Combined assessment of circulating tumor cells (CTCs) and measurable residual disease (MRD) for dynamic risk assessment of patients (Pts) with multiple myeloma (MM)

Abstract

Abstract Background: CTCs and MRD are rare tumor cell populations that respectively disseminate MM and drive relapse after treatment. If its combined assessment improves patients' prognostication, and if CTCs and MRD are genetically related, remains unknown. Aim: Analyze the complementary prognostic value and the genomic profiles of patient-paired CTCs and MRD. Methods: The European CTC consortium is composed of five collaborative groups (Czech Republic, Greece, Italy, HOVON/Netherlands-Belgium, PETHEMA/Spain) that ensembled patient-level data from 2,466 pts with newly diagnosed MM. Of these, 971 had CTC enumeration performed at diagnosis and MRD assessment throughout treatment with EuroFlow (71%) or other flow cytometry methods (29%). Up to 86% of pts were enrolled in clinical trials (GEM-CLARIDEX, CASSIOPEIA, HOVON-143, FORTE, GEM2012MENOS65) while 14% were treated in routine practice (Czech Republic, Greece). Median age was 59 (IQR 52-64); 83% of pts were transplant-eligible and 17% ineligible. Induction treatment included doublet (13%), triplet (77%) and quadruplet (9%) regimens. Progression-free (PFS) was landmarked at the time of MRD assessment. Median follow up was 6 years. Genomic data from tripartite bone marrow (BM) tumor cells and CTCs at diagnosis plus MRD at different time points of treatment was available in 29 MM pts enrolled in the GEM2012MENOS65 and GEM-CLARIDEX clinical trials. BM tumor cells at relapse were available in 3 of the 29 pts. All tumor cell populations were isolated based on patient-specific aberrant phenotypes by FACS. Whole exome sequencing was performed in 144 samples including peripheral blood T cells as germline control. Results: Overall, 641 of the 971 (66%) pts in the European pooled analysis achieved MRD negativity, whereas 330 (34%) remained MRD positive. Median PFS of MRD positive pts was gradually inferior in those with <0.001%, 0.001–0.01%, 0.01–0.1%, 0.1–1% and ≥1% CTCs at diagnosis (65, 38, 29, 30 and 16 months, respectively, p=.001). Of note, these inferior outcomes were not because of progressively higher MRD levels since the respective percentages were 0.01%, 0.02%, 0.04%, 0.01% and 0.01%. MRD negative rates were 76%, 58%, 67%, 61% and 63% in pts with <0.001%, 0.001–0.01%, 0.01–0.1%, 0.1–1% and ≥1% CTCs at diagnosis. Median PFS of MRD negative pts was gradually inferior in those with <0.001%, 0.001–0.01%, 0.01–0.1%, 0.1–1% and ≥1% CTCs: not reached, not reached, 85, 72 and 76 months (p<.001). The most significant differences in PFS were observed when MRD negative pts were stratified according to <0.01% vs ≥0.01% CTCs at diagnosis (median not reached vs 81 months, p<.001). In multivariate analysis of PFS adjusted for transplant-eligibility, induction regimen, R-ISS and 1q gain/amplification, CTC log levels and MRD status showed independent prognostic value. The hazard ratio (HR) of progression and/or death was 1.4 (95% CI 1 – 2, p=.03) in pts with 0.01–0.1% CTCs, 1.8 (95% CI 1.2 – 2.6, p=.003) if 0.1–1% CTCs, and 2.3 (95% CI 1.4 – 3.6, p=.001) if ≥1% CTCs at diagnosis. The HR of progression and/or death in pts with positive MRD was 2.9 (95% CI 2.3 – 3.7, p<.001). Genomic profiles of tripartite BM tumor cells, CTCs and MRD from 29 MM pts uncovered that 36% of single-nucleotide and copy number variants were concordant between CTCs and MRD cells. The Jaccard index of similarity was the lowest between CTCs and MRD (0.36) when compared to the indexes between BM tumor cells at diagnosis and MRD (0.40), or between BM tumor cells and CTCs at diagnosis (0.45). Individual-patient-level observations suggested progressive genomic diversity between CTCs and MRD cells prior to relapse. For example, one patient exhibited mutations in BRAF and DIS3 and del(13q) at diagnosis in BM tumor cells and CTCs, which persisted in MRD. After induction, de novo mutations in NEK11 and SLAMF7 emerged, followed by the appearance of subclonal copy number variants after HDT/ASCT contributing to resistance. Conclusions: While achieving MRD negativity was paramount to ameliorate the poor prognosis of pts with high CTCs, their PFS was not superimposable to those with low CTCs. These results support the integration of CTC levels at diagnosis in clinical-decision making about treatment adaptation in MRD negative pts. Genomic profiles of CTCs and MRD cells show considerable divergency. These results may help explaining why the assessment of CTCs and MRD is independently associated with clinical outcomes.