Drug Response Prediction in High Risk Multiple Myeloma

Abstract

The aim of this study is to test if we could predict progression free survival (PFS) and drug response in high risk myeloma by a new method based on gene expression profiling (GEP). This may provide us with a tool for effective personalized medicine and to prevent any potential non-responders from suffering from drug induced side-effect. A specific Drug Response Predictor (DRPTM) has been developed by Medical Prognosis Institute, DK. The DRP is an assay that based on gene expression analysis of cell lines and tumor samples, can predict the response to specific drugs. High-risk myeloma patients were identified by use of the GEP of 70 genes (GEP70) which has been validated in the transplant, non-transplant, and relapse setting1. We used publically available GEP from patients at diagnosis to predict the response by DRP of drugs used in the treatment of myeloma patients2;3. We stratified patients by GEP70 and by virtual karyotyping of t(14;14) and t(14;16). Furthermore, we used DRP to test for predicted sensitivity to a panel of B-cell active drugs. Multidimensional scaling presented high and low risk myeloma by GEP70 as a continuum pattern rather than two separate clinical populations. The median predicted sensitivity by DRP for all patients varied between 40 to 70, on a scale of 0-100 for melphalan, bortezomib, etoposide, doxorubicin, cisplatin, vincristine, lenalidomide, thalidomide and cyclophosphamide and was higher among high-risk patients. The predicted sensitivity to drugs by DRP was highest for IMiDS, vincristine, doxorubicin and bortezomib. Tumor samples with high expression levels of MMSET had lower predicted sensitivity to all drugs. Among patients with high gene expression levels of cMAF the highest predicted sensitivity was found for IMiDS, bortezomib and vincristine. Twenty-five percent of high-risk patients by GEP70 have a PFS of more than 10 years. The DRP score stratified patients further. Patients with a predicted sensitivity by DRP to melphalan, bortezomib and lenalidomide had a prolonged PFS (melphalan: HR 2.4 [ 1.2-4.9], P= 0.014; bortezomib: HR 5.7[ 1.2-27], P= 0.027; lenalidomide HR 3.8 [ 1.2-13], P= 0.028) and a trend was found for thalidomide HR 2.3 [0.6-8.8] (P= 0.21). Patients with predicted sensitivity to bortezomib had a better response to treatment (P= 0.022). DRP was applied to a panel of B-cell active drugs and a predicted response to treatment was found for all high-risk myeloma patients. In newly diagnosed myeloma patients we find that DRP can predict the outcome to melphalan, bortezomib and lenalidomide and also the response to bortezomib. DRP also predicted response to drugs used in treatment of other B-cell diseases. We will use the results for up-coming clinical trials. (1) Johnson SK et al. Int J Hematol 2011;94:321-333. (2) Shaughnessy JD, et al. A. Blood 2007;109:2276-2284. (3) Terragna C et al. Oncotarget 2016;7:9666-9679.