Advanced Risk Stratification in Multiple Myeloma Beyond Traditional FISH: The First Prospective Real-World Evidence for SKY92 Gene Expression Profiling

Abstract

Background The prognosis of multiple myeloma (MM) has been significantly improved in the last decades. However, high-risk (HR) MM still suffers from early disease progression. Therefore, identifying HR MM patients may be a crucial step to improve the prognosis of this patient group. To date, fluorescence in situ hybridization (FISH) is the most commonly used method for risk stratification in MM. In recent years, gene expression profiling (GEP) approaches, e.g. SKY92, have been developed for detection of HR MM patients. Currently, real-world data of MM risk stratification applying SKY92 alone or in addition to traditional FISH are still missing. In this study, we provide the first prospective real-world evidence of the prognostic value of the combined SKY92 and FISH risk stratification in newly diagnosed (ND) and relapsed/refractory (RR) MM. Methods We prospectively collected bone marrow (BM) samples and clinical data of 147 MM patients. Cytogenetics were analyzed on purified CD138 positive MM cells by FISH, and HR cytogenetics was defined applying the Revised International Staging System (del(17p), t(4;14), and t(14;16)) (Palumbo et al, J Clin Oncol. 2015). SKY92 risk classification was determined with MMprofiler gene expression assay (Kuiper et al, Leukemia 2012). In addition, whole genome sequencing (WGS) was performed, and 150 bp paired-end sequences were generated on the NovaSeq6000 sequencing instruments (Illumina) with 100x coverage. Structural variations (SV) were called using Manta, and copy number variations (CNV) calling was performed with GATK4. Strelka2 variant caller was used for single nucleotide variations (SNV). Variants with global population frequencies >1% were excluded. Results We included 147 patients in our study (NDMM: n=51, RRMM: n=96). SKY92 classification was available for 121 (82.3%) patients. The median follow-up time was 344 days. HR SKY92 was significantly enriched in RRMM (40/76) compared with NDMM (6/45) (P<0.0001). RRMM patients with HR SKY92 showed significantly shorter progression free survival (PFS) (P<0.0001) and overall survival (OS) (P=0.0004) than standard-risk (SR). In NDMM, HR SKY92 also indicated a significantly inferior PFS (P=0.001) in comparison with SR. Of note, samples of 26 (17.7%) patients did not meet the SKY92 quality control criteria due to low BM infiltration (median BM infiltration: 25% vs 55% in other patients, P=0.038). We then combined the SKY92 classification with cytogenetic analyses by FISH, which were performed in 99 patients (NDMM: n=33; RRMM: n=66). We noticed a discrepancy between both risk stratification systems, with 44 (44.4%) and 42 (42.4%) patients being classified as HR by SKY92 and FISH, respectively. In total, 24 (36.3%) RRMM patients and 6 (18.2%) NDMM patients displayed HR in both SKY92 and FISH (double-HR) (Figure). Regarding survival outcome in RRMM, double-HR patients showed the worst PFS (P<0.0001) and OS (P<0.0001) when compared to the other patients (Figure). Noteworthy, RRMM patients with only HR SKY92 demonstrated significantly worse PFS than those with only HR FISH (P=0.007). In addition, double-HR presented a negative prognostic factor for PFS in NDMM (P=0.01). To further elucidate the discrepancy between FISH and SKY92 classifications, we performed WGS in 12 patients who exhibited either HR SKY92 only (n=9) or HR FISH only (n=3). Interestingly, 1 patient with bi-allelic TP53 inactivation (deletion + mutation) and 2 patients harbouring t(4;14) or t(14;16) were determined as SR by SKY92 but as HR by FISH; all the 3 patients were still in remission at the time point of data cut-off (July 2022). Moreover, we found abnormalities at chromosome 1 in 7 out of the 9 patients with HR on SKY92 only (gain1q21: n=6; del1p32: n=2) and, notably, gain1q21 and del1p32 were detected only by WGS in 3 and 2 patients, respectively. The remaining 2 patients did not show any known HR SV/CNV/SNVs, suggesting that HR MM may be associated with some other mechanisms, e.g. epigenetic modifications. Conclusion Altogether, risk classification using SKY92 performed well for detecting HR MM in the clinical routine. Here, we provide the first prospective real-world evidence that advanced risk stratification combining SKY92 and FISH may help to identify the "highest-risk" MM, which requires intensive anti-MM therapy and close monitoring in the course of the disease. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal