Validation of an Adapted NGS‐Based Clonality Assay for MRD Monitoring in B‐ALL Using Archived and Follow‐Up Samples Authors: Moushumi Suryavanshi, Prashant Mehta, Nitin Jain, Dinesh Bhurani, Rakesh Reddy Boya, Pravas Chandra Mishra, Dushyant Kumar, Surender Dhandha, Manoj Kumar, Nupur Das, Sweta Mishra Published: 2025 ABSTRACT ABSTRACTBackgroundMeasurable residual disease (MRD) monitoring is a critical component of risk stratification and therapeutic decision‐making in B‐cell acute lymphoblastic leukemia (B‐ALL). Multiparameter flow cytometry (MFC) is widely used, but its sensitivity is limited to 10−4–10−5. Next‐generation sequencing (NGS) platforms offer greater sensitivity (up to 10−6) but are often limited by cost, infrastructure requirements, and reliance on FFPE tissue. We evaluated the performance of a commercially available NGS‐based assay, the Oncomine B‐cell Clonality panel, adapted at our center for MRD monitoring using diverse sample types.MethodsIn this retrospective study, 32 B‐ALL patients were analyzed. Diagnostic clonotypes were identified from archival stained bone marrow aspirate smears in all cases, and FFPE bone marrow tissue in five. Longitudinal MRD monitoring was performed on bone marrow aspirates collected in EDTA. Serial dilution experiments (10−2–10−6) were conducted to assess sensitivity; reproducibility was tested in triplicates across runs. MRD results from NGS were compared to MFC, and statistical agreement was evaluated using Spearman's correlation and Cohen's kappa.ResultsBaseline clonotypes were detected in 29/32 patients (90.6%). The assay demonstrated a limit of detection of 10−6 and high reproducibility (R 2 = 0.991). NGS and MFC showed 86.5% concordance; NGS detected MRD earlier than MFC in eight cases (44–124 days). Sensitivity and specificity of NGS relative to MFC were 100% and 82.98%, respectively. Clonal detection succeeded in both FFPE and stained slides.ConclusionThis adapted NGS assay offers a sensitive, reproducible, and sample‐flexible MRD monitoring tool for B‐ALL, with potential for broader clinical application. Trial Registration: The authors have confirmed clinical trial registration is not needed for this submission. Introduction The detection of measurable residual disease (MRD) has emerged as a crucial component in the clinical management of hematologic malignancies, particularly B‐cell neoplasms such as acute lymphoblastic leukemia (ALL), while it is assuming increasing importance now in chronic lymphocytic leukemia (CLL), and multiple myeloma (MM) [1, 2]. MRD refers to the persistent malignant clone/s below the threshold of detection of conventional morphological assessment, yet these residual cells frequently serve as the source of disease recurrence. The ability to detect and quantify MRD has fundamentally changed how clinicians assess treatment response, predict relapse, and formulate therapeutic strategies for patients with hematological malignancies [3, 4, 5]. Its accurate measurement has transformed modern clinical practice by enabling precise treatment response evaluation, relapse prediction, and therapy adaptation. Early MRD detection is clinically actionable because it can trigger preemptive interventions, such as allogeneic stem cell transplantation, treatment intensification, or the administration of targeted therapies like blinatumomab. This allows clinicians to intervene before clinical relapse occurs, improving long‐term survival outcomes [6]. Recent data suggest that MRD status is one of the most powerful prognostic markers in ALL: patients who achieve MRD negativity after induction therapy have a 5‐year event‐free survival rate exceeding 90%, compared to less than 50% in those with persistent MRD [7]. MRD‐guided risk stratification now informs frontline therapy decisions across multiple treatment protocols. Traditionally, multiparameter flow cytometry (MFC) has been the predominant method for MRD detection due to its widespread availability, relatively rapid turnaround time, and established clinical utility [8, 9]. MFC relies on the identification of aberrant immunophenotypic patterns to distinguish malignant from normal cells. However, this approach is constrained by sensitivity limitations (typically 10−4–10−5), vulnerability to immunophenotypic shifts during treatment, and variability in operator expertise, standardization, and instrument calibration [10, 11]. In contrast, next‐generation sequencing (NGS) technologies has revolutionized the detection of MRD by targeting unique immunoglobulin (Ig) and T‐cell receptor (TCR) gene rearrangements, which serve as molecular signatures of lymphoid malignancies [12, 13]. Various NGS‐based assays, including the ClonoSEQ Assay (Adaptive Biotechnologies) and assays developed by Invivoscribe, demonstrate superior sensitivity (up to 10−6) and clinical validity across multiple hematologic malignancies [14]. These techniques have established NGS‐based MRD assessment as a new standard, with findings showing that achieving MRD negativity through such sensitive methods. One pediatric study found that NGS identified MRD in 91% of patients, capable of detecting disease at levels as low as 0.0001%, which correlated with relapse risk more effectively than flow cytometry. Notably, MRD positivity at various treatment milestones significantly correlated with worse leukemia‐free survival [15]. Another study indicated that NGS‐based MRD detection could predict relapse in B‐ALL up to 25 weeks before clinical signs, outperforming multi‐color flow cytometry. This early detection facilitates timely preventive treatment strategies [16]. In patients undergoing CAR‐T cell therapy (tisagenlecleucel), detectable MRD by NGS after three months was a strong predictor of relapse, even before B‐cell reappearance. This suggests that MRD tracking could inform the timing of further interventions, such as a second CAR‐T infusion, stem cell transplantation or maintenance [17]. A 2023 study explored blinatumomab's effectiveness in eliminating ultra‐low MRD (as low as 10−6) after chemotherapy, with NGS able to detect residual disease that flow cytometry missed, and 68% of patients becoming MRD‐negative after blinatumomab [18]. Beyond its application in MRD detection, NGS is reshaping diagnosis, risk stratification, and personalized treatment approaches in ALL, offering insights into genetic subtypes and helping to guide targeted therapies [19]. Despite these advances, the wider implementation of NGS‐based MRD testing faces challenges including cost, accessibility, and the need for additional validated platforms [20]. The development of alternative NGS‐based assays that can offer comparable clinical utility while potentially addressing these limitations represents an important direction for the field. We have validated a B‐cell clonality NGS panel, which offers a promising NGS‐based platform for lymphoid clone detection targeting the immunoglobulin heavy chain (IGH) and light chain (IGK, IGL) rearrangements to achieve sensitive and specific MRD detection. In this study, we present a comprehensive validation of an MRD testing workflow utilizing the NGS based B cell clonality assay. Our validation strategy encompasses rigorous analytical assessment through serial dilution experiments to determine sensitivity limits, repeatability testing across multiple independent runs to establish reproducibility, and direct comparison with flow cytometry results to contextualize performance against a widely used clinical method. By subjecting this platform to these stringent validation criteria, we aim to establish its analytical validity, clinical feasibility, and potential utility as an alternative as well as complement for MRD assessment in B cell lymphoid malignancies. Materials and Methods Patient Cohort This retrospective study included 32 patients diagnosed with B‐cell acute lymphoblastic leukemia (B‐ALL), with a median age of 27 years (range: 6–90). The cohort comprised 23 males and nine females. Patients were selected consecutively, based on availability of diagnostic and follow‐up samples suitable for MRD analysis. Sample Collection and Processing For baseline MRD clonality assessment, stained, archived stained bone marrow aspirate smears were used in all 32 patients to identify the diagnostic clonal sequence. In addition, formalin‐fixed paraffin‐embedded (FFPE) bone marrow biopsy blocks were concurrently processed in five of these cases to compare clonal detection across specimen types. For all subsequent MRD time points, bone marrow aspirate samples collected in EDTA tubes were used. For NGS analysis, genomic DNA was extracted using the QIAamp DNA Blood Mini Kit (Qiagen, Hilden, Germany) for bone marrow samples and the QIAamp DNA FFPE Tissue Kit for FFPE samples, following the manufacturer's instructions. NGS‐Based MRD Detection MRD detection was performed using the Oncomine B‐cell Clonality Assay (Thermo Fisher Scientific), which targets the IGH and light chain (IGK, IGL) gene rearrangements. Library preparation was conducted according to the manufacturer's protocol. Briefly, 10–100 ng of genomic DNA was used for multiplex PCR amplification of the rearranged Ig loci. The resulting amplicons were purified, quantified, and pooled for sequencing on the Ion GeneStudio S5 System (Thermo Fisher Scientific). Analysis of the Complementarity‐Determining Region 3 (CDR3) of each IGH/IGK/IGL rearrangement showed that the CDR3 sequence for each patient was distinct. The criteria for determining clonality were outlined as follows: A clone was recognized by having at least five identical sequences obtained through sequencing, and the frequency of that clone had to be greater than 5% to be considered a marker for monitoring minimal residual disease (MRD). Sequences identified in bone marrow samples during follow up assessments were compared to the clonal sequences from their baseline samples, which were designated as index clonal sequences. Samples were deemed MRD‐positive if the same sequences were found in the follow up marrow samples [21, 22]. To assess the feasibility of NGS for monitoring MRD in patients with acute lymphoblastic leukemia, we examined the sensitivity and reproducibility of NGS in detecting MRD. Serial Dilution Experiments To validate assay sensitivity and determine the limit of detection (LOD), serial dilution experiments were conducted using a known positive diagnostic DNA sample. This sample contained a dominant clone with an estimated 80% clonal frequency. It was serially diluted with DNA from a healthy donor to simulate decreasing tumor burdens from 102−2 to 10−6. For each dilution point, 400 ng of total DNA was used for library preparation to ensure consistency across samples. Library quality and yield were verified using fluorometric quantification. A fixed input of 50 pM per library was used for sequencing. Sequencing depth was scaled according to the dilution factor to maintain confidence in rare clone detection. Specifically, 2 million reads for undiluted sample and up to 30 million reads for 10−6 dilution. This proportional sequencing strategy ensured sufficient coverage at each dilution point and allowed a robust assessment of assay sensitivity. Bioinformatics analysis included read alignment, clonotype assembly, and quantitative tracking of the diagnostic clone. The LOD was defined as the lowest dilution at which the index clonotype was consistently and reproducibly detected in all replicates (Figure 1a). (a) shows a linear relationship between sample dilution ratio and clonal frequency (both on log10 scale). The strong linear fit (R 2 = 0.9910) demonstrates excellent assay linearity across the tested dilution range. Error bars likely represent standard deviation across replicates. (b) Displays consistency in clonal frequency detection across three replicates (R1, R2, R3), comparing inter‐run (Blue X) and intra‐run (red circle) measurements. The consistent values of ∼1.0 across all replicates indicate high assay reproducibility. Repeatability Testing To evaluate the reproducibility of the assay, selected samples were analyzed in triplicate across three independent runs. This approach allowed assessment of both intra‐run and inter‐run variability in clonal detection and quantification (Figure 1b). Flow Cytometry MRD Detection Bone marrow aspirates collected in EDTA were processed for flow cytometric immunophenotyping using stain‐lyse‐wash technique as per the standard operating procedure followed in our laboratory. For MRD evaluation, first‐pull bone marrow aspirate sample was used to avoid hemodilution and processed using bulk‐lysis protocol as per Euro flow consortium guidelines. The samples were processed within 4–6 h of collection to ensure optimum viability of cells. Single tube‐13‐color panel was used for immunophenotyping of MRD samples. Data acquisition was done on a flow cytometer with three laser set‐up (Dx‐flex, Beckman Coulter, Brea, CA, USA) and analyzed using Kaluza software (v2.0, Beckman Coulter). We acquired 3–5 million total events per tube. Based on a 20‐event LOD and 50‐event lower limit of quantification (LLOQ), this corresponds to an LOD of ∼0.0004%–0.0007% and an LLOQ of ∼0.001% MRD among nucleated cells. Data Analysis For NGS data analysis, the Ion Reporter Software (Thermo Fisher Scientific) was used to identify and quantify clonal populations. A clonotype was considered significant if it constituted > 5% of all reads for a given immune receptor gene. For MRD assessment, the dominant clonotype identified at diagnosis was tracked in follow‐up samples. MRD positivity by NGS was defined as the detection of the diagnostic clonotype at a level ≥ 0.0001% of all reads for the corresponding immune receptor gene. Statistical analyses were performed using R software (version 4.5.0). We evaluated the correlation between MFC‐MRD and NGS‐MRD results with Spearman's correlation coefficient. Statistical analyses were conducted using GraphPad Prism 9.5.1 (GraphPad Software Inc., San Diego, CA, USA) and R Software version 4.5.0. Results Patient Demographics and Diagnosis Baseline demographic and clinical characteristics of the 32 patients are summarized in Table 1. TABLE 1Summary of patient demographics and treatment modalities. HSCT ‐allogeneic hematopoietic stem cell transplant, CAR T‐ chimeric antigen receptor T‐cell therapy. Variable Value Total patients32Median age (range)27 years (6–90)Sex (Male/Female)23/9 Treatment modality Chemotherapy8 patientsHSCT18 patientsCAR T‐cell therapy6 patients Summary of patient demographics and treatment modalities. HSCT ‐allogeneic hematopoietic stem cell transplant, CAR T‐ chimeric antigen receptor T‐cell therapy. Comparison of Clone Detection Across Different Sample Types To evaluate the consistency of clonal detection across varying specimen types, we compared results from DNA extracted from two commonly available archival sources: aged, stained bone marrow aspirate smears and FFPE bone marrow biopsy tissue. This comparison was conducted in five patients diagnosed with B‐cell ALL, three of them exhibiting a dominant lymphoid clone. In two cases no dominant clone was detected by both sample types. In other three cases, the same dominant clone was successfully identified in both sample types using the NGS‐based detection method. Despite the inherent differences in sample preservation and potential degradation (especially in stained smears), clonal concordance was maintained. Comparison With Flow Cytometry Among 32 patients, baseline NGS clonotypes were successfully identified in 29 patients (90.6%), and MRD tracking was performed at 56 total follow‐up time points. MRD detection by NGS and MFC was concordant at 48 of 56 time points (86.5%). NGS identified MRD in eight samples where MFC was negative. No cases were found where MFC detected MRD while NGS did not. “Discordant” results refer to cases where one method detects MRD and the other does not. In this study, all discordant results involved NGS detecting MRD that was missed by MFC, indicating the higher sensitivity of the NGS assay. The Cohen's kappa coefficient for agreement between the two methods was 0.706 (95% CI: 0.61–0.80), indicating substantial agreement. The sensitivity and specificity of NGS relative to flow cytometry were 100% and 82.98%, respectively (Figure 2). Longitudinal monitoring of minimal residual disease (MRD) in 32 B‐ALL patients using NGS and flow cytometry (FCM). Each horizontal line represents a patient, annotated with age, gender, and treatment modality (HSCT, chemotherapy, or CAR‐T). MRD status over time is shown using next‐generation sequencing (NGS) and flow cytometry (FCM) results: NGS+ (solid red triangle), NGS– (open red triangle), FC+ (solid green circle), and FC– (open green circle). Blue open triangles indicate cases where no baseline clone was detected. Clinical outcomes are marked as alive (solid blue diamond) or deceased (Blue X). Notable genetic alterations, such as TP53, BCR::ABL1 P210, P2RY8::CD99P1 fusion, Ph+, and del(17p)+ complex karyotype, are annotated. This figure illustrates the differential sensitivity and timing of MRD detection by NGS and FCM, along with patient‐specific clinical trajectories. Time to MRD Detection and Clinical Correlation In all eight discordant cases, MRD was later confirmed by MFC at subsequent time points, with detection lag times ranging from 44 to 124 days, further supporting the superior sensitivity of the NGS method. Among the cohort of 32 patients, four patients died during the study period. Three deaths were disease‐related (Cases 1, 10, and 23), all of whom were MRD‐positive by both NGS and MFC. One death was therapy‐related (Case 7, post‐CAR T‐cell therapy), in a patient who remained MRD‐negative by both methods (Figure 2). Case 28, a 62‐year‐old female with Philadelphia‐like ALL post‐CAR T therapy, NGS identified MRD a full month before MFC. Conversely, Case 29, a 19‐year‐old male with p53‐mutated ALL, maintained MRD negativity by both NGS and MFC but succumbed to CAR T‐related neurotoxicity. A further nuance is demonstrated in Case 26, a Ph‐positive ALL patient, where real‐time PCR detected weak BCR::ABL1 transcript levels while NGS MRD was negative. This patient ultimately achieved remission with concordant negativity across both platforms. Discussion This study validates the use of Oncomine B‐cell clonality NGS panel as a sensitive and reproducible method for MRD detection in B‐ALL. Our findings align with a growing body of evidence supporting NGS as a powerful tool in the clinical management of hematologic malignancies, with implications for both relapse prediction and treatment guidance. The analytical performance of the NGS assay demonstrated excellent sensitivity detecting MRD down to 10−6 and high reproducibility across technical replicates. These performance characteristics compare favorably with established NGS platforms, as seen in prior studies [22, 23]. Importantly, our NGS panel offers practical advantages over previously used NGS‐MRD assays. While those platforms are limited to extracting baseline clones only from archival FFPE tissue, our panel successfully detects baseline clones from both FFPE blocks and archival stained bone marrow slides. This is a meaningful innovation, particularly in cases where FFPE blocks may have undergone decalcification or suboptimal fixation, compromising DNA quality for sequencing [24]. In contrast, we demonstrated in five patient cases that dominant lymphoid clones could be identified with equal fidelity from both FFPE and stained slide material, making our panel more versatile in real‐world sample availability scenarios. In three patients, baseline clone detection was unsuccessful, preventing longitudinal MRD tracking. This limitation is shared across all IGH/IGK/IGL‐targeted NGS MRD platforms [15, 25]. Such failures typically stem from somatic hypermutation in Ig heavy and light chain loci, which hinders primer binding during PCR amplification. These mutations are known to introduce mismatches that reduce the efficiency or specificity of clonal detection. Beyond technical aspects, our comparative analysis with MFC highlighted a concordance of 86.5%. Importantly, in eight timepoints, MRD was detectable by NGS earlier than by MFC by as much as 124 days. This is in line with emerging evidence that NGS can serve as a more sensitive early warning tool. Case 28, a 62‐year‐old female with Philadelphia‐like ALL post‐CAR T therapy, exemplified this advantage, with NGS identifying MRD a full month before MFC. Such early detection has clear clinical value, potentially prompting timely therapeutic escalation [23, 26]. Conversely, Case 29, a 19‐year‐old male with p53‐mutated ALL, maintained MRD negativity by both NGS and MFC but succumbed to CAR T‐related neurotoxicity. This case illustrates that MRD status, while highly prognostic, is not wholly predictive, and clinical monitoring especially post‐CAR T must include vigilance for treatment‐related adverse effects [27]. A further nuance is demonstrated in Case 26, a Ph‐positive ALL patient, where real‐time PCR detected weak BCR::ABL1 transcript levels while NGS MRD was negative. Ultimately, the patient achieved remission with concordant negativity across both platforms. This underlines the importance of harmonizing molecular and sequencing approaches in complex genotypic settings [28, 29]. Taken together, these findings affirm the clinical feasibility of our NGS panel and highlight its potential advantages in accessibility, flexibility, and sensitivity. Despite these promising results, several limitations must be acknowledged. First, the sample size is relatively small (n = 32) and may limit the generalizability of findings across all B‐ALL subtypes and treatment contexts. Second, we did not include standardized cell lines for assay calibration, relying instead on patient‐derived samples for dilution experiments. While informative, this limits the ability to benchmark performance against known clonal architecture and allele burden. Third, potential PCR amplification bias including preferential primer binding or dropout due to somatic hypermutation could have influenced the representation of clonal sequences, especially at lower tumor burdens. Future studies should address these limitations through prospective validation in larger, multicenter cohorts. Another essential direction is the inclusion of longitudinal survival analysis (e.g., event‐free and overall survival), which would clarify the prognostic value of ultra‐sensitive MRD detection across treatment phases. This is particularly relevant as therapies such as CAR T cells and blinatumomab become more widely used in MRD‐guided treatment decisions. In summary, this study contributes to the growing evidence that NGS‐based MRD detection is not only feasible but also potentially superior in sensitivity and flexibility compared to traditional flow cytometry. Its compatibility with various sample types and its ability to detect relapse earlier underscore its value in personalized care in B‐ALL. The Oncomine B‐cell clonality NGS panel provides a robust, flexible, and sensitive alternative for MRD monitoring in B‐ALL. Its unique compatibility with both FFPE and stained bone marrow slides increases its practical utility, especially when baseline samples are limited or suboptimal. Though challenges remain, including primer‐binding failures in highly mutated clones and the need for broader validation, this assay represents a meaningful step toward more accessible and accurate MRD surveillance. In the future, combining NGS with flow cytometry, PCR, and targeted therapies will likely offer the most comprehensive view of disease burden, empowering clinicians to tailor interventions and improve outcomes in hematologic malignancies. Author Contributions Moushumi Suryavanshi: conceptualization, methodology, formal analysis, visualization, writing – original draft, supervision, project administration. Prashant Mehta: conceptualization, resources/clinical oversight, supervision. Nitin Jain: resources/clinical oversight. Dinesh Bhurani: resources/clinical oversight. Rakesh Reddy Boya: resources/clinical oversight. Pravas Chandra Mishra: resources/clinical oversight. Dushyant Kumar: investigation (laboratory and data acquisition), data curation. Surender Dhandha: methodology, investigation (laboratory and data acquisition), data curation. Manoj Kumar: investigation (laboratory and data acquisition), data curation, formal analysis, visualization. Nupur Das: methodology. Sweta Mishra: investigation (laboratory and data acquisition). Funding This research received no specific grant from any funding agency in the public, commercial, or not‐for‐profit sectors. Ethics Statement The study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Ethics Committee, Amrita Institute of Medical Sciences (AIMS), Faridabad, India (Approval No. AIMS‐IEC‐BAS‐11‐25‐002). Conflicts of Interest The authors declare no conflicts of interest. Suryavanshi M., Mehta P., and Jain N., et al. ““Validation of an Adapted NGS‐Based Clonality Assay for MRD Monitoring in B‐ALL Using Archived and Follow‐Up Samples.”” eJHaem (6, no. 6 (2025): e70202. 10.1002/jha2.70202 Data Availability Statement De‐identified clinical datasets, flow cytometry summaries, and NGS MRD outputs underlying this study are available from the corresponding author on reasonable request, subject to Institutional Ethics Committee approval and data‐sharing agreements. References Associated Data Data Availability Statement De‐identified clinical datasets, flow cytometry summaries, and NGS MRD outputs underlying this study are available from the corresponding author on reasonable request, subject to Institutional Ethics Committee approval and data‐sharing agreements.