Prospective real-world evaluation of SKY92 for risk stratification in multiple myeloma: Comparison with updated ims/IMWG criteria in the prommis study

Abstract Introduction: Accurate risk stratification in multiple myeloma (MM) is essential to guide treatment decisions and improve outcomes, particularly for high-risk (HR) patients who continue to face poor prognoses despite therapeutic advances. Current clinical practice relies predominantly on cytogenetic markers to define risk, which may not fully capture disease heterogeneity. Gene expression profiling (GEP), such as the SKY92 classifier, offers a complementary tool to refine risk assessment. The PROMMIS study (NCT02911571) is a prospective, multicenter trial evaluating the added clinical value of SKY92 in real-world settings. Recently, the IMS/IMWG proposed revised criteria for high-risk MM (HRMM), incorporating refined DNA-level definitions of high-risk lesions—including TP53 mutations, 1q amplification, and 1p loss— while also placing greater emphasis on renal function and gene expression profiling. Specifically, HRMM is now defined by one or more of the following: (1) del(17p) >20% clonal fraction, and/or TP53 mutation; (2) one of t(4;14), t(14;16), or t(14;20) along with gain(1q) and/or del(1p32); (3) monoallelic del(1p32) along with gain(1q) or biallelic del(1p32); or (4) β2M ≥5.5 mg/L with creatinine <1.2 mg/dL. We aimed to evaluate the prognostic value of SKY92 in comparison to the revised IMS/IMWG criteria, using a proxy based on available variables within PROMMIS. Methods: GEP was performed on CD138+ plasma cells from 251 newly diagnosed MM (NDMM) patients using SKY92 (SkylineDx, San Diego). RNA was extracted via immunomagnetic separation with ≥80% purity. As not all revised variables were available, a proxy for IMS-HR was defined by: (1) del(17p), (2) co-occurrence of gain(1q) with t(4;14) and/or t(14;16) (detected by iFISH), or (3) β2M ≥5.5 mg/L with creatinine <1.2 mg/dL. This data was available for 237 patients. Progression-free survival (PFS) and overall survival (OS) were available for 243 patients (median follow-up: 60 months). Cox proportional hazard models assessed the prognostic value of SKY92 and the IMS proxy. Results: Among 251 patients, SKY92 classified 180 (72%) as standard-risk (SR) and 71 (28%) as high-risk (HR). SKY92-HR patients showed significantly inferior PFS (HR=2.3, 95% CI: 1.6–3.7; p<0.0001) and OS (HR=3.4, 95% CI: 1.8–6.5; p<0.0001). Using the IMS proxy, 61 out of 237 patients (26%) were identified as HR (del[17p], n = 39; t(4;14)+gain(1q), n = 14; t(14;16)+gain(1q), n=7; β2M + Creatinine, n=4). However, no significant difference in PFS (HR=0.90, 95% CI: 0.6–1.4; p=0.60) or OS (HR=0.85, 95% CI: 0.40–1.8; p=0.70) was observed between IMS-HR and IMS-SR patients. Risk category overlap showed that 26 patients (11%) were HR by both IMS proxy and SKY92, 35 (15%) were IMS-HR but SKY92-SR, and 43 (18%) were SKY92-HR but IMS-SR. The latter group (IMS-SR;SKY92-HR) had outcomes comparable to patients classified as HR by both. Conversely, IMS-HR; SKY92-SR patients showed survival similar to those classified as SR by both. In a multivariate analysis SKY92 remained independently prognostic for both PFS (HR=2.3, 95% CI: 1.6–3.4; p<0.001) and OS (HR=3.6, 95% CI: 1.9–6.8; p<0.001). Conclusions: This study presents the first comparison of the SKY92 GEP against the updated IMS/IMWG criteria for HRMM. Due to incomplete data availability in PROMMIS, we approximated the IMS/IMWG definition using a proxy based on accessible variables. While this limits direct comparability to the full criteria, it enabled the introduction of key updated elements, such as gain(1q), and the β2M/creatinine combination. SKY92 identified a clinically relevant subgroup of patients with inferior outcomes that were not captured by the IMS/IMWG proxy. Conversely, patients classified as HR by IMS/IMWG criteria but SR by SKY92 had outcomes comparable to SR patients overall, raising the possibility of overclassification by cytogenetics alone. In multivariate models, SKY92 retained independent prognostic value for both PFS and OS. These findings highlight the complementary role that SKY92 can play in risk stratification when used alongside cytogenetics and clinical variables. Further studies are needed to validate SKY92 against the complete IMS/IMWG definition, including features not assessed here (i.e., TP53 mutation, del(1p32) and t(14;20)). Nonetheless, this analysis supports the integration of gene expression profiling into routine clinical risk assessment to enable more accurate and personalized treatment strategies for patients with MM.