Serum interleukin-10 and clinical outcomes in primary gastrointestinal lymphoma: a retrospective cohort study

Primary gastrointestinal lymphoma (PGIL) comprises histologically diverse extranodal lymphomas with markedly different biological characteristics and clinical outcomes. This study evaluated the association between pretreatment serum cytokine concentrations and clinical outcomes in PGIL, with particular attention to interleukin-10 (IL-10). We retrospectively reviewed 165 adults with PGIL treated between 2012 and 2022. Pretreatment serum cytokine measurements were available for 85 patients, who comprised the cytokine cohort. Overall survival (OS) and progression-free survival (PFS) were estimated using the Kaplan–Meier method. Serum IL-10 was analyzed as a log-transformed continuous covariate in Cox proportional-hazards models. Conventional receiver operating characteristic (ROC) analysis of vital status was performed as an exploratory assessment of discrimination for overall mortality. Histology-stratified IL-10 distributions, a DLBCL-specific continuous Cox analysis, and a sensitivity analysis excluding MALT and follicular lymphoma were also performed. In the full cohort, the reverse Kaplan–Meier median follow-up was 33.4 months, and the 3-year OS rate was 85.1%. Within the cytokine cohort, baseline IL-10 and IL-6 concentrations were higher in deceased patients than in survivors. Log-transformed IL-10 was associated with OS in univariate Cox analysis (hazard ratio [HR] 1.66 per log-unit increase, 95% confidence interval [CI] 1.14–2.43; P = 0.009), but the association was attenuated after adjustment for Ann Arbor stage (HR 1.46, 95% CI 1.00–2.15; P = 0.053). In the DLBCL subgroup ( n = 34), the continuous estimate was highly sensitive to the patient with the highest IL-10 concentration. In the sensitivity cohort excluding MALT and follicular lymphoma ( n = 49), higher log-transformed IL-10 showed a nonsignificant trend toward inferior OS (HR 1.39, 95% CI 0.94–2.06; P = 0.099). Log-transformed IL-10 was not associated with PFS (HR 1.20, 95% CI 0.87–1.65; P = 0.273). The exploratory conventional ROC analysis yielded an area under the curve of 0.722 (95% CI 0.550–0.893), with a data-derived Youden cutoff of 5.31 pg/mL. Higher pretreatment serum IL-10 may represent a preliminary adverse prognostic signal in PGIL, particularly in aggressive histologic subtypes. However, an association independent of disease stage and histology was not established in this cohort. These findings require validation in larger, histology-specific cohorts with appropriate comparator groups.