Circulating tumor cells (CTCs) for dynamic risk assessment of patients (Pts) with smoldering multiple myeloma (SMM)

Abstract

Abstract Background: A single risk assessment using the IMWG 20/2/20 model may be insufficient for individualized management of SMM pts. Serial assessments could improve risk-stratification through the identification of stable vs evolving patterns, but bone marrow (BM) aspirates cannot be performed repeatedly for quantification and characterization of tumor cells. Aim: Investigate the role of CTCs as an alternative to BM tumor cells for dynamic risk assessment of SMM. Methods: The iMMunocell study enrolled 324 untreated SMM pts from 2017 to 2024. A total of 1,250 blood samples were collected every 6 months during 3 years to monitor CTCs and immune profiles using next-generation flow cytometry. CTCs were isolated by FACS for longitudinal exome sequencing. Clinical data were updated every 6 months, resulting in a total of 13,659 longitudinal datapoints that included serological biomarkers, BM assessment of tumor burden with conventional microscopy and flow cytometry, cytogenetics, and imaging with MRI and/or PET/CT. Dynamic risk assessment using single or combined features measured in ≥3 timepoints was performed with CoNNector. Baseline and longitudinal immune monitoring data will be presented at the meeting. Results: Median age of the 324 SMM pts was 67 years. Patient distribution into low, intermediate and high-risk disease according the IMWG 20/2/20 model was 40%, 36% and 24%. With a median follow-up of 46 months, 108 of the 324 (33%) pts progressed to active MM. At baseline, CTCs were detected in 235 of the 324 (73%) SMM pts (median 0.002%, range 0.0002% - 0.42%). The optimal CTC cutoff to stratify pts in two risk groups was 0.008% (HR: 3.5, p<.001). Pts having >0.008% CTCs at baseline showed median time-to-progression (TTP) of 26 months vs not reached in those with ≤0.008% CTCs. Higher CTC cutoffs such as ≥0.25% defined a subgroup of pts who all progressed to active MM in 2y. In a multivariate analysis of TTP including the risk factors >2g/dL M-component, >20 sFLC ratio, >20% BM tumor cells and >0.008% CTCs, all except BM tumor cells showed independent prognostic value. The C-indexes of the 20/2/0.008 and the 20/2/20 models were 0.85 and 0.81. Being minimally-invasive, the 20/2/0.008 model could be reassessed every 6 months. It was noteworthy that 37% of pts had their risk of transformation modified over time. Of particular interest were the 26% of pts who converted from low/intermediate into high-risk SMM according to the 20/2/0.008 model. These showed a median TTP similar to pts who were high-risk at baseline (23 and 26 months). In 206 pts with ≥3 CTC assessments (1,055 in total), CoNNector identified two subgroups with undetectable/low vs high CTC levels over time. The latter pts showed a 4.8-fold increased risk of progression (p=.008). Besides CTCs, only serial hemoglobin levels were associated with transformation into active MM (HR: 12.5, p<.001). Serial levels of the M-component, sFLC ratio, β2m and creatinine were not significantly associated with risk of progression. Other features such as BM tumor burden, cytogenetics and imaging were rarely available in ≥3 timepoints and could not be analyzed over time. A dynamic model using serial CTC and hemoglobin levels stratified pts having 0, 1 and 2 risk factors with progression rates of 0%, 5.5% and 54% at 2y (p<.001). Longitudinal characterization of CTC exomes in a total of 40 samples from 14 pts enabled genomic profiling with unprecedented periodicity. For example, a patient with stable SMM and CTC exomes in six consecutive samples showed relatively stable mutational and copy number burden, while the cancer cell fraction of mutated NRAS and +1q21 changed over time. Surprisingly, SMM pts who progressed to active MM did not show a notable accumulation of mutations and copy number alterations in longitudinal CTC exomes. Conclusions: iMMunocell is one of the largest, comprehensive and dynamic studies performed in SMM. Our results support the complementary evaluation of the 20/2/20 and 20/2/0.008 risk models at baseline, and the longitudinal evaluation of the minimally-invasive 20/2/0.008 model for improved prediction of outcomes. Of note, the combined assessment of serial CTC and hemoglobin levels resulted in the most effective dynamic model and enabled the identification of SMM pts with virtually no risk of progression at 2y. Furthermore, CTCs can be used for periodic evaluation of genetic alterations towards a better understanding of its impact in SMM evolution.