The Use of Biomarker Ferritin and Albumin As Well As HCT-CI, but Not PAM, Was Predictive of Suvival, Gvhd and, Relapse in Myeloid Malignancies

Abstract

Developing the optimal models for pretransplant assessment as more reduced intensity conditioning regimens are implemented and the higher median age of transplant. The pre-transplant assessment of mortality (PAM) assesses all cause mortality at 2 years and the Hematopoietic Stem cell Transplantation specific comorbidity index (HCT-CI) which previously used Charlson comorbidity index (CCI) to assessed NRM based on history and organ function. Previous studies have shown that certain biomarkers such as ferritin and albumin may help optimize risk assessment before transplant. We tested the validity of PAM, HCT-CI, serum ferritin and albumin to predict GVHD, relapse and overall survival (OS) in our older patient population. IRB approval was obtained to retrospectively review clinical outcomes in patients undergoing allogeneic stem cell transplantation at our institution between, 2010 - 2013. We examined a total of 94 patients with myeloid malignancies (see Table 1 for patient characteristics). Risk variables assessed included calculated PAM, HCT-CI and KPS scores and pre HCT serum ferritin and albumin levels. We examined the predictive value of these biomarkers with respect to GVHD, relapse, and overall survival. Of the 94 patients 46% are alive and in remission, 20% have relapse and GVHD developed in 56% of patients. To determine the optimal cutoff for classifying into high and low groups for each risk measure, first the association between the measure and clinical outcomes was determined to obtain the logistic regression. Then area under the curve (AUC) of the receiver operating characteristic (ROC) curve for each clinical outcome against the risk measure was calculated. The cutoff that maximized the sum of the ROC-AUC across all measured was selected. Patients with HCT-CI > 4, (cut off 4), had a less likelihood of survival than those with HCT < 4 OR 4.3 (95% CI: 1.6-11.2) (p=.0033). Similarly, patients with low albumin, (cutoff 4.1), were less likely to than those with high albumin survive OR 0.4 (95% CI: 0.2-0.9) (p=0.02). Patients in the high ferritin group, (cut off 2425) were less likely to survive than those with low ferritin OR 4.0 (95% CI:1.3-2.3) (p=0.016). The high HCT-CI group, (cut off 4) was more likely to relapse than lower HCT-CI OR 2.9, (95% CI: 1.1-8.1) (p=0.04) and patients with a high KPS, (cutoff 81), were less likely to relapse than lower KPS OR 0.3 (95% CI: 0.1-0.8) (p=0.22). In the high ferritin group, (cutoff 2020) patients were more likely to have GVHD than low ferritin group OR 0.2 (95% CI: 0.1-0.7) (p=0.005). (Table 2) We conclude that HCT-CI was more predictive of mortality than the PAM score in our cohort or patients with median age 65 years old. Our results are concordant with previously reported data showing serum ferritin and albumin is useful biomarkers to predict GVHD, relapse and survival. We are now incorporating objective markers such as ferritin, albumin, and other functional assessments to objectively quantify risk in our current patients presenting for allogeneic HSCT.Table 1Patient CharacteristicsAll Patients (N = 94)N (%)Patient Age< 5033 (35)50-5935 (37)60-7026 (28)RaceCaucasian76 (81)African American14 (15)Other4 (4)DonorRelated40 (43)Unrelated54 (57)HLA MismatchNo71 (76)Yes10 (11)Missing13Stem Cell SourceBM5 (5)PBSC89 (95)ConditioningMyeloablative47 (50)Reduced Intensity47 (50)ATGNo13 (14)Yes81 (86)Patient CMV Sero statusNegative41 (44)Positive48 (51)Missing5KPS< 9017 (18)90-10077 (82)BM indicates bone marrow; PBSC, peripheral blood stem cells; CMV, cytomegalovirus.Table 2Biomarkers and Clinical OutcomesBiomarkersGVHDRelapseSurvivalCut-offp-valueCut-offp-valueCut-offp-valuePAM240.1222240.4911210.2077HCT-CCI20330840.049540.0033Ferritin20200.005410700.381824250.0164KPS810.1679810.0226910.1289Albumin4.20.11034.20.27504.10.0220 DisclosuresNo relevant conflicts of interest to declare.