Circulating tumor cells (CTCs) combined with frailty index offer an improved stratification system for transplant-ineligible newly diagnosed multiple myeloma patients

Abstract

Abstract Background Higher levels of circulating tumor cells (CTCs) have recently emerged as a negative independent prognostic factor in transplant-eligible newly-diagnosed Multiple Myeloma (NDMM) patients (pts). However, the prognostic impact of CTCs in transplant-ineligible (TI) NDMM pts is not adequately reported, especially in the context of their frailty status. Our aim was to evaluate the clinical relevance of CTCs in TI pts, associate their levels with frailty index, and define a novel CTC-based algorithm to improve current stratification systems. Methods CTC evaluation was performed with Next-Generation Flow cytometry (NGF) in 405 previously untreated TI pts. A minimum of 20 circulating clonal cells was set as a prerequisite for the detection of CTCs; the median limit of detection reached was 2.3x10-6. All pts had available clinical data regarding demographics, cytogenetics, heavy/light chain restriction, bone marrow (BM) infiltration, white blood cell count, platelet count, levels of serum albumin, creatinine, calcium, LDH and β2-microglobulin and were categorized to fit, intermediate-fit and frail based on the IMWG prognostic system. Pts with CTCs≥5% were defined as having pPCL and were excluded from the study. The median follow-up monitoring was 30 months. 66% of the pts were treated with bortezomib-based triplets (VRD, VCD), 20% with Dara-based quadruplets and 14% with heterogeneous regimens. Results CTCs were detected in 371/405 (91.6%) pts with a median value of 0.012% of total nucleated cells (range 0.0002%-4.9%). Pts experienced a progressively worse outcome per CTC-log increase for both PFS and OS; for pts with CTCs at levels ≤10-6, 10-5, 10-4, 10-3 and 10-2, the 2-year PFS rates were 78%, 70%, 62%, 60% and 47%, and the 2-year OS rates were 92%, 84%, 79%, 74% and 66%, respectively. In line with previous observations, the optimal CTC cut-off point was 0.02%, clearly segregating TI pts in two distinct prognostic categories (median PFS: 38 vs. 48 months, HR: 0.71, 95% CI:0.53-0.97, p<0.01; median OS: NR for both categories below and above 0.02% CTCs, HR: 0.61, 95% CI: 0.42-0.80, p<0.01). In terms of frailty, 38.8%, 35.3% and 25.9% of pts were fit, intermediate-fit and frail with significantly distinct clinical outcomes (median PFS: 48, 41 and 31 months, respectively, p<0.01; median OS: NR, 65 and 42 months, respectively, p<0.0001). Of note, the presence of CTCs was independent of the frailty status (median CTCs: 0.014%, 0.01% and 0.016% for fit, intermediate-fit and frail pts, respectively). This lack of association was further verified in multivariate Cox-regression analyses for both PFS and OS, in models including also R-ISS, and treatment type as covariates, where all variables retained their independent prognostic value. The independent prognostic value of frailty index, CTCs and R-ISS allowed us to incorporate all parameters into a unified model to refine patient stratification. Based on our model, pts were assigned one point per increasing risk status in each prognostic category (1, 2, 3 points fit, intermediate-fit and frail pts, respectively; 1, 2 points for CTCs<0.02% and ≥0.02%, respectively; 1, 2, 3 points for R-ISS1, R-ISS2 and R-ISS3, respectively) for a final scoring system between 3-8 points. A total of 7.7%, 18.3% 26.5%, 26.8%, 17.5% and 3.1% of pts had scores 3, 4, 5, 6, 7 and 8, respectively. The model showed its maximum performance for both PFS and OS in a 3-scale mode, where scores 3-4 consisted group A, scores 5-6 group B and scores 7-8 group C. As such, the median PFS was 53, 40 and 29 months for pts in group A, B and C, respectively (p<0.001), and the median OS was NR for pts in groups A and B, and 39 months for pts in group C (p<0.0001). The c-index value of our model was 0.67 for PFS and 0.72 for OS, and proved superior than the relevant performance of R-ISS (0.63 and 0.70, respectively) and/or frailty index alone (0.58 and 0.6, respectively) or any other dual combination thereof. Conclusions The evaluation of CTCs offers a non-invasive examination with a strong prognostic impact for TI NDMM pts, independent from other established biomarkers. The integration of CTCs together with the patients' frailty status and the well-established R-ISS allows for a significant refinement of current predictive and prognostic algorithms, that could be used for a highly accurate therapeutic decision-making in the TI setting.