The Endothelial Activation and Stress Index (EASIX) Score Improves Risk Stratification for Double Expressor Diffuse Large B-Cell Lymphoma (DEL)

Introduction

Double-expressor lymphoma (DEL), defined as Diffuse Large B-cell Lymphoma (DLBCL) with expression of MYC ≥40% and BCL2 ≥50% by immunohistochemistry (IHC) staining has poorer outcomes then non-expressor DLBCL. The R-IPI (revised international prognostic index) was developed to risk-stratify DLBCL patients treated with R-CHOP. DLBCL however is a heterogenous disease and the R-IPI score may not accurately prognosticate survival outcomes in patients with poorer risk DLBCL such as DEL. The EASIX score, a surrogate for endothelial dysfunction, was originally designed to predict an individual's risk of mortality after graft-versus-host disease (GvHD) onset (Luft et al. 2017). Its components of platelet count, serum creatinine, and lactate dehydrogenase (LDH) have been reported prognostic factors for lymphoma. Therefore, we planned this study to determine whether EASIX could improve the prognostication for patients with DEL.

Methods

The records of 251 DEL between January 2013 to December 2023 were reviewed. Only patients who received R-CHOP and had a minimum of 1-year follow-up were included. Patients with double hit, i.e MYC and BCL2 translocation, were excluded. EASIX scores were calculated using serum lactate dehydrogenase (LDH), creatinine levels and the platelet count measured at time of treatment initiation as follows: LDH (U/L) × creatinine (mg/dL)/platelet count (10⁹/L). Optimal cutoff for EASIX score for both progression free survival (PFS) and overall survival (OS) was determined using receiver operating characteristic curve (ROC) analysis. Clinically however, it would be less convenient to use 2 different EASIX score cut-off for risk stratification. Thus, we combined the Youden index for both ROCs for PFS and OS. The EASIX score with the highest Youden index was used to dichotomise the study population into high and low EASIX scores. A score of 1 was assigned to a high EASIX score. Revised International Prognostic index (R-IPI), high/low EASIX score, and a combined R-IPI/EASIX score were then evaluated as prognostic factors for PFS and OS.

Results

A total of 191 patients who met the inclusion criteria were included. Median age was 65 years old. Median follow-up was 4.9 years with a maximum follow up of 10.6 years. Patients with R-IPI 0 (n= 8) had a 5-year PFS of 85.7% and 5-year OS of 100%; patients with R-IPI 1-2 (n=80) had 5-year PFS of 68.7% and 5-year OS of 74.1%; and patients with R-IPI 3-5 (n=101) had 5-year PFS of 39.6% and 5-year OS of 52.0%. Both 5-year PFS and OS were inferior to the published 5-year survival approximates ~ 90% (R-IPI 0/very good), 80% (R-IPI 1-2/good), and 60% (R-IPI 3-5/poor)(Sehn et al. 2007).

Median EASIX score was 1.98 (IQR 1.40-4.77). According to ROC analysis, the area under the curve (AUC) for PFS was 0.642 (p <0.001) and for OS was 0.66 (p<0.001). An optimal EASIX score cut-off of 1.87 gave the highest combined Youden index. High EASIX score (n=102) was an independent prognostic factor for both PFS and OS, with 5-year PFS of 43.2% vs 63.6% and 5-year OS of 54.2% vs 71.9% (p<0.001). We subsequently combined the scores of R-IPI and EASIX resulting in a total score of 6. We categorised the scores into 3 groups: 0-1, 2-4 and 5-6. Patients with a combined R-IPI/EASIX score of 0-1 (n= 38) had a 5-year PFS of 90.1% and 5-year OS of 100%; patients with score of 2-4 (n=110) had 5-year PFS of 50.3% and 5-year OS of 63.3%; and patients with a score of 5-6 (n=43) had 5-year PFS of 31.7% and 5-year OS of 36.8%. The combined R-IPI/EASIX score was better than the R-IPI score in terms of model discrimination for both PFS (C-index: R-IPI/EASIX 0.622 vs R-IPI 0.593) and OS (C-index: R-IPI/EASIX 0.662 vs R-IPI 0.621). Drawing a direct comparison between the 5-years PFS and OS of the 3 categories of the combined R-IPI/EASIX score and that of the R-IPI score, the R-IPI/EASIX score was better able to identify DEL patients who had a very good risk, i.e 5-year PFS and OS of ³90%. It also identified a very poor risk group that had 5-year PFS and OS of £40%. Thus, we propose the following risk stratification with our combined R-IPI/EASIX score - Good risk (score 0-1), Poor risk (score 2-4), Very Poor risk (score 5-6).

Conclusion

We demonstrated that the EASIX score is an independent prognostic factor for patients with DEL. In combination with the R-IPI score it can potentially better risk stratify DEL patients and identify poor-risk patients that will benefit from novel therapies. Further validation studies would be needed.

Chan:SymBio Pharmaceuticals: Research Funding. Goh:Novartis: Consultancy; AstraZeneca: Consultancy; Johnson & Johnson: Consultancy, Honoraria; Antengene: Consultancy; Amgen: Consultancy; EUSA Pharma: Consultancy; Pfizer: Consultancy; MSD: Honoraria; Roche: Honoraria; Astellas: Honoraria; AbbVie: Honoraria; NS Pharma: Consultancy.