Background: We studied changes in blood concentrations of the soluble B-cell activation markers sCD27 and sCD30 in relation to future lymphoma risk, by investigating repeated pre-diagnostic plasma samples from the same individuals. High levels of sCD27 and sCD30 have been associated with lymphoma risk in previous prospective studies based on a single pre-diagnostic blood sample per participant. These studies, however do not inform about the dynamics of the marker-disease association on an individual level hampering its interpretation and potential use in lymphoma diagnostics and disease monitoring.Methods: In the Northern Sweden Health and Disease Study cohort we identified 170 individuals who had donated two pre-diagnostic blood samples and subsequently developed B-cell lymphoma. Blood samples were donated 13.2±4.4 and 5.5±4.0 years (mean±SD) prior to diagnosis, respectively. Cancer-free controls from the same cohort were individually matched to cases on a 1:1 ratio on sex, age and blood draw dates. We investigated associations between lymphoma risk by subtypes and marker concentrations; as measured separately at baseline, at time of the repeated sample, as well as the association with rate of change (slope) while adjusting for the baseline marker concentration.Findings: We observed strong associations between B-cell lymphoma risk and sCD27 and sCD30 concentrations, both at baseline and at time of the repeated blood sample. Associations were evident in samples collected up to 25 years before diagnosis. Blood concentrations of sCD27 and sCD30 increased significantly closer to diagnosis among future B-cell lymphoma cases, while they remained temporally stable among controls. Modeling measures of baseline and slope simultaneously with multivariable conditional logistic regression to estimate odds ratios (ORs), revealed significant associations between both slope and baseline measures and lymphoma risk; sCD27 (ORBaseline=7.2, Ptrend=2.8 x 10-4; ORSlope=2.8, Ptrend=2.2 x 10-4), and sCD30 (ORBaseline=2.9, Ptrend=3.7 x 10-4; ORSlope=2.9, Ptrend=0.001) (ORs 4th vs. 1st quartile; P for linear trend based on median values of analyte quartiles used as a continuous variable). Subtype specific analyses showed that the association between lymphoma risk and slope was restricted to indolent subtypes and mainly driven by chronic lymphocytic leukemia; sCD27 (ORSlope=6.7, Ptrend=1.3 x 10-5), and sCD30 (ORSlope=5.9, Ptrend=4.8 x 10-5), while associations with baseline marker concentration were evident among different subtypes, including diffuse large B-cell lymphoma.Interpretation: B-cell lymphoma risk seems to be influenced by increased sCD27 and sCD30 blood concentrations more than two decades before diagnosis. Increased concentrations of these B-cell activation markers among different future lymphoma subtypes early in life may reflect a constitutional predisposition, suggesting a role of B-cell activation in lymphoma development at early stages across subtypes. The observed increase of sCD27 and sCD30 concentrations closer to diagnosis however, seems to be subtype-specific and restricted to indolent lymphoma cases. Therefore, these markers may reflect early progression of undiagnosed disease and could potentially be utilized to improve lymphoma diagnostics and disease monitoring in indolent lymphoma. DisclosuresNo relevant conflicts of interest to declare.