Assessment of MRD status throughout treatment, as well as during the post-treatment phase, has well-established prognostic significance in paediatric B-ALL [4,5,6]. In this study, we conducted a comprehensive examination of MRD data acquired using the LymphoTrack assay, in conjunction with pertinent clinical parameters, treatment response profiles, and MRD results from other assays (MFC, RT-PCR, and FISH) in real-world clinical practice.

Use of NGS to track clonal rearrangements of immunoglobulin and T-cell receptor genes is widely accepted as a valuable tool for MRD assessment [7, 11, 13]. NGS methodology has several notable advantages, including high sensitivity, broad applicability across diverse patient populations, elimination of the need for patient-specific oligonucleotide primers, and the ability to detect both oligoclonality and clonal evolution dynamics over time. It also offers the practical benefit of obviating the need for freshly acquired patient samples, in contrast to conventional methods such as flow cytometry [11,12,13,14,15,16]. In addition, NGS MRD has shown improvement in measurement precision, and prediction of relapse and survival outcomes of post-treatment MRD in pediatric B-ALL compared to MFC and PCR [11, 18, 19]. NGS MRD also enables the identification of new emerging clones during monitoring of relapsed ALL patients [20].

Nevertheless, there are several limitations to using the NGS clonality assay in patients with B-ALL. One important constraint is the potential for false positives and false negatives because of the assay’s exquisite sensitivity, which may detect minor clonal populations of uncertain clinical significance or miss rare variants. Furthermore, the phenomenon of pseudoclonality presents a challenge, as the emergence of novel clonal populations during treatment may introduce ambiguity regarding their clinical significance [21, 22]. This may be particularly problematic in samples obtained after intensive chemotherapy regimens when lymphocyte counts are diminished. The intricate nature of data analysis and interpretation poses further obstacles, demanding specialised expertise and resources for accurate assessment [22]. Moreover, full clonal heterogeneity may be difficult to detect in highly diversified leukaemia populations. The assay’s prolonged turnaround time and relatively high cost also warrant optimisation. Thus, while NGS holds substantial promise in MRD assessment, its limitations underscore the need to examine the correlation between NGS and other MRD assays in clinical studies and to refine its utility in clinical practice [7, 10, 11, 13, 23, 24].

The average turnaround time of LymphoTrack-MRD analysis included in this study was 11.1 days (range, 4 to 24 days). Despite the longer duration required for LymphoTrack-MRD analysis, its clinical implementation is not deemed problematic. This is because each assessment period for treatment response spans 4 to 8 weeks, and the early phase of intensification treatments exhibits uniformity across risk groups. Nonetheless, efforts should be directed towards reducing the longer turnaround time for early MRD measurement for SR patients in subsequent research.

An index clonal sequence was identified in 97.8% of our study cohort. MFC was used to assess disease status in all patients, and the concordance rate between MFC and LymphoTrack was 74.8%. The presence of positive clonality on high-throughput sequencing and negative results on MFC was previously found to be associated with worse outcomes than negative results on both tests [7, 11]. Our study did not statistically confirm these findings, however, when discordancy occurred, NGS+/MFC- was most common, and relapses occurred exclusively in NGS+/MFC- patients. Discordance can occur when MFC remission is achieved without LymphoTrack (possibly due to false-positive LymphoTrack results) or as an early sign of relapse. Svaton et al. reported that discordant cases between NGS MRD and PCR MRD for immunoglobulin/T cell receptor markers were enriched in low-specificity clonotypes with short junction lengths [18]. In our cohort, there was no significant difference in junction length between patients who experienced relapse and NGS+/MFC- patients who did not experience relapse (independent two-sample t-test, P = 0.4905 for IGH and P = 0.9757 for IGK).

Of the 58 patients lacking target markers for FISH or RT-PCR, 54 had clonality sequences allowing MRD monitoring. Concordance between RT-PCR and LymphoTrack was 70.7%. Furthermore, patients with sustained negative NGS clonality assay results had a favourable prognosis, highlighting its superior sensitivity and clinical utility as an assessment tool for assessing MRD, compared to MFC or RT-PCR [11, 18].

We also found that increased clone quantity or conversion to MRD positivity for IGH or IGK clonality was strongly associated with worse RFS, compared to decreased clone quantity or sustained negativity. These findings highlight the importance of vigilant monitoring of changes in NGS clonality for evaluating disease progression and therapeutic responses. They also suggest the need for more intensified treatment strategies to improve outcomes in patients manifesting an escalation in MRD status.

The 4 patients with bone marrow relapse exhibited clear evidence of elevated MRD levels of clone quantity or positive conversion on NGS clonality assay before morphologic relapse. Although pseudoclonal expansion is possible when lymphoid cell counts are low [25], the possibility of relapse must be seriously considered when new clone is detected. Notably, 2 patients displayed the emergence of new clonal sequences before morphologic relapse, and 1 patient exhibited a new BCR::ABL1 mutation. Furthermore, in 1 patient initially presenting with 3 IGH clones, the least prevalent clone (16.64%) at diagnosis became the exclusive, dominant clone (90.28%) at relapse. These findings are suggestive of blast transformation [26] or expansion of a minor population of blasts during relapse [27]. Our results highlight the robust utility of the NGS clonality assay, which allows for early detection and monitoring of multiple productive gene rearrangements in ALL, especially in the context of MRD assessment.

Among relapsed patients, 2 patients (P35 and P50) showed positivity for IGKV3D-20-Kdel clonotypes. The manufacturer provided information on the low-specificity IGK clonotypes (Intron-Kdel, V3D-20 with any J or Kdel, and V3-11 with any J or Kdel) that may not be suitable for MRD analysis. However, in our patients, IGKV3D-20-Kdel clonotypes decreased or disappeared in CR and increased upon relapse, similar to other index clonotypes. Based on these experiences, we included the clonotype in our analysis. On the other hand, 4 patients in this study were positive only for the low-specificity clonotype and negative for MFC at CR (12 samples from 4 patients). These 4 patients did not relapse and maintained CR. Given that there are cases where low-specificity clonotypes are meaningful as an MRD marker and cases with the possibility of false positives, low-specificity clonotypes must be used in combination with other clonotypes from NGS and other MRD techniques such as RT-PCR and MFC.

This study had several limitations. First, we included all samples in the analysis without limiting minimal sequencing read count or input DNA amount; we did not apply the sensitivity threshold. Sequencing read count and input DNA amount determine the sensitivity of the LymphoTrack assay. In our analysis, 36.2% of the samples did not have sufficient sequencing read count and input DNA count to reach 10− 4 sensitivity with 95% confidence. Therefore, we must consider the possibility that the same level of sensitivity was not obtained in all samples, which affected the results of our analysis. However, our findings are meaningful in suggesting that significant MRD results can be obtained even with a level of assay performance that can be performed in real-world practice. Second, this study is its retrospective cohort design, which may have led to overestimated survival outcomes. Additionally, the treatment approach, including close monitoring and early consideration of HSCT for patients with persistent NGS clonality positivity, may have introduced bias. Furthermore, the short follow-up duration and limited number of patients precluded meaningful analysis of the impact of specific genetic mutations on survival outcomes.

In conclusion, this study highlights the potential of NGS clonality as a valuable tool for evaluating MRD in paediatric patients with B-ALL. Our findings demonstrated its ability to assess a larger number of patients than conventional methods and to predict disease deterioration before bone marrow relapse. Prospective studies with larger patient cohorts are required to explore tailored treatment strategies based on sustained MRD status. These efforts will contribute to advancing our understanding and management of paediatric ALL.