional analyses will assess timing of susceptibility and heterogeneity of observed
associations by NHL histologic subtype, calendar period, sex, age, race/ethnicity, body mass index, smoking
and socio-demographic and neighborhood contextual factors. The proposed studies have strong potential to
yield novel insights on the contribution of passive, low-level exposure to environmental factors to the etiology
of MM and NHL, including on risk to individual histologic subtypes of NHL and molecular subtypes of DLBCL,
and thus to provide urgently needed identification of novel risk factors for these understudied cancers.