AUTHORS
Adi Zoref-Lorenz, MD, PhD, James R. Cerhan, MD, PhD2, Elliot Cahn3, Eric Mou, MD4, George A Cholack, MD, MSc5, Nicholas J Boddicker, PhD6, Yucai Wang, MD, PhD7, Andrew L. Feldman, MD8, Bri Negaard9, Raphael Mwangi, MS10, Joseph Peter Novak, MS11, Janek S. Walker, PhD12, Noa Rabinowicz, PhD13, Hadasa Malka, MsC14, Martin Ellis, MD15, Uri Abadi, MD16, Ronit Gurion, MD17, Pia Raanani, MD18, Thomas M. Habermann, MD19, Brian K Link, MD20, Stephen Maxted Ansell, MD, PhD21, Gaurav Goyal, MD22, Anne J. Novak, PhD23, Gilad Itchaki, MD24, Matthew E. Maurer, ScD25, Thomas Witzig, MD26, Michael Jordan, MD, PhD27
AFFILIATIONS
1 Cincinnati Children's Hospital Medical Center, Cincinnati, United States, Gray Faculty of Medical & Health Sciences, Tel Aviv-Yafo, Israel, Meir Medical Center, Kfar Saba, Israel, Meir Medical Center, Kfar Saba, Israel
2 Mayo Clinic, Rochester, Minnesota, United States
3 Mayo Clinic, Rochester, Minnesota, United States
4 University of Iowa Hospitals and Clinics, Iowa City, Iowa, United States
5 Mayo Clinic, Rochester, Minnesota, United States
6 Mayo Clinic, Rochester, Minnesota, United States
7 Mayo Clinic, Rochester, Minnesota, United States
8 Mayo Clinic, Rochester, Minnesota, United States
9 Mayo Clinic, Rochester, Minnesota, United States
10 Mayo Clinic, Rochester, Minnesota, United States
11 Mayo Clinic, Rochester, Minnesota, United States
12 Mayo Clinic, Rochester, Minnesota, United States
13 Gray Faculty of Medical & Health Sciences, Tel Aviv-Yafo, Israel, Meir Medical Center, Kfar Saba, Israel
14 Felsenstein Medical Research Center, Rabin Medical Center, Petach Tikva, Israel, Gray Faculty of Medical & Health Sciences, Tel Aviv-Yafo, Israel, Meir Medical Center, Kfar Saba, Israel
15 Gray Faculty of Medical & Health Sciences, Tel Aviv-Yafo, Israel, Meir Medical Center, Kfar Saba, Israel, Meir Medical Center, Kfar Saba, Israel
16 Gray Faculty of Medical & Health Sciences, Tel Aviv-Yafo, Israel, Meir Medical Center, Kfar Saba, Israel
17 Gray Faculty of Medical & Health Sciences, Tel Aviv-Yafo, Israel, Rabin Medical Center, Petach Tikva, Israel
18 Felsenstein Medical Research Center, Rabin Medical Center, Petach Tikva, Israel, Gray Faculty of Medical & Health Sciences, Tel Aviv-Yafo, Israel, Rabin Medical Center, Petach Tikva, Israel
19 Mayo Clinic, Rochester, Minnesota, United States
20 University of Iowa Hospitals and Clinics, Iowa City, Iowa, United States
21 Mayo Clinic, Rochester, Minnesota, United States
22 University of Alabama, Birmingham, Alabama, United States
23 Mayo Clinic, Rochester, Minnesota, United States
24 Gray Faculty of Medical & Health Sciences, Tel Aviv-Yafo, Israel, Meir Medical Center, Kfar Saba, Israel
25 Mayo Clinic, Rochester, Minnesota, United States
26 Mayo Clinic, Rochester, Minnesota, United States
27 Cincinnati Children's Hospital Medical Center, Cincinnati, United States, Cincinnati Children's Hospital Medical Center, Cincinnati, United States
ABSTRACT
Background: Prognostic models in large B-cell lymphoma (LBCL) primarily capture tumor burden and tumor-intrinsic biology, but not systemic immune dysregulation. We previously developed the blood-based Optimized HLH Inflammation Index (OHI), a pretreatment index that combines soluble CD25 (sCD25) and ferritin, to identify patients with hematological malignancies and clinically meaningful hyperinflammation and associated early mortality in the haemophagocytic lymphohistiocytosis (HLH) spectrum (Zoref-Lorenz et al., Blood, 2022). Although overt HLH is uncommon in LBCL, we hypothesized that a broader, clinically inapparent hyperinflammatory state contributes to treatment failure and early death. We further refined and validated the OHI for prognosis in the setting of LBCL. Here we present the LBCL Hyperinflammation Index (LHI), based on the OHI framework.
Methods: Adults with newly diagnosed LBCL treated with frontline immunochemotherapy were studied in 3 prospective cohorts: a discovery set (n=331) and an independent internal validation set (n=335), both from the Mayo Clinic/University of Iowa Molecular Epidemiology Resource (MER), and an external validation cohort from Israel (n=42). Optimal thresholds were derived in the discovery set by receiver operating characteristic (ROC) analysis for 500-day mortality and the Youden index. LHIhigh, LHIint, and LHIlow indicated elevation of both, one, or neither marker, respectively. Associations with 500-day mortality, event-free survival (EFS), overall survival (OS), and lymphoma-specific survival (LSS) were evaluated using logistic and Cox regression with adjustment for established clinical and molecular prognostic factors. Protein profiles were determined using SomaLogic proteomics in 9 external-cohort patients and independent immunoassay validation in 80 MER patients. LBCL molecular classification was performed in 210 patients and germline sequencing of 46 HLH-associated genes in 457 patients. Fulfillment of the HLH-2004 diagnostic criteria was assessed.
Results: The optimized thresholds derived by ROC in the discovery cohort were sCD25 >1,600 U/mL and ferritin >250 ng/mL. LHIhigh comprised 23.9%, 20.9%, and 33.3% of the discovery, internal-validation, and external-validation cohorts, respectively. Compared with LHIlow, LHIhigh was associated with markedly higher 500-day mortality in the discovery cohort (32.1% vs 5.0%; OR=8.96; 95% CI 3.84-23.6), internal validation cohort (21.4% vs 6.4%; OR=3.97; 95% CI 1.67-9.97), and external validation cohort (28.6% vs 0%; P=.006). LHIhigh was also associated with inferior EFS, OS, and LSS in both MER cohorts. In the pooled MER cohorts, LHIhigh remained associated with 500-day mortality (OR=3.71; 95% CI 1.88-7.33) and EFS (HR=1.85; 95% CI 1.38-2.49) after adjustment for age and International Prognostic Index. At 2 years, cumulative incidences of lymphoma-related death (23.7% vs 6.1%) and therapy-related death (5.4% vs 1.1%) were higher for LHIhigh than LHIlow.
Proteomic analysis comparing LHIhigh vs. LHIlow demonstrated coordinated activation of cytotoxic T-cell (soluble CD8 alpha and soluble CD137), interferon-gamma (CXCL9), tumor necrosis factor (soluble tumor necrosis factor receptor II), macrophage (MIF and soluble CD14), and complement (iC3b) pathways in LHIhigh patients. All pathways were independently validated in the 80-patient MER cohort (all P<0.001). The strongest associations were observed for CXCL9 and sCD8 (CXCL9: median 8299 vs. 989 pg/mL; sCD8: median 1520 vs. 477 pg/mL; both p<0.0001). LHI status in the combined cohort was not associated with major LBCL molecular classifications, MYC double-hit status, tumor Epstein- Barr virus status, or pathogenic germline variants associated with HLH. Although 149 of 666 MER patients (22.4%) were LHIhigh, only 1-2% met at least five of the eight HLH-2004 diagnostic criteria, demonstrating that LHI captures a substantially broader inflammatory phenotype than overt HLH.
Conclusions: LHI identifies a biologically distinct process in ~25% of LBCL patients, associated with early mortality and treatment failure, independent of conventional clinical risk factors. The coordinated activation of interferon-gamma, TNF, cytotoxic T-cell, macrophage, and complement pathways provides a biologic rationale for testing inflammation-directed therapy alongside lymphoma treatment in LHIhigh patients. LHI offers a simple strategy for biologically grounded risk stratification and prospective trial planning.
DISCLOSURES
Adi Zoref-Lorenz, MD, PhD: Electra - Research Funding; Sobi - Research Funding; Electra - Consultancy (Includes expert testimony); Sobi - Consultancy (Includes expert testimony); Sobi - Membership on a Board/Advisory Committee