Blood Flow: a new applicable and ultra-sensitive test to monitor measurable disease in peripheral blood of multiple myeloma

Abstract

BACKGROUND AND AIMS OF THE STUDY. Assessing Minimal Residual disease (MRD) in bone marrow (BM) using next-generation flow (NGF) or sequencing (NGS) precludes periodic evaluations because of its invasiveness. MRD assessment in PB could overcome this limitation, but its prognostic value is not established and the negative predictive value (NPV) when compared to BM is < 70%. This is because tumor burden in PB is ~3log lower than in BM, and methods capable of detecting MRD below 10-6 are thus needed for improved concordance. We first aimed at investigating the prognostic value of MRD assessment in PB using NGF. Our second aim was to develop a new method with increased sensitivity.

PATIENTS AND METHODS. MRD was evaluated using NGF in PB of 138 MM patients enrolled in the GEM2014MAIN trial. PB samples were collected after the second year of maintenance, when patients stopped treatment if MRD negative (in BM), or continued on therapy for three additional years if MRD positive at that time. Reaching a minimum sensitivity of 10-7 requires analyzing ≥ 2x108 cells (~50mL of PB). To avoid high staining costs and impractically long acquisition periods, a new method integrating immunomagnetic enrichment using MACS® MicroBeads prior NGF was developed and coined as BloodFlow. Large PB volumes were magnetically labeled and processed via MACS® columns, and ~100µL aliquots enriched with circulating plasma cells (PCs) were analyzed using EuroFlow NGF. The concordance between MRD assessment using BloodFlow (in PB) vs NGF (in PB and/or BM) was analyzed in 389 samples from 351 MM pts.

RESULTS. Of the 138 patients enrolled in the GEM2014MAIN trial having MRD assessed in PB, 15 (12%) showed positive MRD. Their median progression-free survival (PFS) since MRD testing was 22 months, which was significantly inferior vs those with undetectable MRD in PB (median not reached). The respective rates of PFS at two years were 50% and 98% (HR: 11.7; p<.00001). Among the 123 patients with undetectable MRD in PB, 33 (27%) showed persistent MRD in BM and inferior PFS vs those with undetectable MRD in PB and BM. The respective rates of PFS at two years were 62% and 100% (p<0.0001). The results from this part of the study confirmed the prognostic value of MRD assessment using NGF in PB, and emphasized the importance of increased sensitivity to reduce the number of false-negative results in PB vs BM. In the second part of the study, an optimized BloodFlow protocol was developed after comparing various lysing methods and MicroBeads combinations for optimal enrichment of circulating PC. Initial testing in PB samples from healthy individuals showed on average an 82-fold increment in the number of circulating normal PC with BloodFlow vs NGF. Dilution experiments with MM cell lines showed detection of up to 1x10-7 tumor cells. The performance of BloodFlow vs NGF in PB was compared in 353 samples. BloodFlow detected MRD in 33/353 (9%) samples. The lowest MRD level was 6x10-8. Of the 33 samples with detectable MRD using BloodFlow, 19 (58%) were negative by NGF. All cases with undetectable MRD according to BloodFlow were also negative by NGF. Subsequently, we compared the performance of BloodFlow vs NGF in 199-paired PB and BM samples. Concordance was observed in 137 (69%) double-negative and 19 (9.5%) double-positive samples. MRD was detected in BM and not in 41 (20.5%) PB paired-samples, while 2 (1%) were negative in BM but positive in PB (both showing MRD reappearance in BM soon after). Thus, BloodFlow showed a NPV of 77% when compared to NGF in BM. MRD assessment during induction and intensification was the feature more frequently associated with a false-negative result using BloodFlow (26/41 [63%]), followed by reduced PB cellularity (15/41 [37%]) and MRD levels below 10-5 in BM (12/41 [29%]). In the GEM2014MAIN trial, 2 of 4 pts with positive MRD in PB using BloodFlow progressed, whereas none of the 29 patients having undetectable MRD relapsed thus far.

CONCLUSIONS. MRD assessment in PB using NGF was prognostic in patients under maintenance or observation. Notwithstanding, a new method (BloodFlow) with an unprecedented sensitivity to detect MRD down to 10-8 in PB of MM patients, was developed to increase the NPV. BloodFlow detected MRD in PB more frequently than NGF, with a consequent decrease in the number of cases with persistent MRD in BM while undetectable in PB, which were more frequent during early and intensive stages of treatment. These results suggest the possibility of periodic and ultra-sensitive MRD assessment in PB during maintenance or observation.