Multiple myeloma (MM), a terminally differentiated plasma cell malignancy, is one of the most clinically and genetically heterogeneous cancers. Even in the era of novel therapeutic agents and improved survival, MM treatment remains challenging, and myeloma remains an incurable disease. In that regard, much effort has been put into understanding the pathogenesis and predicting prognosis by dissecting MM into different classifications through variables contributing to its disease course. This led to the emergence of many studies that identified criteria involved in staging and risk-stratification. While clinical lab values have been associated with prognosis and treatment outcome, the genomic landscape of myeloma seemed to be a prominent determinant of prognosis. As such, the International Myeloma Working Group (IMWG) has stratified MM patients into three distinct risk classifications: standard risk (hyperdiploidy, t[11;14], t[6;14]), intermediate risk (t[4;14], del[13q]), and high risk (del[17p], t[14;16], and t[14;20]). Notably, the International Staging System (ISS), which previously staged patients purely according to clinical variables, was later revised to include lactate dehydrogenase (LDH) level and cytogenetics in its assessment, and has been the gold standard staging system for a long time. While an overlap does exist, unfortunately, there is still a significant difference in prognosis when using different systems for risk-stratification, which negatively affects clinical decision-making.
With the huge advancement in genomic technologies during the past decade, more specific molecular classifications were discovered and proposed for risk-stratification. Although gene expression profiling (GEP) on microarrays was heavily studied and yielded high-risk signatures that were incorporated into the University of Arkansas for Medical Sciences (UAMS) and the Intergroupe Francophone du Myélome models, it is rarely used as a prognostic tool. Conversely, large data sets of next-generation DNA sequencing revealed a relatively small set of commonly mutated genes in MM with associated prognostic value on patient survival. These included KRAS, NRAS, BRAF, FAM46C, DIS3, TRAF, CYLD, and LTB with a neutral prognosis; TP53, ATM, and ATR with a poor prognosis; and IRF4 and EGR1 with a favorable prognosis.1 However, given the genomic complexity of myeloma, we believe that more alterations will be discovered at the epigenetic level as well as in the intronic regions, with ongoing whole-genome sequencing studies being done on tumor samples of large cohorts, which could in turn be used as prognostic biomarkers.
Around two years ago, the Myeloma Genome Project (MGP) was announced at the 58th ASH Annual Meeting. The project aims to integrate high-quality genomic data of newly diagnosed treatment-naïve MM patients, collected by UAMS Myeloma Institute, the Myeloma XI trial (United Kingdom), Intergroupe Francophone du Myélome/Dana-Farber Cancer Institute, and the Multiple Myeloma Research Foundation (MMRF). Because of this initiative, we were able to discover more about the intricate genomic network in myeloma, as apparent in many of the published articles. Recently, Dr. Brian A. Walker and colleagues studied 784 eligible subjects of the 1,273 MM patients who are part of MGP. Their work demonstrated a new class of myeloma patients, labeled as the “high-risk, double-hit group.” This is actually a follow-up study to another recently published article by the same group, in which analysts used integrative genomics on the same dataset of 1,273 MM patients to identify 63 driver genes, including novel somatic mutations, which are associated with worse outcomes and dependent on particular genomic alterations such as copy number alterations, translocations, and hyperdiploidy.2
In this article, Dr. Walker and colleagues implemented a recursive partitioning model on existing sequencing data to discover that a worse prognosis could be predicted in MM patients in the presence of a bi-allelic inactivation of TP53 or the amplification (≥4 copies) of CKS1B (1q21) in the context of ISS III staging. This subset of patients, the double-hit group, has a significantly diminished median progression-free survival (PFS) of 15.4 months and median overall survival (OS) of 20.7 months compared with the low- and intermediate-risk patients. Their model defined “low-risk” as patients having low or intermediate disease stage without any genetic factors, while the intermediate-risk was defined by either ISS I plus t(4;14) or del(17p), or ISS III with no genetic factors. Interestingly, the bi-allelic inactivation of TP53 turned out to be a prominent determinant of PFS and OS prognosis; upon review, del(17p) is no longer prognostically relevant, signifying the higher importance of DNA sequencing of TP53 compared to FISH analysis of a 17p loss. Conversely, an amplification of more than four copies of CKS1B, which constituted a small fraction of 6.3 percent of patients, was associated with an even worse PFS and OS compared to just a gain of CKS1B, which was found in 21.9 percent of patients. While other mutations were found to be significantly associated with poorer outcomes at a univariate level, they were excluded in the multivariate model. Though this could truly mean that there is no significant effect of these mutations on prognosis, it could also signify that longer follow up is warranted to truly discern the effect of these genetic alterations on progression, relapse, and treatment resistance.
In Brief
The importance of including the mentioned genetic factors mentioned here into a future risk-stratification model is illustrated by the fact that patients in the double-hit group performed poorly, irrespective of IMWG risk stratification, while patients classified as high-risk by IMWG criteria and low-risk by recursive partitioning had OS and PFS similar to that of those classified as low-risk by IMWG criteria and intermediate-risk by recursive partitioning. This raises a longstanding concern: not having a concrete risk-stratification model that clearly guides clinical management in MM. However, we are on the right path to achieving that goal. Most notably, the MMRF’s Clinical Outcomes in MM to Personal Assessment of Genetic Profile (CoMMpass, NCT0145429) study3 is a significant initiative to collect clinical and genomic data on newly diagnosed MM patients with long-term follow up and sequential collection of biological samples to monitor genomic alterations with time. To date, researchers have collected genomic data on more than 1,200 patients, including whole-exome, whole-genome, and targeted DNA sequencing data, as well as RNA sequencing of tumor samples. It is with such big initiatives that we will be able to map out the genomic landscape of this complex disease to predict survival and treatment response, and devise targeted therapies accordingly. Indeed, because of the current genomic studies of myeloma, several clinical trials that aim at targeting defective genes and pathways have already been conducted or are still ongoing (i.e., RAS/MAPK, PI3K/Akt, Cyclin D, and the BCL2 pathway). Just imagine what would happen when we further decipher the significance of intronic alterations and identify additional epigenetic modulators in thousands of newly diagnosed, relapsed, and treatment-refractory myeloma patients: The newly identified double-hit group would be just the beginning of many risk-stratifiers to come.
References
Competing Interests
Dr. Mouhieddine and Dr. Ghobrial indicated no relevant conflicts of interest.
