P-239: Dynamic Risk Model: A Novel Approach Incorporating Functional High Risk Factors for Predicting Survival Outcomes in Patients with Relapsed/Refractory Multiple Myeloma

Introduction: Early relapse and refractoriness to novel agents have been recognized as high-risk features associated with resistance to subsequent therapy and reduced overall survival in patients with multiple myeloma (MM). Newly identified risk factors following such treatments are referred to as functional high risk or dynamic high risk. In this study, we propose a novel risk model aimed at predicting survival outcomes in patients with relapsed/refractory MM (RRMM), incorporating functional high risk factors.

Methods: We conducted a retrospective review of medical records for RRMM patients who underwent a second line of therapy at Kyungpook National University Hospital and Chungnam National University Hospital between 2011 and 2021. Early relapse was defined as biochemical or overt relapse occurring within 18 months after the initiation of frontline therapy. We assigned weighted scores based on the range of hazard ratio (HR) derived from the Cox hazard model for each risk factor. Subsequently, we classified risk groups by summing the scores for each risk factor.

Results: A total of 141 patients with RRMM were included in the analysis. The median age at diagnosis was 65.7 years (range: 33-80). Early relapse was observed in 77 patients (54.6%). Following relapse, all of the patients received carfilzomib (72.3%)- or ixazomib (27.7%)-based treatment. Early relapse was identified as a significant risk factor for survival prediction (HR 2.56, p= 0.002). Other variables at relapse, including hemoglobin (Hb), platelet count, LDH, albumin, beta 2 microglobulins (B2GM), and plasmacytoma, were determined to be relevant risk factors affecting survival outcomes. The scores assigned were 2 for early relapse, LDH, and plasmacytoma, 0.5 for Hb, and 1 for albumin. Platelet count and B2GM were excluded from the weighting score calculation due to their relatively insignificant HR values. The risk model for survival prediction, referred to as the Dynamic Risk Model, classified RRMM patients into three stages: stage I (total score 0-2), stage II (2.5-4), and stage III (4.5-7.5). Seventy patients were included in the survival outcome analysis using the Dynamic Risk Model, after excluding 71 patients due to missing laboratory or imaging data. Of the analyzed patients, 29 (41.4%) were classified as stage I, 25 (35.7%) as stage II, and 16 (22.9%) as stage III. The new model effectively differentiated survival outcomes among the stages in RRMM patients, with 24-month OS rates of 95.7%, 42.3%, and 16.2% in stages I, II, and III, respectively (p < 0.001). The 24-month progression-free survival rates were 57.3%, 35.7%, and 0.0% in stages I, II, and III, respectively (p < 0.001).

Conclusions: The integration of early relapse and readily accessible variables within the Dynamic Risk Model demonstrated a remarkable ability to prognosticate survival outcomes in patients with RRMM. Significantly, this model represents the first application of functional high risk in reevaluating risk stratification within the RR setting.