Next-Generation Sequencing to Assess the Impact of Low Dose Melphalan on Residual Disease before High-Dose Melphalan and Stem Cell Transplant in Multiple Myeloma

High-Dose Melphalan (HDM) followed by stem cell transplant (SCT) remains the standard-of-care for transplant-eligible patients newly-diagnosed with multiple myeloma (MM). However, ~1/3 of patients relapse <2 years after undergoing HDM-SCT, indicating that melphalan-sensitivity is limited to a subset of patients and is currently not predictable. Currently, models that predict melphalan-resistance before proceeding to transplant are lacking. Rather, transplant-eligibility is defined mostly based on adequate organ function and performance status. Therefore, there is an urgent and unmet clinical need to develop strategies that accurately predict melphalan sensitivity among MM patients prior to HDM-SCT and save melphalan-resistant patients from undergoing this highly morbid procedure, if no demonstrable benefit is expected from it.

Traditional disease-measurement methods based on International Myeloma Working Group (IMWG) criteria rely on the secretory function of myeloma cells and measure monoclonal protein levels. Following induction therapy, pre-transplant monoclonal protein levels are usually very low, and further reduction in myeloma secretory function are not detectable. In addition, the long half-life of monoclonal proteins makes assessing short-term disease changes problematic. Methods to accurately detect minor changes in disease burden following a low dose of melphalan (LDM) as a marker of melphalan sensitivity are needed to better predict patient responses to LDM. Next-generation sequencing (NGS), is an alternative approach that may allow for the highly sensitive, rapid, real-time detection of minuscule changes in tumor volume that are not influenced by the long half-life of monoclonal proteins.

Here, we propose to use NGS-based tumor assessment to evaluate changes in disease volume following LDM before proceeding to HDM-SCT. Evidence is lacking to determine whether a single LDM generates a decrease in myeloma burden that is measurable by NGS. Our central hypothesis is that NGS of bone marrow aspirates from newly-diagnosed, post-induction transplant-eligible MM patients will provide a method to precisely determine the effect of LDM on disease burden. ClonoSEQ assay is an FDA-cleared, highly sensitive, specific, and standardized method to detect and monitor MRD, in MM patients. clonoSEQ leverages the power of NGS and offers an accurate and reliable way to assess how disease burden changes over time in response to treatment. Therefore, we propose a proof-of-principle study to assess the validity of this strategy and to provide essential data for future trial design investigating individualized approaches based on NGS sequencing and low doses of therapeutic agents. We will test the central hypothesis that LDM, administered at 16 mg/m ², generates a detectable reduction in tumor burden measured by NGS. A detectable reduction in tumor burden is defined as a ≥ 20% decrease in NGS clonal count in at least 30% of subjects. We will administer propylene glycol-free melphalan formulation (EVOMELA) due to greater stability upon reconstitution than AlKERAN formulation in order to diminish the variability in the effective administered dose. The primary and secondary objectives and endpoints of the study are listed in Table-1,2.

Statistical Considerations: Clonoseq detects measurable residual disease at the level of a single cell given sufficient sample input. The specific hypothesis of this pilot trial is LDM produces a measurable disease reduction that is readily detectable by clonoSEQ with at least a 20% reduction in at least 30% of patients. Assuming a 100% yield for VJD clonal sequencing and calibration efficacy by clonoSEQ, the sample size required to test the null hypothesis of 5% patients with positive MRD test against alternative 30% patients with positive MRD test is 16 patients. The sample size estimation is using two-sided chi-square test with 80% power. The sample size estimation is n = 21, when power = 90% based on one sample Binomial distribution theory. We will assume 20% failure rate for VJD clonal sequencing and calibration efficacy by clonoSEQ. Therefore, by enrolling 20 patients, we expect that at least 16 patients will have MRD assessable by NGS method.

Figure 1

View largeDownload slide

Figure 1

View largeDownload slide

Close modal