Abstract Introduction Angioimmunoblastic T-cell lymphoma (AITL) is an aggressive subtype of peripheral T-cell lymphoma (PTCL) with poor survival outcomes. Characterized by a unique inflammatory and immune response, its aggressiveness is highly dependent on the tumor microenvironment (TME). Although cytokines are critical messengers within the TME, their prognostic role in AITL is poorly understood. This study aimed to analyze the relationship between serum cytokine profiles and clinical outcomes in patients with AITL. Methods This multicenter, retrospective study enrolled newly diagnosed adult AITL patients between August 2015 and December 2023. Serum levels of seven cytokines were serially analyzed at diagnosis and subsequent timepoints. Optimal cutoff values for overall survival (OS) were determined using maximally selected rank statistics. Results The cohort of 267 patients presented with high-risk features, including a median age of 67, a male predominance (57.7%), advanced-stage disease (97.4%), and a majority having poor-risk prognostic scores (IPI ≥ 3: 65.9%; PIT ≥ 2: 73.0%; PIAI ≥ 2: 66.3%). At diagnosis, the median serum cytokine levels were 0.9 pg/mL for interleukin (IL)-2, 0.9 pg/mL for IL-4, 17.5 pg/mL for IL-6, 9.1 pg/mL for IL-10, 1.9 pg/mL for tumor necrosis factor (TNF)-α, 2.4 pg/mL for interferon (IFN)-γ, and 0.1 pg/mL for IL-17A. Elevated levels of IL-6 and IL-10 were both associated with poorer ECOG PS, higher LDH and β2-MG levels, lower albumin, thrombocytopenia, B symptoms, bone marrow involvement, and higher-risk prognostic scores. High IL-10 was also associated with a more advanced disease stage and serous effusion, whereas other cytokines showed limited correlations. Of the seven cytokines analyzed, only IL-6 and IL-10 demonstrated prognostic significance for OS, with optimal cutoffs of 14.6 pg/mL and 16.4 pg/mL, respectively (both p .001). High baseline IL-6 (n=147) and IL-10 (n=91) were associated with significantly shorter median OS (16.8 vs. 65.1 months and 9.4 vs. 65.1 months, respectively; both p .001). Concurrent elevation of both cytokines conferred the worst OS (p .001). Furthermore, IL-6 and IL-10 levels correlated with the response to first-line chemotherapy. Among 241 evaluable patients, low IL-6 levels were associated with significantly higher complete response (CR) and overall response rates (ORR) (CR: 60.4% vs. 33.8%, p .001; ORR: 80.2% vs. 66.2%, p = .015). This therapeutic advantage was even more pronounced for patients with low IL-10 levels (CR: 57.2% vs. 21.3%; ORR: 81.3% vs. 53.3%; both p .001). In univariate analysis, both elevated IL-6 and IL-10 levels significantly predicted worse OS (both p .001). However, in multivariate analysis adjusted for PIAI score, only high IL-10 (HR, 1.966; p .001) and low albumin (HR, 1.595; p 0.05) remained independent prognostic factors for poor OS. We therefore compared the prognostic ability of various dynamic IL-10 metrics (absolute level, absolute change, and relative change). The results showed that only the absolute IL-10 levels after course 1 (cutoff, 23.8 pg/mL) and at the end of treatment (cutoff, 5.1 pg/mL) were independent prognostic indicators. While the end-of-treatment IL-10 model demonstrated a superior statistical fit (AIC: 543.1 vs. 598.9; Harrell's C-statistic: 0.768 vs. 0.761), the absolute IL-10 level after course 1 offered greater clinical utility for early prognostication. In subgroup analyses, this early marker was prognostic regardless of baseline IL-10 status; however, it did not predict early recurrence (≤ 12 months). Additionally, IL-10 levels were significantly higher at relapse compared to post-treatment levels (median: 11.5 vs. 3.4 pg/mL, p .001), an increase observed in 64.3% (32/39) of patients who had initially responded to therapy (CR/PR). Conclusions Serum levels of IL-6 and IL-10 correlate with adverse clinical features, chemoresistance, and poor survival in AITL. An elevated IL-10 level at diagnosis serves as an independent adverse prognostic factor. Dynamic monitoring of IL-10, particularly after the first course of therapy, provides a powerful and clinically valuable tool for early prognostication.
Zhao et al. (Mon,) studied this question.