Session: 622. Lymphomas: Translational – Non-Genetic: Poster III
Hematology Disease Topics & Pathways:
Clinical Practice (Health Services and Quality), Lymphomas, Diseases, Aggressive lymphoma, Lymphoid Malignancies
We retrospectively collected 295 secondary HLH patients diagnosed in our institute from Jan. 2019 to Dec. 2023, all patients had data of cytokine expression at the time of diagnosis. The diagnosis of HLH was made according to the HLH-2004 protocol. All statistical analyses were performed using SPSS version 25.0 and GraphPad Prism version 8.0. P< 0.05 was statistically significant.
Of the 295 HLH patients detected cytokine expression at baseline, the median age was 37 years old, 46 patients (15.6%) were older than 60. 116 patients (39.3%) with HLH were caused by malignancies, mostly lymphoma. Among these, natural killer/T cell lymphoma (32/116, 27.6%) was the first-ranked entities that contribute to secondary HLH, followed by diffuse large B cell lymphoma (23/116, 19.8%), acute natural killer cell leukemia (13/116, 11.2%) and peripheral T cell lymphoma, non-otherwise specific (11/116, 9.5%).
At a median follow-up of 673 days, 116 (39.3%) patients died. The median OS of the whole cohort was not reached (range: 1-1563 days), the 60-day OS rate was 70.5%. 1 year survival of the cohort was 61.5%, 3 years survival of the cohort was 57.3%. For different underlying diseases, 1 year OS was 87.1%, 76.7%, 61.0%, 33.1% for rheumatic diseases, infection, others and malignancies-related HLH, respectively (P=0.000).
In term of clinical characteristics, we found that IL-1β was more likely to be highly expressed in malignancies (M-HLH) compared with rheumatic-related HLH (R-HLH). Similarly, IL6 was more frequently upregulated in M-HLH compared with either R-HLH or infection-related HLH (I-HLH). In line with previous reports, IL10 was much higher in M-HLH than all other diseases caused HLH. In addition, we analyzed cytokine profile between lymphoma and non-lymphoma related HLH, and the result showed that IL6, IL10 and IFN-γ were all significantly higher in patients with HLH caused by lymphoma. Both IL10 and IFN-γ were remarkably high-expressed in HLH patients with positive EBV-DNA compared with those negative.
Among these cytokines, we found that high IL6, IL8 and IL10 significantly associated with poor survival, therefore we focused on these three cytokines. It revealed that patients with adult, age>60y, decreased Hb, decreased PLT, ANC<1.0*109/L, high TIBL, high DIBL, decreased ALB and high creatinine have much higher frequency of high IL6. Similarly, patients with high IL8 are tended to have decreased WBC, decreased PLT, decreased AMC, high AST, high TIBL and high DIBL. High IL10 were associated with adult, decreased Hb, decreased PLT, decreased ALC, high TIBL, high DIBL, decreased ALB, high creatinine, high β2MG, triglyceride>3mmol/L, high sCD25, FIB<1.5g/L, EBV positive and BM infiltration.
Univariate survival analysis was performed, it showed that there are many factors predict inferior OS, including age>60y, Hb<90g/L, high TIBL, high creatinine, high β2MG, BM infiltration, treatment initiated within 2 weeks after diagnosis etc. Multivariate analysis demonstrated Hb<90g/L, high AST, BM infiltration and high IL10 are independent adverse predictors.
For these four independent risk factors, each one was scored 1, and the cohort was grouped into 4 subgroups (score 0-1: low risk, score 2: low-mediate risk, score 3: high-mediate risk, score 4: high risk). Totally, there were 78 patients in low-risk group, 105 patients in low-mediate risk group, 81 patients in high-mediate risk group, 31 patients were grouped into high risk. The 1-year OS for these four groups were 81.1%, 65.2%, 48.6% and 19.4%, respectively. This risk model, we named it as HLH score, would well distinguish the four risk subgroups.
Cytokines of secondary HLH displayed different patterns in terms of underlying diseases, clinical characteristics and responses. High IL6, IL8 and IL10 significantly associated with inferior survival, especially IL10, which is an independent risk factor of reduced OS. Moreover, we promoted a risk mode, which would well predict OS in secondary HLH.
Disclosures: No relevant conflicts of interest to declare.
See more of: Oral and Poster Abstracts