The Easix Score Predicts Overall Survival and Late Relapse in Patients Undergoing Reduced Intensity Conditioned Allogeneic Stem Cell Transplantation for Myeloid Disease

Abstract

Background Allogeneic hematopoietic stem cell transplantation (alloSCT) is an effective treatment for myeloid malignancies, however outcomes post-transplant are variable and predictive scores limited. The Endothelial Activation and Stress Index (EASIX) measured at key timepoints pre- and post-transplant has been shown to predict outcomes in alloSCT, including overall survival, non-relapse mortality and incidence of sinusoidal obstructive syndrome (SOS). It can be readily calculated using standard biomarkers (LDH (U/L) x creatinine (mg/dL)/platelets (x109/L). We sought to examine its predictive role in a homogenous population of patients. Methods We conducted a retrospective national review of 117 patients undergoing reduced-intensity alloSCT for myeloid disease including AML (n=56), MDS (n=35) and MPN (n=26). All included received fludarabine, busulfan and rabbit anti-thymocyte globulin (Flu/Bu/ATG) conditioning from 2015-2020. We calculated the EASIX score on the day prior to conditioning, on day 0, at onset of acute graft versus host disease and 1 year post transplant. Univariate analysis was conducted in the STATA statistical package. Results At the pre-conditioning time-point the mean EASIX was 3.11 (0.3-17.54). Increased scores correlated with and were reliably predictive of reduced overall survival (p=0.034) (fig. 1). Higher scores at this time-point did not predict chronic GVHD or relapse, however an increased rate of acute GVHD was seen in those with lower EASIX (p=0.012). At day 0 the mean EASIX was 15.38 (1.07-179.02), however we could not demonstrate a predictive effect on SOS as no events occurred in this cohort. Nor did we identify a predictive value at this point for GVHD risk or relapse. At onset of acute GVHD (n=67, mean 3.52, 0.44-51.43), the EASIX score was predictive of overall survival (p=0.027) but did not predict risk of chronic GVHD or relapse. At 1 year post-transplant (n=88, mean 2.09, 0.48-13.59) a higher score showed a trend towards decreased overall survival (p=0.1) but was associated with an increased risk of subsequent relapse (p=0.01)(fig. 2). Conclusions The EASIX score is an easily applicable clinical tool using real world biomarkers readily available at serial time-points in transplant recipients. When applied to our homogenously treated cohort it retains its value in predicting overall mortality at the key time points of pre-conditioning and at onset of GVHD. Adoption of this simple score may be of benefit in risk stratifying the transplant process for reduced intensity conditioned allograft recipients, particularly when combined with established prediction tools for non-relapse mortality such as the HCT-CI. In addition to this, we have shown that the EASIX score may be a useful predictive tool for late relapse in those patients who are alive at 1 year post transplant. We suspect that this effect is primarily reflective of those patients with persistent thrombocytopenia and raised LDH, both markers of potential disease activity. Importantly, the use of this score in the reduced intensity population may be particularly relevant due to the increasing age of transplant eligibility and its use as a surrogate of endothelial dysregulation and potential for end organ complications through the transplant process. To assess the underlying T cell mediated and cytokine secretome effects on endothelial dysfunction, an observational study is ongoing at our institution, examining the interplay of immunomodulatory therapy, graft T cell content, early T and NK cell recovery and the cytokine secretion patterns at various time points before, during and after transplant in reduced intensity conditioned patients with myeloid disease. We plan to correlate these data with serial EASIX assessments to identify the determinants of the underlying biological process and identify potential time-points and biological targets for future study.