Consultant Haematologist and Adjunct Research Fellow Alfred Health-Monash University, Melbourne, VIC, Australia, Australia
Introduction: Extramedullary disease (EMD) affects 30% of multiple myeloma (MM) patients, predicting poor overall survival due to aggressive disease kinetics and multi-drug resistance. Current methods for detection and monitoring include PET/CT scans; alternatives such as bone marrow (BM) biopsies and consensus response criteria (CRC) are limited as EMD is frequently non-secretory or with minimal BM involvement. EMD is associated with driver mutations (DM) in the MAPK pathway (KRAS, NRAS and BRAF). We have shown DM are detectable in cell free DNA (cfDNA) in EMD patients and now sought to clarify the role of cfDNA characterisation in EMD patients.
Methods: Plasma was collected in Streck DNA tubes. cfDNA was extracted using QIAGEN QIAamp® Circulating Nucleic Acid Kit. DM were identified by whole genome sequencing or whole exome sequencing. Once identified, droplet digital PCR was used to detect DM in cfDNA at additional timepoints, including prior to EMD development, after treatment and at relapse.
Results: 13 patients were included. 100% had the EMD DM detected in the blood at the time of EMD diagnosis with variant allele frequency (VAF) ranging from 0.05% to 37.63%. At diagnosis, low cfDNA VAF associated with low disease burden on PET/CT, however high VAFs do not necessarily reflect PET/CT burden and may reflect factors such as tumour necrosis.
Eight patients had at least two cfDNA timepoints. Changes in the cfDNA VAF over time mirrored PET-CT-defined response to therapy and relapse. We then compared cfDNA levels after therapy with other response parameters- PET/CT, CRC and EuroFlow minimal residual disease (MRD). Response assessments by cfDNA appeared complementary to PET/CT, with 3 cases showing incongruity between PET/CT and cfDNA; 2 had cfDNA positivity with PET/CT negativity and 1 cfDNA negativity but PET/CT positivity. In patients with secretory disease, cfDNA was more sensitive compared to CRC, with DM detectable in cfDNA in 3 patients in a CR. cfDNA assessment also outperformed MRD testing, with 40% of MRD negative timepoints demonstrating cfDNA positivity.
Importantly, changes in VAF were prognostic. Patients achieving cfDNA negativity had the longest remissions (median 23.5 months) whereas those who did not demonstrated short time to progression (median 6 months). In 2 patients who experienced EMD relapse from cfDNA negativity, cfDNA became positive prior to overt relapse by PET/CT, CRC, or BM MRD. Finally, DM were detectable in cfDNA prior to the initial development of EMD, in one case two years before EMD diagnosis. Liquid biopsies from an additional 24 patients are being analysed.
Conclusions: In EMD patients MAPK activating DM are detectable in cfDNA. cfDNA is complementary to PET/CT scanning, more sensitive than BM MRD and CRC, and is a predictor of EMD relapse. DM exist in cfDNA prior to EMD, and their detection may allow tailored monitoring strategies and early intervention.
