A 5-Stage System for ASCT Multiple Myeloma

652. Multiple Myeloma: Clinical and Epidemiological

INTRODUCTION:

In the last three decades, significant advancements in multiple myeloma (MM) treatment, including immunomodulators, targeted therapies, chemotherapy combinations, autologous stem cell transplantation (ASCT), and CAR-T therapy, have emerged. ASCT offers hope for MM patients by using high-dose chemotherapy with stem cell support. However, it carries risks of severe toxicity. Despite evaluating factors like age, performance status, and co-morbidities for ASCT eligibility, outcomes for MM patients still vary greatly. Various models, such as the ISS-series systems, have been developed to assess risk factors for progression-free and overall survival We conducted a study on 5,259 ASCT patients since 1989, identifying 7 easy variables (ISS stage, diagnosis-transplant interval, ferritin, transferrin, LDH, age, and gender) and creating a 5-stage system. This system helps evaluate the risk of MM patients before ASCT.

METHODS:

In this retrospective study, data from 5,259 MM patients receiving ASCT (1989-2022) were collected (median follow-up: 57 months). We assessed 30 clinical variables pre-ASCT. Multiple imputation was used to deal with the missing values. Patients were split into a 70% training set and a 30% validation set. Univariate/multivariate Cox regression analysis identified 7 clinical and demographic signatures. Signatures were scored based on hazard ratios normalized to ISS III vs. ISS I&II. K-adaptive partitioning generated a 5-stage system, ATM5S, evenly distributed and validated in the training set. Harrel’s c-index was used to compare ATM5S with the current myeloma systems.

RESULTS:

3683 patients (70%) in the test cohort had a median age of 59.6, median PFS 38.8 month and median OS 56.1 month. IgG was the most common isotype (55%). Univariate Cox analysis showed that 27 variables were significant predictors for both PFS and OS. Multivariate Cox analysis showed that time from diagnosis to transplant, ferritin, transferrin, LDH, ISS stage, lymphocytes, platelets, calcium, gender and age were independent prognostic factors for both OS and PFS. We selected the most easy, understandable and well-distributed (with high-risk group>20%) 7 variables: ISS, time from diagnosis to transplant, ferritin, transferrin, LDH, age and gender and developed a score system, which resulted into a 5-stage system (ATM5S) by k-adaptive partitioning. This staging system also demonstrated a significant discriminatory effect in the validation set (30%, 1576 patients).

CONCLUSIONS:

The Arkansas Transplant-Myeloma 5-Stage System (ATM5S) showed a promising performance in both training and validation cohorts.