P-174: Early relapse within 18 months (ER18) is a powerful dynamic predictor for prognosis and could revise static risk distribution in patients with new-diagnosed multiple myeloma

Introduction: Due to the wide application of novel agents, the survival outcomes of patients with newly-diagnosed multiple myeloma (NDMM) have significantly improved in recent years. However, almost NDMM patients eventually relapse owing to the acquisition of drug resistance and sustained residual disease. However, the definitions of ER in MM vary from study to study.

Methods: The study was based on the National Longitudinal Cohort of Hematological Diseases in China (NCT04645199). A total of 629 patients diagnosed with NDMM between January 2013 and December 2019 were included in this study. We evaluated different ER definitions and investigated the underlying relationship between ER with static risk distribution.

Results: We first summarized all studies about ER published to date and found that relapse within 18 months from the initial treatment or within 12 months from SCT was mostly used to define early PD in many studies. Our findings indicate that ER18 could effectively balance the bias of ER definitions for patients with or without ASCT, as well as fulfill all requirements for a dynamic risk predictor. The ER18 population (114/587; 19.4%) presented with more aggressive biologic features compared to a reference cohort (P < 0.001), with a significant short median overall survival (OS) of 28.9 months. We also described the specific transitions from static risk profile to dynamic risk distribution in our cohort and then constructed a mixed-risk pattern to identify four novel populations with distinct survival outcomes (P < 0.001). A total of 367 (67.0%) patients presented with at least one high-risk feature: those with ISS III stage, elevated LDH, and HRCAs were classified as baseline high-risk (BHR) group; 86 (23.4%) patients experienced ER18 and further defined as mixed high-risk (MHR) population, and the remaining patients were classified as static high-risk (SHR). Within patients without BHR features, the dynamic risk event (ER18) occurred in 21(11.6%) patients, defined as functional high-risk (FHR). As expected, the MHR population had the worst outcomes with a median OS of 25.9 months (P < 0.001). Of note, FHR patients had similar survival compared to SHR patients (OS: not reached vs. 71.4 months; P=0.235; Figure 3B), supporting the importance of dynamic risk factor (ER18) as a prognostic factor and further suggesting that lack of durable response could exert an adverse impact on prognosis similar to baseline high-risk variables. Moreover, we confirmed that ER18 refined the predictive accuracy of the Revised International Staging System stage (R-ISS).

Conclusions: ER18 maintains its significance as a predictor of prognosis in the multivariate analysis. The underlying correlation between clinical features at diagnosis and dynamic predictor (ER18) was demonstrated. Our results further indicated that ER18 was a refining factor for the R-ISS staging system and that the refined model could actualize a more valuable risk distribution.