A total of 63 patients with cardiac AL were included in this study. There were 21 (33.3%) patients who achieved early MRD negativity, and 42 (66.7%) patients who did not. The median time from treatment initiation to early MRD test was 2.27 (inter-quartile range, IQR 1.97–3.98) months in the entire cohort, and was 2.86 (IQR 2.27–4.59) months in early MRD negative and 2.27 (IQR 1.84–2.42) months in early MRD positive group, respectively.

Table 1 depicts the baseline characteristics of the entire cohort and the comparison between early MRD negative and early MRD positive patients. In the entire cohort, as for the features relating underlying plasma cell dyscrasia, involved light chain was lambda in 81% (n = 51) of patients; the median difference in involved and uninvolved light chains (dFLC) was 116.3 mg/L (IQR 56.0–252.1) and the median bone marrow plasma cell percentage was 10% (IQR 6–15%); the most commonly found cytogenetic abnormality was t(11;14), accounting for 45.8% (27/59). As for the features relating cardiac involvement, no patient was Mayo 2004 stage I because of the inclusion criteria (NT-proBNP > 650 pg/ml) and the majority of patients were classified as Mayo 2004 stage III, with end-stage (Mayo 2004 stage IIIb) cardiac involvement accounting for 12.7%. First-line treatment was proteasome inhibitor-based therapy for all patients, while two patients in the early MRD positive group received first-line daratumumab, both harboring t(11;14) at baseline. No patient in the cohort had undergone autologous stem cell transplantation (ASCT) at the last follow up.

Patients who achieved early MRD negativity were less likely to be male (57.1% vs 81.0, P = 0.045), and less likely to harbor t(11;14) at baseline (21.1% vs 57.5%, P = 0.009). There were no significant differences in other baseline characteristics of patients who were early MRD negative or positive, such as age, cardiac staging and biomarkers, renal or hepatic involvement. Of note, a trend toward numerically lower dFLC and bone marrow plasma cell percentage could be observed for early MRD negative group.

HemCR/VGPR/PR/NR was achieved in 33.3% (n = 21), 38.1% (n = 24), 14.3% (n = 9), 14.3% (n = 9) of patients in the entire cohort at the time of MRD assessment, with 71.4% (45/63) of patients achieving ≥ HemVGPR. In the 21 patients with early MRD negativity, HemCR/VGPR was achieved in 14/7 patients respectively, i.e., all 21 patients in early MRD negative group achieved ≥ HemVGPR. In the 42 patients with early MRD positivity, Hem CR/VGPR/PR/NR was achieved in 7/17/9/9 patients respectively. The MRD negative rates amongst patients in HemCR were 66.7% (14/21) and amongst patients in HemVGPR were 29.2% (7/24). The MRD negative rates amongst patients in HemVGPR or better were 46.7% (21/45). The hematologic response distribution in early MRD negative and positive group is shown as in Table 2.

Throughout first-line therapy, CarCR/VGPR/PR/NR was achieved in 1.6% (n = 1), 19.0% (n = 12), 27.0% (n = 17), 52.4% (n = 33) of patients in the entire cohort, with 47.6% (30/63) of patients achieving ≥ CarPR and 20.6% (13/63) of patients achieving ≥ CarVGPR. Within the first four cycles, CarCR/VGPR/PR/NR was achieved in 1.6% (n = 1), 11.1% (n = 7), 27.0% (n = 17), 60.3% (n = 38) of patients in the entire cohort, with 39.7% (25/63) of patients achieving ≥ CarPR and 12.7% (8/63) of patients achieving ≥ CarVGPR. The only patient who achieved best response CarCR was a patient with Mayo 2004 IIIa cardiac stage at baseline (cTnT 0.054 ng/ml, NT-proBNP 771.7 pg/ml) and no cytogenetic abnormality, achieved best hematologic response HemVGPR and MRD test positive after two cycles of CyBorD, and achieved CarCR after 4 cycles of CyBorD (NT-proBNP decreased to 256.5 pg/ml).

Early MRD negativity was associated with a higher likelihood of cardiac response ≥ CarPR (> 30% reduction in NT-proBNP from baseline, which is also the attainment of cardiac response using the binary category of response) throughout first-line therapy [66.7% (14/21) vs 38.1% (16/42), P = 0.032] and within four cycles [61.9% (13/21) vs 28.6% (12/42), P = 0.011] (Fig. 1). The trend was consistent in patients with ≥ HemPR/ ≥ VGPR/ ≥ CR, with a numerical higher but statistically nonsignificant proportion favoring early MRD negative group (Supplementary Table S1). Recording the date and time of attaining cardiac response ≥ CarPR in first-line therapy (patients who did not were censored at treatment switching or last follow up), the cumulative incidence curve of cardiac response ≥ CarPR is shown as in Fig. 2, with a significant difference between early MRD negative and positive group (P = 0.034). Among those who attained ≥ CarPR in first-line therapy, the median time to CarPR was 1.8 months in early MRD negative group and 3.9 months in early MRD positive group. The cumulative incidence curve in patients with ≥ HemPR/ ≥ VGPR/ ≥ CR is shown in Supplementary Figure S1, with a consistent trend favoring early MRD negative group.

Proportion of cardiac response ≥ CarPR throughout first-line therapy and within four cycles

Cumulative incidence curve of cardiac response ≥ CarPR (x-axis maximum set to 36 months)

Early MRD negativity was also associated with a higher likelihood of cardiac response ≥ CarVGPR (> 60% reduction in NT-proBNP from baseline) throughout first-line therapy [38.1% (8/21) vs 11.9% (5/42), P = 0.023] and a numerically higher proportion within four cycles [23.8% (5/21) vs 7.1% (3/42), P = 0.104] (Fig. 3). The trend was again consistent in patients with ≥ HemPR/ ≥ VGPR/ ≥ CR, with a numerical higher but statistically nonsignificant proportion favoring early MRD negative group (Supplementary Table S2). Recording the date and time of attaining cardiac response ≥ CarVGPR in first-line therapy (patients who did not were censored at treatment switching or last follow up), the cumulative incidence curve of cardiac response ≥ CarVGPR is shown as in Fig. 4, with a significant difference between early MRD negative and positive group (P = 0.026). Among those who attained ≥ CarVGPR in first-line therapy, the median time to CarVGPR was 3.4 months in early MRD negative group and 6.5 months in early MRD positive group. The cumulative incidence curve in patients with ≥ HemPR/ ≥ VGPR/ ≥ CR is shown in Supplementary Figure S2, with a consistent trend favoring early MRD negative group.

Proportion of cardiac response ≥ CarVGPR throughout first-line therapy and within four cycles

Cumulative incidence curve of cardiac response ≥ CarVGPR (x-axis maximum set to 36 months)

The proportion of certain depth of cardiac response throughout first-line therapy was investigated in certain clinically relevant subgroups. Both the proportion of patients achieving ≥ CarPR and ≥ CarVGPR tended to be numerically higher for early MRD negative group in patients with or without t(11;14), with statistical significance reached for proportion of ≥ CarPR in t(11;14) positive subgroup (100.0% vs 39.1%, P = 0.041). Of note, even if patients were harboring t(11;14), when they attained early MRD negative, they were highly likely to achieve cardiac response (100% in our cohort). Similar trends could be observed in patients with Mayo 2004 IIIb stage or not and BMPCs ≥ 10% or not, both favoring early MRD negative group, with statistical significance reached in certain subgroups such as not Mayo 2004 IIIb (66.7% vs 37.8% for ≥ CarPR, P = 0.044) and BMPCs ≥ 10% (87.5% vs 40.0 for ≥ CarPR, P = 0.039; 50.0% vs 8.0% for ≥ CarVGPR, P = 0.020) (Fig. 5). The specific proportion of ≥ CarPR and ≥ CarVGPR for early MRD negative and positive group and P value in clinically relevant subgroups were listed in Supplementary Table S3.

Proportion of cardiac response ≥ CarPR (A,C,E) and ≥ CarVGPR (B,D,F) throughout first-line therapy in clinically relevant subgroups: t(11;14)- vs t(11;14) + (A,B), Not Mayo 2004 IIIb vs Mayo 2004 IIIb (C,D), BMPCs < 10% vs ≥ 10% (E,F)

The median follow-up time from treatment initiation was 27.2 (IQR 9.2–50.3) months. The median EFS was significantly longer in early MRD negative group compared to early MRD positive group (not reached vs 19.9 months, P = 0.031), as well as median PFS (not reached vs 31.3 months, P = 0.033) (Fig. 6). The median OS was both not reached in either group, with no statistical difference (P = 0.786).

Event-free survival (EFS) (A) and progression-free survival (PFS) (B) from treatment initiation