C041 | A COMPREHENSIVE GERIATRIC ASSESSMENT SCORE FOR PREDICTING THE OUTCOME FOLLOWING ALLOGENEIC STEM CELL TRANSPLANTATION IN PATIENTS AGED OVER 50 YEARS: RESULTS FROM A PROSPECTIVE ITALIAN SINGLE-CENTER STUDY

Background: Allogeneic stem cell transplantation (allo-SCT) is increasingly used in older adults, and the Comprehensive Geriatric Assessment (CGA) has proven valuable for a more comprehensive health evaluation of allo-SCT recipients. Methods: We conducted a prospective CGA on 135 patients aged ≥50 years who underwent allo-SCT from 2020 to 2023. The CGA included Basic and Instrumental ADL/IADL, G8 Geriatric Screening Tool, VES-13, FIL score, Clinical Frailty Score, Gait Speed, Chair Stand, Hand Grip tests, Mini-Cog, Mini-Mental State Examination (MMSE), and HCT-CI. We performed a two-factor analysis with oblimin rotation and Bartlett estimation, using CGA components and sex, with continuous variables transformed into quartiles. Based on the estimated weights of each component a score was developed: [Gait Speed] + 2*[Hand Grip] + G8 + 1.5*[Sex]. The ROC analysis was used to classify patients into three categories based on increasing levels of fitness: low (n=13; score ≤13), moderate (n=62; score >13 and ≤22.5), and high (n=24; score >22.5). Results: The 1 and 2 year OS, NRM, and CIR were calculated for each CGA domain. Among physical domains, reduced hand grip strength was associated with lower OS(p<0.01) and higher NRM(p=0.05). In cognitive assessments, MMSE impairments correlated with lower OS (p=0.03). Disability-related domains, VES-13 and FIL scores, indicated reduced OS for unfit patients (p=0.04 and <0.01, respectively) and higher CIR (p=0.01 and <0.01, respectively). No significant differences were seen in OS, NRM, or CIR for other CGA domains. The HCT-CI, commonly used for patient selection in allo-SCT, failed to predict outcomes. The simplified CGA stratified patients into low, moderate, and high fitness groups. The low-fitness group had reduced OS (p=0.025) and a significantly higher risk of relapse (p<0.01) compared to the intermediate- and high-fitness groups (Figure 1). Factors such as diagnosis, disease risk index, previous therapy lines, conditioning intensity, donor type, age, time from diagnosis to allo-SCT, and graft-versus-host disease incidence were balanced across groups. Conclusions: Frailty assessment is not routinely used in hematology due to time constraints. We identified a simplified score based on three CGA domains and sex, which effectively predicts OS and CIR risk. Limitations include the relatively small sample size and diversity of hematological disorders, warranting further validation in larger cohorts.