System and method for measuring and analyzing minimal residual disease in childhood b-precursor acute lymphoblastic leukemia by multiparameter flow cytometry

Abstract

The present invention relates to a system and a method for measuring and analyzing minimal residual disease (MRD) in pediatric B-cell precursor acute lymphoblastic leukemia (B-ALL) using multiparameter flow cytometry (MPFC). The invention finds application in clinical diagnostics and hematology-oncology for quantifying MRD in B-ALL patients with high sensitivity and specificity, needed for risk stratification, monitoring treatment response, and informing therapeutic decisions. The system comprises interconnected subsystems including an acquisition subsystem with an MPFC instrument, a control and file generation subsystem, and an analytical subsystem. The analytical subsystem incorporates modules for sequential data reduction, automated data cleaning, automated unsupervised data clustering, and interactive cluster analysis. Key advantages include high MRD detection sensitivity (e.g., 10⁻⁵ or 0.001%) and high specificity, without reliance on reference samples or supervised machine learning models, making it applicable in laboratories with different measuring equipment and using different panels of antibodies for identification of leukemic cells.

Description

[0001] SYSTEM AND METHOD FOR MEASURING AND ANALYZING MINIMAL RESIDUAL DISEASE IN CHILDHOOD B-PRECURSOR ACUTE LYMPHOBLASTIC LEUKEMIA BY MULTIPARAMETER FLOW CYTOMETRY RELATED APPLICATIONS This application claims priority to B.G. Provisional Patent Application No. BG / P / 2024 / 113931, filed July 19, 2024 the entire disclosure of which is hereby incorporated herein by reference. FIELD The present invention pertains to systems and methods for the measurement and analysis of minimal residual disease (MRD) associated with pediatric B-cell precursor acute lymphoblastic leukemia (BCP-ALL). Provided herein are systems and methods employing multidimensional automated cluster analysis of cellular characteristics, as measured by multiparameter flow cytometry, to identify rare aberrant cell populations within a biological sample. The disclosed system and method are applicable to risk assessment, disease monitoring, and therapeutic decision-making in pediatric BCP-ALL, relevant to the fields of pediatrics, hematology, and clinical immunology. Furthermore, the principles underlying the system and method possess broader utility in diverse biological and medical applications requiring the detection of rare aberrant cells via multidimensional data analysis. BACKGROUND Acute lymphoblastic leukemia (ALL) represents the most prevalent pediatric malignancy, constituting approximately 25% of diagnosed childhood cancers. Approximately 60% of ALL cases manifest in individuals under 20 years of age, with an annual incidence exceeding 90 cases per 1 million persons in this demographic. Within the pediatric population, B-cell precursor acute lymphoblastic leukemia (BCP-ALL) is the predominant subtype. Current cytostatic or cytotoxic treatment protocols may not achieve complete eradication of malignant cells in all patients afflicted with BCP-ALL. Consequently, in a subset of patients, a small population of residual leukemic cells persists, typically within the bone marrow, which can subsequently proliferate and lead to disease relapse. This residual population of malignant cells is termed minimal residual disease (MRD). The quantitative assessment of MRD currently serves as the most significant prognostic indicator for predicting survival outcomes in pediatric ALL. The identification of MRD leukemic cells is conventionally accomplished through the analysis of their immunophenotype, characterized by a specific pattern of expression of cell surface and intracellular proteins known as phenotypic markers, often referred to as "cell markers" or "CD molecules" (Cluster of Differentiation antigens). A cell's phenotype is typically defined by a panel comprising 10 to 20 such markers. Detection of these markers is facilitated by the use of monoclonal antibodies possessing high specificity for individual markers. These antibodies are conjugated to distinct fluorochromes. Upon appropriate excitation, the fluorescence emitted by these fluorochrome-conjugated antibodies permits the detection and quantification of marker expression. The accurate measurement of MRD necessitates methodologies exhibiting high sensitivity, capable of detecting target cells at frequencies as low as 1 in 10,000 to 1 in 100,000 total cells (equivalent to 0.01% to 0.001%). Multiparameter flow cytometry currently fulfills this requirement for high-sensitivity detection. In flow cytometry, individual cells in suspension are hydrodynamically focused to pass sequentially through one or more focused laser beams, typically at rates ranging from 5,000 to 10,000 cells per second. As each cell intersects the laser beam(s), it scatters incident light in the forward direction (Forward Scatter, FSC) and orthogonal to the beam (Side Scatter, SSC). Concurrently, fluorochromes bound to the cell (e.g., via conjugated antibodies) are excited and emit fluorescence at specific wavelengths. The scattered light (FSC, SSC) and the fluorescence emissions across various wavelengths are captured by dedicated optical systems comprising filters and detectors. Each distinct detection pathway is referred to as a "channel" or "parameter". Contemporary flow cytometers are capable of measuring between 12 and 50 parameters simultaneously. The process of acquiring data from a sample using a flow cytometer is termed "acquisition". For each cell interrogated by the instrument, the signals captured by the detectors for each parameter are digitized and compiled into a data structure known as a digital "event". These events are stored electronically in a standardized file format, typically the FCS 3.0 (Flow Cytometry Standard) format established by the International Society for Advancement of Cytometry (ISAC). FCS files are compatible with various specialized software packages designed for flow cytometry data analysis. Within the context of this invention, the measured cells are represented as events, and the terms "cell" and "event" are used interchangeably. Conventional analysis of flow cytometry data typically involves graphical representation of events using one-dimensional "histograms" or two-dimensional diagrams termed "bivariate plots" or "bi-plots". A histogram displays the distribution of events based on the signal intensity of a single parameter (X-axis) versus the count of events (Y-axis). A bivariate plot typically depicts events as individual points, where the coordinates of each point correspond to the signal intensities of two selected parameters. Consequently, bivariate plots are also commonly referred to as "dot plots". Within a bivariate plot, events (cells) exhibiting similar coordinate values often aggregate into visually distinct groupings referred to as "clouds" or "clusters". In the present disclosure, the term "cluster" denotes such a grouping, representing a phenotypically distinct cell population. A specific cell population identified on a plot can be computationally isolated or selected using an electronic boundary termed a "gate". Gating allows for the extraction of quantitative information pertaining to the enclosed cell population or enables the selection of this population for subsequent analysis using further histograms or bivariate plots, wherein different combinations of parameters are visualized. The detection of MRD via this classical approach often requires sequential analysis of multiple bivariate plots, applying successive gates to progressively isolate the rare MRD cell cluster. The number of unique bivariate plots that can be generated from a dataset acquired using N distinct fluorochromes (parameters, excluding scatter) is given by the combination formula N(N- 1) / 2. For instance, an experiment utilizing 6 parameters necessitates the potential analysis of 15 unique bivariate plots. This number increases geometrically with the number of parameters; for example, a 16-parameter dataset yields 120 potential bivariate plots. Concomitantly, the theoretical number of potential distinct cell phenotypes (clusters) defined by combinations of positive / negative expression for N markers increases exponentially (2^N). For example, with 10 parameters, there are 1024 (2^10) potential phenotypic combinations, and with 16 parameters, this number increases to 65,536 (2^16). This conventional analytical strategy, relying on sequential gating across numerous bivariate plots, while often functional, presents several limitations. Firstly, it is inherently time-consuming and introduces subjectivity at multiple decision points where gates are manually drawn based on interpretations of marker expression patterns to discriminate between normal and aberrant cell populations. Secondly, as the number of parameters and corresponding bivariate plots increases, the analysis becomes fragmented, rendering comprehensive interpretation of the high-dimensional data challenging. Thirdly, and significantly, the assessment of individual cell clusters in isolation through sequential gating fails to provide insights into the phenotypic relationships or proximities between different cell populations within the multidimensional space. A further disadvantage of classical gating methods for MRD quantification pertains to the reliable identification and enumeration of extremely rare cell populations (e.g., requiring detection of at least 40 target cells amongst 4 x 10^6 total acquired cells for statistical robustness). Bivariate plots represent two-dimensional projections of the multidimensional data. Consequently, very small cell populations (0.01% - 0.001%) may be difficult to visually discern or accurately delineate using manual gates, particularly if they overlap with larger populations or are obscured by background noise in these limited dimensional views. Owing to these aforementioned limitations, the application of classical, manual gating-based analysis becomes increasingly complex, subjective, and potentially unreliable for flow cytometry studies involving more than approximately 8 parameters. This presents a significant challenge for effectively leveraging the data generated by modern high-parameter flow cytometry instrumentation for sensitive applications such as MRD detection. The detection of rare leukemic blast populations, characteristic of Minimal Residual Disease (MRD), necessitates the simultaneous assessment of numerous cell markers. Consequently, multiparameter flow cytometry (MPFC) utilizing 10 to 16 parameters is currently employed for MRD assessment, with an ongoing trend towards incorporating even higher numbers of parameters. This increase in dimensionality, however, renders data analysis using conventional manual gating strategies increasingly challenging. Consequently, there is a recognized need for alternative methods and systems capable of efficiently analyzing the large, high-dimensional datasets generated by MPFC. In recent years, computer-aided automated analysis techniques, including both unsupervised and supervised machine learning algorithms, have emerged as promising approaches for managing and interpreting the complexity inherent in high-parameter flow cytometry data [1, 2]. Patent document EP2347352B1 [3], published November 6, 2019, addresses certain limitations associated with the manual analysis of numerous bivariate plots derived from MPFC data. This document discloses a system and method wherein cell populations are organized within a hierarchical tree structure to facilitate their identification. The disclosed methods and systems enable a user to visualize and explore data corresponding to more than 512 potential phenotypes (arising from 9 or more parameters / fluorochromes) within a single interactive hierarchical tree graph. Furthermore, the system permits the use of multiple interconnected graphs. Subpopulations identified within one graphical representation can be selected, effectively acting as gates, to refine the visualization or analysis within another connected graph, potentially allowing exploration of up to 1024 phenotypes (corresponding to 10 parameters). The construction of these hierarchical graphs, however, relies on manual gating applied to histograms or bivariate plots, mirroring the process used in conventional flow cytometry data analysis, and is predicated on the operator's expert assessment of cell subpopulations. Consequently, the resulting cluster definitions are directly influenced by the number and placement of these manual gates, introducing significant operator-dependent subjectivity. The operator can interactively manipulate the defined clusters. A primary advantage of this approach lies in its ability to present quantitative information about multiple clusters and their corresponding cell subpopulations within a unified graphical interface, thereby enhancing visualization. However, this method does not provide automation for the initial definition or separation of cell clusters. While aiming to facilitate classical analysis, this approach, due to its reliance on manual gating derived from conventional plots, remains primarily suited for identifying and analyzing relatively abundant cell populations (typically >1%) and is not optimized for the detection of rare populations (<0.01%). Such rare populations are often difficult for an operator to visually discern and accurately gate on standard histograms or bivariate plots. In contrast, the present invention utilizes a fully automated clustering process, independent of operator subjectivity, enabling the detection of cell populations present at frequencies of 0.01% or even as low as 0.001% of the total measured cells. Patent document US 2003 / 0078703 A1 [4], published April 24, 2003, discloses a cytometric network system incorporating a central database designed to facilitate the analysis and review of substantial volumes of cytometric data originating from multiple flow cytometers. Data from multiple instruments are uploaded into the database where combined compensation and calibration matrices are automatically applied to correct for both spectral overlap between acquisition channels and instrument-dependent variations. Automated gating is subsequently applied to the data using common, predefined templates. Users accessing the system via client computers can view the processed data and perform statistical analyses on selected data subsets. A significant limitation of this approach is its reliance on predefined common templates for automated gating. This necessitates that target cell subpopulations exhibit consistent phenotypes and occupy similar locations within the multidimensional space (and thus, similar positions on bivariate plots) across all analyzed samples. This requirement is often not met in clinical samples from BCP-ALL patients, as leukemic cell populations frequently display phenotypic heterogeneity and occupy variable locations within the data space across different individuals or time points. Consequently, this template-based system is generally unsuitable for MRD assessment in BCP-ALL. In one embodiment of the present invention, while a database may be utilized for file storage, it functions as an auxiliary component. The core of the present invention resides in the specific method employed for analyzing multidimensional data to enable the detection of rare MRD cell populations. Patent document CN 103942415 A

[0009] published July 23, 2014, discloses a method for the automatic analysis of flow cytometry data. The method involves two primary steps: (1) determining the optimal number of clusters within the dataset, potentially using a criterion such as the Bayesian Information Criterion (BIC), and (2) subsequently performing automatic clustering or grouping of the data points, for example, using a t-mixture model, based on the determined number of clusters. The stated aim is to provide automatic and rapid analysis via computer software. A significant limitation of the method disclosed in CN 103942415 A is the lack of demonstrated application for identifying rare cell subpopulations or specifically for quantifying MRD. In the present invention, automated clustering represents only one component of the overall method. The method of the present invention incorporates additional steps beyond clustering, which are specifically designed and necessary for the robust identification and quantification of MRD cells in the context of BCP-ALL. Patent US 5,605,805 [5], issued February 25, 1997, discloses a method for the automated determination of acute leukemia lineage. The method utilizes data acquired from eight separate sample tubes, each stained with a specific antibody cocktail: 1. unstained; 2. isotype controls; 3. CD10 FITC, CD19 PE; 4. CD20 FITC, CD5 PE; 5. CD3 FITC, CD22 PE; 6. CD7 FITC, CD33 PE; 7. HLADR FITC, CD13 PE; and 8. CD34 FITC, CD38 PE. Data acquisition is performed using a flow cytometer, generating eight data files, typically comprising four parameters each (e.g., two fluorescence channels plus FSC and SSC). During data analysis, cell clusters within each file are identified using a computational algorithm. These identified clusters represent distinct cell populations. Characteristic features of each cell population (e.g., mean parameter values) are then compared against a predefined decision tree designed to identify known normal cell populations. Leukemic populations are subsequently identified by computationally excluding the recognized normal populations; the remaining aberrant populations are then classified into major leukemia lineages (B-lineage ALL, T-lineage ALL, AML, AUL, B-CLL) or designated as unknown. This disclosed method exhibits several limitations relevant to the present context: 1. The method, including its clustering algorithm and decision tree, is specifically tailored to the predefined 8-tube panel, which notably lacks markers typically required for sensitive MRD detection in BCP-ALL.2. The system is designed for broad lineage classification of the primary leukemic population and is not configured or validated for the quantification of rare MRD populations characteristic of BCP-ALL post-treatment. In contrast, the method of the present invention is designed to be flexible, accommodating a user-defined number of markers, including those currently established for MRD analysis and potentially incorporating novel markers as they become available. Furthermore, the system and method of the present invention are specifically developed to enable the measurement of rare MRD populations in BCP-ALL with high specificity and sensitivity. Patent document EP 1785899 A2 [6], published August 27, 2008, discloses a computational system for identifying event populations within multidimensional datasets acquired via flow cytometry. The system processes data using components including: a) a finite mixture model (FMM) comprising a weighted sum of multidimensional Gaussian probability density functions to represent cell populations; b) an expectation-maximization (EM) algorithm, potentially operating on data subsets, used to estimate the parameters of the FMM density functions, thereby performing clustering of the multidimensional data; and c) incorporation of 'expert knowledge', which may encompass data transformations or logical rules derived from prior understanding of the biological system. The integration of expert knowledge with the FMM is intended to yield more robust and accurate automated classification of data into distinct populations. This expert knowledge, in the context of flow cytometry, often relates to expected locations or characteristics of specific cell populations when viewed in certain two-dimensional projections (bivariate plots) based on prior biological understanding (e.g., from an expert hematologist). This knowledge might pertain to typical cluster locations, geometric shapes within specific bivariate projections, or relative positions between different clusters. The overall system aims to provide automated cell classification. The results of cluster identification (i.e., population definition) within the multidimensional dataset can be presented in human-interpretable formats, such as quantitative summaries or graphical visualizations (e.g., diagrams with color-coding to distinguish discrete populations). Relative to the present invention, the system described in EP 1785899 A2 exhibits certain limitations: 1. The demonstrated implementation focuses on classifying major, abundant leukocyte populations in peripheral blood (granulocytes, monocytes, lymphocytes). Its applicability to the detection and quantification of rare populations, such as MRD cells (often found in bone marrow), is not established or demonstrated.2. While the underlying FMM / EM approach is theoretically applicable to fluorescence data, the document may not explicitly demonstrate robust clustering based predominantly on multi-parameter fluorescence measurements crucial for immunophenotyping and MRD detection.3. The system implementation is described in the context of data with approximately 7 parameters. Its scalability and performance characteristics when applied to significantly higher-dimensional data (e.g., 14-16 parameters or more), typical of modern MRD assays, are not addressed. The present invention, in contrast, provides a solution specifically designed such that: a) rare populations (down to frequencies of 10^-5 or 0.001%) can be reliably detected, particularly within complex matrices like bone marrow; b) analysis effectively utilizes both light scatter and multiple fluorescence parameters; and c) the methodology is demonstrably capable of analyzing high-dimensional MPFC data (e.g., 16 parameters or more). U.S. Patent No.8,214,157 B2 [7], published July 3, 2012, and patent WO2007117423A2, sharing the same title and inventors, disclose methods and apparatus for processing flow cytometry data. These documents address challenges associated with analyzing multiple flow cytometry datasets by proposing a method and apparatus for representing multidimensional data distributions through a structured set of segments. This segmentation is achieved by partitioning the data space (potentially visualized via bivariate plots or analyzed directly in multidimensional space) using a technique described as involving hierarchical clustering, potentially incorporating principles of multidimensional minimum variance and equal probability assignment. The segmentation proceeds through multiple hierarchical levels (e.g., 5 levels), with the number of segments typically increasing at each subsequent level (e.g., doubling, potentially reaching 32 segments in total). The resulting segmented data, representing aspects of the multidimensional probability density, can then be converted into a linear numerical sequence, termed a "fingerprint." This fingerprint serves as a concise, one-dimensional representation of the information contained within the original multidimensional dataset. Furthermore, the described apparatus and methods include algorithms that utilize the segmentation framework derived from one dataset to process and represent a second dataset, thereby generating a comparable fingerprint. Comparison of these fingerprints allows for quantitative assessment of similarity or dissimilarity between two or more flow cytometry measurements. However, this approach exhibits several disadvantages with respect to its application for MRD measurement using MPFC: 1) If the segmentation process relies heavily on lower-dimensional representations like bivariate plots, which inherently capture only a fraction of the multidimensional information, there is a significant risk that extremely rare cell populations, such as MRD cells, may not be adequately resolved or distinctly captured within any single segment. This could lead to their omission from the analysis or their aggregation with larger, unrelated populations.2) The primary utility emphasized in these documents appears to be the comparison of overall sample profiles via fingerprints, rather than the specific detection and quantification of rare subpopulations within individual samples, which is the goal of MRD assessment.3) The methodology was demonstrated using relatively low-dimensional (e.g., 4- parameter) flow cytometry data, primarily for analyzing major lymphocyte subpopulations in peripheral blood. Its scalability, computational feasibility, and analytical effectiveness for processing the significantly higher-dimensional datasets (involving numerous parameters and potentially hundreds of bivariate plot combinations) generated by modern MPFC assays for MRD remain unproven. In contrast, the present invention directly analyzes the complete multidimensional dataset to computationally define distinct clusters based on inherent data structure. This approach enables the identification of clusters corresponding specifically to rare MRD populations and is explicitly designed for applicability to high-dimensional MPFC data. Patent document WO 2006 / 089190 A2 [8], published December 28, 2006, discloses a system intended to facilitate the detection of low-frequency tumor cells by characterizing phenotypic differences relative to their normal counterparts, through comparative analysis of data derived from normal and aberrant cell populations. The present invention relates to a system comprising a computer and algorithms that define centroids and associated radii within the n-dimensional space of flow cytometry data. These centroid-radius structures delineate clusters intended to represent stages of normal cellular maturation for a specific cell lineage, established through the analysis of a 'normal' cell dataset. The radii, extending from each centroid, conceptually encompass variation across all measured parameters, including fluorescence intensities. A notable aspect is that the definition of these clusters, potentially involving adjustments to centroid location or radius characteristics, incorporates manual intervention. Individual cells are mapped as points in this n-dimensional space based on their measured parameter values (e.g., fluorescence intensities). Cells falling within the defined centroid-radius regions constitute the clusters representing normal maturation stages. These multidimensional clusters are typically visualized graphically, for example, on a two-dimensional plot. Crucially, this methodology necessitates separate analyses: first, establishing the 'normal map' using data acquired from healthy individuals ('normal samples'), and second, analyzing data from patient samples ('pathological samples'). Identification of potentially aberrant (tumor) cells is subsequently achieved by comparing the patient data map to the established normal map, often involving visual inspection or computational subtraction, to highlight cells falling outside the defined normal patterns. An implementation is described wherein the normal map, derived from multiple (e.g., 25) normal samples, can be computationally overlaid or subtracted from the patient sample map, thereby potentially visualizing rare pathological cells, such as MRD cells. Relative to the present invention, this approach presents several disadvantages: 1) The reliance on manual adjustment for defining the centroids and radii of the normal maturation clusters introduces operator subjectivity into the process.2) The fundamental requirement to establish a reference 'normal map' using multiple samples from healthy donors presents significant practical challenges and potential ethical considerations, especially concerning the acquisition of pediatric bone marrow samples. Furthermore, such a reference map is inherently sensitive to variations in flow cytometry instrumentation, reagent lots (including antibody panels), staining protocols, and compensation settings. This sensitivity severely limits the transferability of a 'normal map' between different laboratories or even within the same laboratory over time if protocols or panels change. A new normal map would likely need to be generated for each specific antibody panel and potentially for each distinct instrument configuration.3) The described implementation utilized datasets comprising approximately 200,000 events. It remains uncertain whether the two-dimensional graphical representations employed would maintain sufficient resolution and clarity to reliably distinguish rare MRD populations when analyzing the substantially larger datasets (e.g., 4 x 10^6 events or more) typically acquired to achieve the requisite sensitivity (e.g., 10^-5) for clinical MRD detection.4) The methodology was demonstrated using 6- parameter flow cytometry data. Its performance, the feasibility of constructing meaningful normal maps, and the interpretability of the resulting comparisons with significantly higher dimensionality (e.g., 10-16 parameters or more) are not established. The present invention overcomes these limitations by: 1) utilizing a fully automated clustering process that does not rely on manual intervention or subjective decisions for cluster definition; 2) eliminating the requirement for separate 'normal' reference samples and the associated practical and ethical difficulties; and 3) being specifically designed and validated for analyzing large, high- parameter datasets (e.g., 4 x 10^6 events, 16 parameters or more) characteristic of modern MPFC assays, thereby enabling high-sensitivity MRD detection in contexts such as BCP-ALL. Scientific publications by Jafari et al. (2018)

[0010] , Kárai et al. (2021)

[0011] , and Shopsowitz et al. (2022) [12, 13] describe the application and optimization of Radar Plots for visualizing and analyzing flow cytometry data, including for MRD assessment in B-ALL. A Radar Plot is a graphical technique for displaying multidimensional data within a two-dimensional diagram. Individual parameters are represented as axes radiating from a common central point. The configuration of the plot, including the center position, the arrangement and scaling of axes, can often be adjusted (potentially manually) to optimize the visual separation of different cell clusters projected onto this 2D representation. Gates may sometimes be applied to selected populations visualized on the Radar Plot for further analysis. Kárai et al.

[0011] focused on optimizing the Radar Plot configuration (center position, axis arrangement, direction, and magnitude) to achieve improved differentiation of cell populations, developing specific Radar Plot templates tailored for visualizing pathological blasts in 6- parameter flow cytometry data relevant to MRD. It was noted that these optimized templates are specific to the particular antibody panel employed. Shopsowitz et al. [12, 13] further explored the optimization of Radar Plots using machine learning techniques. Their approach involved generating a library of candidate Radar Plot configurations. Each configuration was used to transform a balanced training dataset, and classifiers (e.g., Support Vector Machines, SVM) were subsequently trained to discriminate between normal (e.g., mature B-cells or hematogones) and abnormal events within each transformed Radar Plot view. The performance of each SVM model, assessed using metrics like the Area Under the Curve (AUC) from Receiver Operating Characteristic (ROC) analysis, served as a benchmark to rank the candidate Radar Plot configurations. The best-performing configurations were selected iteratively for further refinement until the performance plateaued. This data-driven approach aimed to identify an optimized Radar Plot representation for discriminating normal from abnormal cells. However, this work was also primarily demonstrated using 6-parameter cytometry data, and the resulting optimized plots remain specific to the training data and the antibody panel used. Compared to the present invention, these Radar Plot-based solutions exhibit certain limitations: 1) Optimization of the Radar Plot visualization often involves either manual adjustment of parameters or a template selection process (even if guided by metrics like AUC), which can introduce operator dependence or require substantial effort to determine the most effective view for a given dataset or task.2) The resulting optimized plots or templates are typically specific to the particular antibody panel and potentially the flow cytometer characteristics used during optimization. This inherent specificity poses a significant barrier to the general applicability and standardization of the method across different laboratories or when protocols are updated.3) While demonstrated primarily with 6-parameter data, the effectiveness and interpretability of Radar Plots tend to diminish as the dimensionality of the data increases. Practical experience and observations, such as those noted by Shopsowitz et al., indicate that designing informative Radar Plots for high-parameter data (e.g., 14 or 16 parameters and beyond) becomes exceedingly challenging due to the vast combinatorial space of possible axis arrangements and transformations. Relying on intuition or trial-and-error becomes impractical, and it is difficult to ascertain whether any given Radar Plot configuration is truly optimal for the analytical task with high-dimensional data. In contrast, the present invention employs fully automated multidimensional clustering algorithms that operate directly on the high-dimensional data, circumventing the visualization bottlenecks and optimization complexities associated with methods like Radar Plots when applied to high- parameter datasets. The system and method of disclosed herein are designed for direct application to MPFC data comprising 12-16 parameters or14-16 parameters (or more), facilitating sensitive MRD measurement in BCP-ALL. A key advantage is that the core analytical methodology is designed to be robust and independent of specific antibody panel compositions or variations in flow cytometer types, thereby enhancing its potential for widespread adoption and standardization across diverse laboratory environments. A scientific publication by Shopsowitz K. et al.

[0012] , dated 2024, describes a system designated MAGIC-DR for the measurement of Minimal Residual Disease (MRD) in acute myeloid leukemia (AML). The MAGIC-DR system incorporates two primary components: a supervised machine learning subsystem utilizing an XGBoost classifier, and a subsystem employing Uniform Manifold Approximation and Projection for Dimension Reduction (UMAP) for unsupervised data representation, typically generating a two-dimensional visualization. XGBoost employs a gradient boosting methodology, constructing a predictive model through sequential addition of decision trees until further performance improvements are minimal. As a supervised learning technique, the training of the XGBoost classifier necessitates annotated data indicating the true cellular composition of training samples. Accordingly, the classifier is trained to distinguish leukemic cells by analyzing a cohort of reference samples, comprising both samples confirmed negative for AML leukemic cells and samples confirmed positive. The curation of training data and subsequent model training generally involves operator oversight. Following successful training, the XGBoost classifier is reported to reliably identify immunophenotypically heterogeneous AML blasts and promonocytes with high predictive accuracy. The output of the XGBoost classifier may be integrated with the UMAP subsystem, potentially providing classifications or probabilities for cells visualized by UMAP. The UMAP subsystem functions to perform unsupervised dimensionality reduction, projecting the high-dimensional flow cytometry data onto a lower-dimensional space, typically a two-dimensional graph, wherein distinct cell populations may manifest as separate visual clusters. Identified leukemic populations can be distinctly visualized (e.g., via color-coding) and potentially selected via gating for subsequent statistical analysis. Relative to the present invention, the MAGIC-DR system exhibits certain limitations: 1) A significant constraint inherent to supervised learning models like XGBoost is their sensitivity to the specific panel of antibodies (markers) and potentially the instrument configuration used during training. Even minor alterations to the antibody panel or flow cytometer settings could potentially degrade system performance, necessitating retraining of the classifier with newly acquired, appropriately labeled data. This dependence on specific training configurations presents a substantial challenge for the widespread implementation and standardization of such a system across different laboratories or protocols.2) The UMAP subsystem, while effective for visualizing complex datasets, also presents interpretive challenges. Common to many dimensionality reduction techniques, UMAP projections can yield complex patterns of clusters whose precise biological identity, size, and relative positions may vary considerably between samples. Consequently, when analyzing MRD, where the target population is rare and may possess an aberrant phenotype, it is often difficult to prospectively identify the specific location of the MRD cluster within the UMAP visualization without prior knowledge or additional guidance (such as from a paired classifier like XGBoost). In contrast, the system and method of the present invention are designed to be independent of specific antibody panel compositions and flow cytometer types, thereby facilitating broader applicability across diverse laboratory settings without the need for model retraining specific to each configuration. A scientific publication by Reiter M. et al.

[0014] in 2019 describes a supervised machine learning approach for MRD assessment, employing a combination of multiple Gaussian Mixture Models (GMMs) adapted into a parametric density modeling framework. This approach aims to determine optimal weighting factors for a linear combination of stored GMMs to effectively represent new, previously unseen samples by interpolating information from the stored reference samples. The supervised training utilizes flow cytometry data derived from samples previously analyzed with manual gating to establish ground truth labels. The described training dataset comprised 6-parameter flow cytometry data from 337 pediatric B-ALL bone marrow samples, acquired at day 15 of induction therapy across three distinct laboratories. In assessing the method's sensitivity, the authors determined a practical resolution limit for accurate MRD quantification at a threshold of 0.05%. The authors posit that their composite GMM approach demonstrated superior performance compared to alternative machine learning techniques, including Support Vector Machines (SVM), Deep Neural Networks (DNN), and a single GMM approach, within the specific context of pediatric B-ALL MRD assessment using the available flow cytometry data. While advantageously focused on B-ALL MRD measurement, the method described by Reiter et al. presents several limitations: 1) A critical requirement for clinical utility is robust predictive performance, even when analyzing samples processed on different flow cytometers, potentially with unknown instrument settings or variations in staining protocols. Although the training data incorporated samples from multiple laboratories to address this, the authors observed that model performance was optimal when test and training samples originated from the same system (instrument and staining panel). Reduced performance was noted when applying the model across different systems, indicating potential difficulties in achieving consistent performance upon deployment in diverse laboratory environments.2) The method's ability to reliably detect rare or atypical MRD phenotypes encountered in B-ALL may be constrained. Samples exhibiting uncommon phenotypes (e.g., associated with pro-B-ALL or lineage switching events) might not be adequately represented by the interpolation capabilities of the model derived from the specific training dataset.3) The model development and validation were performed using 6-parameter flow cytometry data. Its applicability and performance characteristics when applied to higher- dimensional MPFC data (e.g., 14-16 parameters or more) remain undetermined.4) The reported MRD quantification sensitivity limit of 0.05% falls short of the current clinical standard, which typically requires sensitivity of at least 0.01%. The present invention addresses these limitations. Its core clustering methodology operates automatically on multiparameter data without reliance on supervised machine learning or interpolation from stored reference samples. This independence facilitates implementation across different laboratories. Furthermore, the method is capable of detecting rare phenotypes. The present invention has been implemented and validated using data from 14-parameter flow cytometry, achieving MRD detection sensitivity exceeding the 0.01% threshold. Based on the review of the current state of the art, it is apparent that existing methodologies and systems for measuring MRD in B-ALL face challenges in providing consistently effective, stable, reliable, and accurate results, particularly concerning automation, objectivity, sensitivity, and applicability across diverse high-parameter datasets and laboratory settings. The foregoing analysis indicates a persistent need for improved systems and methods that overcome the limitations inherent in prior art solutions. Specifically, there is a need for a system and method capable of identifying rare subpopulations of cells within a larger cell population derived from a biological sample, in a reproducible and objective manner, without dependence on pre-existing reference samples or panel-specific trained models. The objective of the present invention is to address these deficiencies by providing a system and method that enable the detection of rare cells with high reproducibility and sensitivity, without requiring supervised machine learning for the core analysis. The disclosed system and methods are particularly applicable to the measurement of MRD in B-ALL using high-parameter flow cytometry data. Description of the Figures The present invention will be further described by way of example only, with reference to the accompanying diagrams and figures, wherein: ^ Figure 1: A schematic representation of the system for analyzing multiparameter flow cytometry data to measure MRD in B-ALL (System 100). ^ Figure 2: A simplified block diagram of Subsystem-1 (101). ^ Figure 3: A simplified block diagram of Subsystem-2 (103). ^ Figure 4: A simplified block diagram of Subsystem-3 (106). ^ Figure 5: A detailed block diagram of the downstream analysis process, according to an embodiment of the invention. ^ Figure 6: Illustrations of graphs generated by the Data Reduction Module (DRM) 108. ^ Figure 7: Illustrations of graphs generated by the Data Cleaning Module (DCM) 109. ^ Figure 8: Illustrations of graphs generated by the Data Clustering Module (DClM) 110. ^ Figure 9: Illustrations of graphs generated by the Cluster Analysis Module (CAM) 111. ^ Figure 10: Illustrations of the sensitivity and specificity analysis results of the system and method (100). Reference Numerals in the Figures: ^ 101: Subsystem-1 for measurement ^ 102: Patient sample ^ 103: Subsystem-2 for data acquisition ^ 104: FCS file ^ 105: File repository ^ 106: Subsystem-3 for data analysis ^ 107: External server ^ 108: Data Reduction Module (DRM) ^ 109: Data Cleaning Module (DCM) ^ 110: Data Clustering Module (DClM) ^ 111: Cluster Analysis Module (CAM) ^ 112: Communication network (LAN) ^ 113: Internet ^ 201: Multiparameter flow cytometer ^ 202: Digital data converter ^ 203: Network interface ^ 301: Central processing unit (CPU) ^ 302: Graphics processing unit (GPU) ^ 303: User interface ^ 304: Input device ^ 305: Primary memory ^ 306: Secondary memory ^ 401: Central processing unit (CPU) ^ 402: Graphics interface ^ 403: FCSdr file ^ 404: FCScd file ^ 405: CG file ^ 406: CS file ^ 407: AG file ^ 408: AS file ^ 501: Gate-1 for leukocytes ^ 502: Gate-2 for single cells ^ 503: Gate-3 for B-cells ^ 505: Data cleaning ^ 507: Graph of raw events ^ 508: Graph of cleaned events ^ 509: Statistical result ^ 510: Data clustering ^ 511: Cluster graph – CG file ^ 512: Heatmap – HM file ^ 513: Cluster analysis ^ 514: Cluster graph ^ 515: Gate placement ^ 516: Cluster selection ^ 517: Cluster statistics ^ 518: MRD result for B-ALL ^ 601: Bivariate plot FSC vs. SSC ^ 602: FSC – forward scatter ^ 603: SSC – side scatter ^ 605: Leukocytes ^ 606: Cell debris ^ 607: Bivariate plot FSC-H vs. FSC-A ^ 608: FSC-H – intensity ^ 609: FSC-A – area ^ 610: Single cells ^ 611: Doublets ^ 612: Bivariate plot FSC vs. CD19 ^ 613: CD19 ^ 614: B-cells ^ 701: Data anomaly ^ 702: Time parameter ^ 801: Cluster tree ^ 802: Cluster map ^ 803: Heatmap ^ 804: MRD cluster for B-ALL ^ 805: Metacluster ^ 901: Cluster graph ^ 902: Flow cytometer parameters ^ 903: Fluorescence intensity ^ 904: Cluster lines ^ 905: Cluster gate ^ 906: Cluster gate ^ 1001: Sensitivity and specificity by classical DIVA method ^ 1002: Sensitivity and specificity via the invention SUMMARY OF THE INVENTION AND DETAILED DESCRIPTION The technical problems identified in the prior art, including the specific issues associated with existing MRD analysis techniques, are addressed by the system and method disclosed herein for analyzing multiparameter flow cytometry (MPFC) data to measure minimal residual disease (MRD) in B-cell acute lymphoblastic leukemia (B-ALL). The features, advantages, structure, and operation of various embodiments of the present invention are described in detail below, with reference to the accompanying drawings. In one aspect the invention relates to a computer-implemented method for analysing minimal residual disease (MRD) associated with paediatric B-cell precursor acute lymphoblastic leukaemia (B-ALL) in a sample from a patient using multiparameter flow cytometry (MPFC) data, said method comprising: i) isolating a subset of data corresponding to a target cell lineage; ii) applying an automated algorithm to identify and remove data anomalies from the data subset to produce a cleaned data subset; iii) applying an automated, unsupervised clustering algorithm to group events in the cleaned data subset into a plurality of clusters based on the multiparameter data; and iv) analysing statistical properties of the plurality of clusters to identify at least one cluster corresponding phenotypically to MRD cells. In some embodiments, the patient is a human. In some embodiments, the patient has previously been diagnosed with pediatric B-cell precursor acute lymphoblastic leukemia. In some embodiments, the patient has previously been treated for pediatric B-cell precursor acute lymphoblastic leukemia. In some embodiments, step iii) further comprises generating graphical representations of the clusters. In some embodiments, step i) is performed by a data reduction module (DRM). In some embodiments, step ii) is performed by a data cleaning module (DCM). In some embodiments, step iii) is performed by a data clustering module (DClM). In some embodiments, step iv) is performed by a cluster analysis module (CAM). In some embodiments, the method comprises: i) isolating a subset of data corresponding to a target cell lineage using a Data Reduction Module (DRM), ii) applying an automated algorithm to identify and remove data anomalies from the data subset to produce a cleaned data subset using a Data Cleaning Module (DCM), iii) applying an automated, unsupervised clustering algorithm to group events in the cleaned data subset into a plurality of clusters based on the multiparameter data using a Data Clustering Module (DClM); and iv) analyzing statistical properties of the plurality of clusters to identify at least one cluster corresponding phenotypically to MRD cells using a Cluster Analysis Module (CAM), wherein the DRM, the DCM, the DClM and the CAM comprise an Analytical Subsystem. In some embodiments, the method comprises: i) transmitting a file to the Analytical Subsystem or a file repository; ii) within the Data Reduction Module (DRM) of the Analytical Subsystem, retrieving the file and isolating a subset of data corresponding to a target cell lineage; iii) within a Data Cleaning Module (DCM) of the Analytical Subsystem, retrieving the reduced data file and applying an automated algorithm to identify and remove data anomalies from the data subset to produce a cleaned data subset; iv) within a Data Cleaning Module (DCM) of the Analytical Subsystem, retrieving the reduced data file and applying an automated, unsupervised clustering algorithm to group events in the cleaned data subset into a plurality of clusters based on the multiparameter data; and v) within a Cluster Analysis Module (CAM) of the Analytical Subsystem, retrieving the cluster statistics and analysing statistical properties of the plurality of clusters to identify at least one cluster corresponding phenotypically to MRD cells. In some embodiments, the DRM isolates a subset of data corresponding to a target cell lineage by applying a data reduction process. In some embodiments, the data reduction process comprises identifying a total leukocyte population within a reference gate and isolating a B-cell population within a subsequent gate, thereby generating a reduced data file containing data for the B-cell population and storing a count of the total leukocyte population. In some embodiments, the CAM determines a quantity of cells within the identified at least one cluster corresponding to MRD cells and calculates a percentage of MRD cells by dividing the quantity of cells determined by the DRM by the count of the total leukocyte population stored by the DRM and multiplying by 100. In some embodiments, the method comprises: i) transmitting a file to the Analytical Subsystem or a file repository; ii) within the Data Reduction Module (DRM) of the Analytical Subsystem, retrieving the file and applying a data reduction process, comprising identifying a total leukocyte population within a reference gate and isolating a B-cell population within a subsequent gate, thereby generating a reduced data file containing data for the B-cell population and storing a count of the total leukocyte population; iii) within a Data Cleaning Module (DCM) of the Analytical Subsystem, retrieving the reduced data file and applying an automated data cleaning algorithm to remove data anomalies, thereby generating a cleaned data file; iv) within a Data Clustering Module (DClM) of the Analytical Subsystem, retrieving the cleaned data file and applying an automated, unsupervised clustering algorithm to partition events within the cleaned data file into a plurality of clusters based on their multiparameter data; v) within the Cluster Analysis Module (CAM), analysing the cluster statistics and optionally visualizing cluster expression profiles to identify at least one specific cluster exhibiting a multiparameter phenotype characteristic of leukemic MRD cells for B-ALL; and vi) within the Cluster Analysis Module (CAM), (1) determining a quantity of cells within the identified at least one specific cluster corresponding to MRD cells; and calculating a percentage of MRD cells by dividing the quantity of cells identified in step (1) by the count of the total leukocyte population stored in step (ii) and multiplying by 100. In some embodiments, following step iv), the DClM further generates graphical representations of the plurality of clusters and generates cluster statistics characterizing each cluster and stores the cluster statistics. In some embodiments, the file transmitted to the Analytical Subsystem or the file repository is an FCS file. In some embodiments, the data reduction process performed by the DRM comprises displaying data from the file on the user interface, thereby enabling an operator to perform steps including: a) defining Gate-1 on a first bivariate plot to encompass the leukocyte population and storing the count of events within Gate-1 for subsequent MRD percentage calculation; and b) defining Gate-3 on a subsequent bivariate plot to encompass the B-cell population, wherein the data corresponding to events within Gate-3 constitutes the reduced data file utilized for subsequent clustering and identification of leukemic MRD cells. In some embodiments, the automated data cleaning algorithm applied by the DCM, in addition to generating the cleaned data file, causes the display on the user interface of at least one graphical representation comparing data characteristics before cleaning and after cleaning. In some embodiments, the display on the user interface of at least one graphical representation comparing data characteristics before cleaning and after cleaning is accompanied by corresponding event counts, thereby enabling an operator to visually assess the effectiveness of the cleaning process and determine if adjustments to preceding steps, such as gating within the DRM, are warranted. In some embodiments, the graphical representations generated by the DClM and displayed on the user interface comprise at least one of a cluster tree, a cluster map, and a heatmap, wherein: a) the cluster tree visually represents hierarchical relationships among the plurality of clusters, potentially illustrating groupings into metaclusters and indicating relative phenotypic distances between clusters; and b) the heatmap displays the plurality of clusters against the measured parameters, utilizing a color scale to represent the expression level of each parameter within each cluster, thereby providing a comprehensive visualization of the immunophenotype associated with the cell population within each respective cluster. In further embodiments, the heatmap displays the plurality of clusters as rows and / or the heatmap displays measured parameters as columns and / or the expression level is median or mean intensity. In some embodiments, the analysis performed by the CAM using the cluster statistics comprises displaying cluster information on the user interface via a graphical representation, wherein: a) the graphical representation displays axes corresponding to the measured parameters and signal intensity; b) each cluster of the plurality of clusters is represented as a profile line interconnecting points corresponding to the characteristic signal intensity value for that cluster for each respective parameter; and c) the method further comprises enabling the application of one or more analytical gates based on signal intensity thresholds for selected parameters, wherein logical combinations of said gates using operators such as AND, OR, or NOT can be employed to computationally filter or select clusters matching a predefined MRD phenotype. In some embodiments, utilizing algorithms within DRM, the data can be graphically displayed on the user interface. For example, a bivariate plot showing forward scatter (FSC) versus side scatter (SSC) can be generated. On this plot, the operator may apply an initial gate, to isolate the total leukocyte population while excluding non-cellular debris and platelets. The events within the leukocyte gate can then be visualized on a subsequent bivariate plot, perhaps plotting FSC height (FSC-H) versus FSC area (FSC-A), to discriminate between single cells and doublet events. A second gate selects the single cell population. These single leukocyte events are further displayed on another bivariate plot, typically plotting a scatter parameter (e.g., FSC 602) against a key lineage marker, such as the B-cell specific marker CD19. A third gate is applied to select CD19-positive cells, which encompass the B-lineage cells, including the target leukemic MRD cells in B-ALL. In some embodiments, the Data Cleaning Module (DCM) removes residual data anomalies that might have occurred during the acquisition from patient sample. In some embodiments, data anomalies are associated with technical artifacts, such as those caused by air bubbles, cell aggregates that passed the doublet discrimination gate, transient changes in flow rate, or electronic noise. In some embodiments, the Data Clustering Module (DClM) executes the core task of grouping the high-dimensional data for each of the cleaned cells into distinct clusters. In some embodiments each of the distinct clusters represents a phenotypically homogeneous cell subpopulation. In some embodiments, the CAM analyzes the defined clusters to identify the specific cluster(s) exhibiting the phenotypic characteristics of the leukemic MRD cells relevant to the patient's B- ALL diagnosis. In some embodiments, the DRM generates a FCSdr file. In some embodiments, the DCM generates a FCScd file. In some embodiments, the DCM is configured to apply an automated algorithm selected from the group consisting of flowClean, flowAI, and PeacoQC. In further embodiments, the automated algorithm identifies and removes events likely associated with technical artifacts. In some embodiments, the events likely associated with technical artifacts are selected from the group consisting of air bubbles, cell aggregates that passed the doublet discrimination gate, transient changes in flow rate, or electronic noise. In some embodiments, the DClM is configured to apply an automated, unsupervised clustering algorithm selected from the group consisting of FlowSOM, Phenograph, X-shift, flowMeans, and flowClust. In some embodiments, the CAM generates a CS file. In some embodiments, the Cluster Analysis Module is ClusterExplorer. In a further aspect, the invention relates to a method for analysing minimal residual disease (MRD) associated with paediatric B-cell precursor acute lymphoblastic leukaemia (B-ALL) in a sample from a patient using multiparameter flow cytometry (MPFC) data, the method comprising: i) treating the sample from the patient with a panel of fluorochrome-conjugated antibodies specific for a plurality of cellular markers relevant to distinguishing leukemic B-ALL cells from non-leukemic B-ALL cells; ii) measuring the treated sample from the patient using an MPFC instrument to generate measured signals for each of a plurality of events across multiple parameters, and converting said measured signals into a digital data format; and iii) using a processor, performing a computer-implemented method as disclosed herein. In some embodiments, the patient is a human. In some embodiments, the patient has previously been diagnosed with pediatric B-cell precursor acute lymphoblastic leukemia. In some embodiments, the patient has previously been treated for pediatric B-cell precursor acute lymphoblastic leukemia. In some embodiments, the computer-implemented method of step iii) is a computer- implemented method for analysing minimal residual disease (MRD) associated with paediatric B- cell precursor acute lymphoblastic leukaemia (B-ALL) in a sample from a patient using multiparameter flow cytometry (MPFC) data, said method comprising: i) isolating a subset of data corresponding to a target cell lineage; ii) applying an automated algorithm to identify and remove data anomalies from the data subset to produce a cleaned data subset; iii) applying an automated, unsupervised clustering algorithm to group events in the cleaned data subset into a plurality of clusters based on the multiparameter data; and iv) analysing statistical properties of the plurality of clusters to identify at least one cluster corresponding phenotypically to MRD cells. In some embodiments, step iii) further comprises generating graphical representations of the clusters. In some embodiments, step i) is performed by a data reduction module (DRM). In some embodiments, step ii) is performed by a data cleaning module (DCM). In some embodiments, step iii) is performed by a data clustering module (DClM). In some embodiments, step iv) is performed by a cluster analysis module (CAM). In some embodiments, the computer-implemented method comprises: i) isolating a subset of data corresponding to a target cell lineage using a Data Reduction Module (DRM), ii) applying an automated algorithm to identify and remove data anomalies from the data subset to produce a cleaned data subset using a Data Cleaning Module (DCM), iii) applying an automated, unsupervised clustering algorithm to group events in the cleaned data subset into a plurality of clusters based on the multiparameter data using a Data Clustering Module (DClM); and iv) analyzing statistical properties of the plurality of clusters to identify at least one cluster corresponding phenotypically to MRD cells using a Cluster Analysis Module (CAM), wherein the DRM, the DCM, the DClM and the CAM comprise an Analytical Subsystem. In some embodiments, the computer-implemented method comprises: i) transmitting a file to the Analytical Subsystem or a file repository; ii) within the Data Reduction Module (DRM) of the Analytical Subsystem, retrieving the file and isolating a subset of data corresponding to a target cell lineage; iii) within a Data Cleaning Module (DCM) of the Analytical Subsystem, retrieving the reduced data file and applying an automated algorithm to identify and remove data anomalies from the data subset to produce a cleaned data subset; v) within a Data Cleaning Module (DCM) of the Analytical Subsystem, retrieving the reduced data file and applying an automated, unsupervised clustering algorithm to group events in the cleaned data subset into a plurality of clusters based on the multiparameter data; and iv) within a Cluster Analysis Module (CAM) of the Analytical Subsystem, retrieving the cluster statistics and analysing statistical properties of the plurality of clusters to identify at least one cluster corresponding phenotypically to MRD cells. In some embodiments, the DRM isolates a subset of data corresponding to a target cell lineage by applying a data reduction process. In further embodiments, the data reduction process comprises identifying a total leukocyte population within a reference gate and isolating a B-cell population within a subsequent gate, thereby generating a reduced data file containing data for the B-cell population and storing a count of the total leukocyte population. In some embodiments, the CAM determines a quantity of cells within the identified at least one cluster corresponding to MRD cells and calculates a percentage of MRD cells by dividing the quantity of cells determined by the DRM by the count of the total leukocyte population stored by the DRM and multiplying by 100. In some embodiments, the computer-implemented method comprises: i) transmitting a file to the Analytical Subsystem or a file repository; ii) within the Data Reduction Module (DRM) of the Analytical Subsystem, retrieving the file and applying a data reduction process, comprising identifying a total leukocyte population within a reference gate and isolating a B-cell population within a subsequent gate, thereby generating a reduced data file containing data for the B-cell population and storing a count of the total leukocyte population; iii) within a Data Cleaning Module (DCM) of the Analytical Subsystem, retrieving the reduced data file and applying an automated data cleaning algorithm to remove data anomalies, thereby generating a cleaned data file; iv) within a Data Clustering Module (DClM) of the Analytical Subsystem, retrieving the cleaned data file and applying an automated, unsupervised clustering algorithm to partition events within the cleaned data file into a plurality of clusters based on their multiparameter data; v) within the Cluster Analysis Module (CAM), analysing the cluster statistics and optionally visualizing cluster expression profiles to identify at least one specific cluster exhibiting a multiparameter phenotype characteristic of leukemic MRD cells for B- ALL; and vi) within the Cluster Analysis Module (CAM), (1) determining a quantity of cells within the identified at least one specific cluster corresponding to MRD cells; and calculating a percentage of MRD cells by dividing the quantity of cells identified in step (1) by the count of the total leukocyte population stored in step (ii) and multiplying by 100. In some embodiments, following step iv), the DClM further generates graphical representations of the plurality of clusters and generates cluster statistics characterizing each cluster and stores the cluster statistics. In some embodiments, the file transmitted to the Analytical Subsystem or the file repository is an FCS file. In some embodiments, the data reduction process performed by the DRM comprises displaying data from the file on the user interface, thereby enabling an operator to perform steps including: a) defining Gate-1 on a first bivariate plot to encompass the leukocyte population and storing the count of events within Gate-1 for subsequent MRD percentage calculation; and b) defining Gate-3 on a subsequent bivariate plot to encompass the B-cell population, wherein the data corresponding to events within Gate-3 constitutes the reduced data file utilized for subsequent clustering and identification of leukemic MRD cells. In some embodiments, the automated data cleaning algorithm applied by the DCM, in addition to generating the cleaned data file, causes the display on the user interface of at least one graphical representation comparing data characteristics before cleaning and after cleaning, optionally accompanied by corresponding event counts, thereby enabling an operator to visually assess the effectiveness of the cleaning process and determine if adjustments to preceding steps, such as gating within the DRM, are warranted. In some embodiments, the graphical representations generated by the DClM and displayed on the user interface comprise at least one of a cluster tree, a cluster map, and a heatmap, wherein: a) the cluster tree visually represents hierarchical relationships among the plurality of clusters, potentially illustrating groupings into metaclusters and indicating relative phenotypic distances between clusters; and b) the heatmap displays the plurality of clusters against the measured parameters, utilizing a color scale to represent the expression level of each parameter within each cluster, thereby providing a comprehensive visualization of the immunophenotype associated with the cell population within each respective cluster. In further embodiments, the heatmap displays the plurality of clusters as rows and / or the heatmap displays measured parameters as columns and / or the expression level is median or mean intensity. In some embodiments, the analysis performed by the CAM using the cluster statistics comprises displaying cluster information on the user interface via a graphical representation, wherein: a) the graphical representation displays axes corresponding to the measured parameters and signal intensity; b) each cluster of the plurality of clusters is represented as a profile line interconnecting points corresponding to the characteristic signal intensity value for that cluster for each respective parameter; and c) the method further comprises enabling the application of one or more analytical gates based on signal intensity thresholds for selected parameters, wherein logical combinations of said gates using operators such as AND, OR, or NOT can be employed to computationally filter or select clusters matching a predefined MRD phenotype. In some embodiments, utilizing algorithms within DRM, the data can be graphically displayed on the user interface. For example, a bivariate plot showing forward scatter (FSC) versus side scatter (SSC) can be generated. On this plot, the operator may apply an initial gate, to isolate the total leukocyte population while excluding non-cellular debris and platelets. The events within the leukocyte gate can then be visualized on a subsequent bivariate plot, perhaps plotting FSC height (FSC-H) versus FSC area (FSC-A), to discriminate between single cells and doublet events. A second gate selects the single cell population. These single leukocyte events are further displayed on another bivariate plot, typically plotting a scatter parameter (e.g., FSC 602) against a key lineage marker, such as the B-cell specific marker CD19. A third gate is applied to select CD19-positive cells, which encompass the B-lineage cells, including the target leukemic MRD cells in B-ALL. In some embodiments, the Data Cleaning Module (DCM) removes residual data anomalies that might have occurred during the acquisition from patient sample. In some embodiments, data anomalies are associated with technical artifacts, such as those caused by air bubbles, cell aggregates that passed the doublet discrimination gate, transient changes in flow rate, or electronic noise. In some embodiments, the Data Clustering Module (DClM) executes the core task of grouping the high-dimensional data for each of the cleaned cells into distinct clusters. In some embodiments each of the distinct clusters represents a phenotypically homogeneous cell subpopulation. In some embodiments, the CAM analyzes the defined clusters to identify the specific cluster(s) exhibiting the phenotypic characteristics of the leukemic MRD cells relevant to the patient's B- ALL diagnosis. In some embodiments, the DRM generates a FCSdr file. In some embodiments, the DCM generates a FCScd file. In some embodiments, the DCM is configured to apply an automated algorithm selected from the group consisting of flowClean, flowAI, and PeacoQC. In further embodiments, the automated algorithm identifies and removes events likely associated with technical artifacts. In some embodiments, the events likely associated with technical artifacts are selected from the group consisting of air bubbles, cell aggregates that passed the doublet discrimination gate, transient changes in flow rate, or electronic noise. In some embodiments, the DClM is configured to apply an automated, unsupervised clustering algorithm selected from the group consisting of FlowSOM, Phenograph, X-shift, flowMeans, and flowClust. In some embodiments, the CAM generates a CS file. In some embodiments, the Cluster Analysis Module is ClusterExplorer. In some embodiments, the sample from a patient comprises bone marrow cells. In some embodiments, the sample from a patient comprises peripheral blood cells. In some embodiments, the sample from a patient comprises at least 4x106cells. In some embodiments, the panel of fluorochrome-conjugated antibodies specific for a plurality of cellular markers comprises a combination of antibodies directed against markers, wherein the combination of antibodies comprises: i) antibodies targeting markers CD45, CD20, CD34, CD38, CD10, CD58, CD66c, CD73, CD81, CD123, CD304, CD44, CD86, CD99 and CD371, and ii) antibodies targeting markers CD19 and / or CD22, wherein the antibodies are conjugated with fluorochromes. In some embodiments, the panel of fluorochrome-conjugated antibodies specific for a plurality of cellular markers further comprises an antibody targeting CD36. In some embodiments, the panel of fluorochrome-conjugated antibodies comprises the panel of Table 1. In some embodiments, the MPFC instrument is a flow cytometer. In some embodiments, the measured signals are fluorescence emissions of antibodies conjugated to distinct fluorochromes, upon appropriate laser excitation. A further aspect of the invention relates to a system comprising a memory and a processor configured to: i) isolate a subset of data corresponding to a target cell lineage; ii) apply an automated algorithm to identify and remove data anomalies from the data subset to produce a cleaned data subset; iii) apply an automated, unsupervised clustering algorithm to group events in the cleaned data subset into a plurality of clusters based on the multiparameter data; and iv) analyse statistical properties of the plurality of clusters to identify at least one cluster corresponding phenotypically to MRD cells. In some embodiments, the system is for analysing minimal residual disease (MRD) associated with paediatric B-cell precursor acute lymphoblastic leukaemia (B-ALL) in a sample from a patient using multiparameter flow cytometry (MPFC) data, the system comprising a memory and a processor, wherein the system is configured to: i) isolate a subset of data corresponding to a target cell lineage; ii) apply an automated algorithm to identify and remove data anomalies from the data subset to produce a cleaned data subset; iii) apply an automated, unsupervised clustering algorithm to group events in the cleaned data subset into a plurality of clusters based on the multiparameter data; and, iv) analyse statistical properties of the plurality of clusters to identify at least one cluster corresponding phenotypically to MRD cells. In some embodiments, following step iii), the system is further configured to generate graphical representations of the clusters. In some embodiments, the system comprises a data reduction module (DRM) configured to computationally isolate a subset of data corresponding to a target cell lineage. In some embodiments, the system comprises a data cleaning module (DCM) configured to apply an automated algorithm to identify and remove data anomalies from the data subset. In some embodiments, the system comprises a data clustering module (DClM) configured to apply an automated, unsupervised clustering algorithm to group events in the cleaned data subset into a plurality of clusters based on the multiparameter data. In some embodiments, the system comprises a Cluster Analysis Module (CAM) configured to analyze statistical properties of the plurality of clusters to identify at least one cluster corresponding phenotypically to MRD cells. In some embodiments, the system comprises: i) a data reduction module (DRM) configured to computationally isolate a subset of data corresponding to a target cell lineage; ii) a data cleaning module (DCM) configured to apply an automated algorithm to identify and remove data anomalies from the data subset; iii) a data clustering module (DClM) configured to apply an automated, unsupervised clustering algorithm to group events in the cleaned data subset into a plurality of clusters based on the multiparameter data; and iv) a Cluster Analysis Module (CAM) configured to analyze statistical properties of the plurality of clusters to identify at least one cluster corresponding phenotypically to MRD cells. In some embodiments, the CAM is further configured to generate graphical representations of the clusters. In some embodiments, the system comprises: a) an Acquisition Subsystem comprising an MPFC instrument configured to measure cellular parameters of the sample from the patient and signal processing electronics configured to convert measured signals into a digital data format; b) a Control and File Generation Subsystem operatively coupled to the Acquisition Subsystem, configured to receive the digital data format and transform said digital data into a file format; c) an Analytical Subsystem operatively coupled to receive the file format, configured to analyze the MPFC data contained therein, said Analytical Subsystem comprising: i. a Data Reduction Module (DRM) configured to computationally isolate a subset of data corresponding to a target cell lineage; ii. a Data Cleaning Module (DCM) configured to apply an automated algorithm to identify and remove data anomalies from the data subset; iii. a Data Clustering Module (DClM) configured to apply an automated, unsupervised clustering algorithm to group events in the cleaned data subset into a plurality of clusters based on the multiparameter data; and iv. a Cluster Analysis Module (CAM) configured to analyze statistical properties of the plurality of clusters to identify at least one cluster corresponding phenotypically to MRD cells; d) a file repository operatively coupled for storing data files; and e) at least one communication network operatively connecting the Control and File Generation Subsystem, the Analytical Subsystem, and the file repository and the external server. In some embodiments, the CAM is further configured to generate graphical representations of the clusters. In some embodiments, the system further comprises an external server operatively accessible for providing analysis algorithms or related resources. In some embodiments, the analytical subsystem comprises four interconnected functional modules configured to operate sequentially. In some embodiments, the Control and File Generation Subsystem and the Analytical Subsystem are implemented on one or more computing devices, each comprising at least one central processing unit (CPU), primary memory, secondary memory, a network interface, a user interface, and at least one input device. In some embodiments, the one or more computing devices further comprises at least one graphics processing unit (GPU). In some embodiments, the interconnected functional modules of the Analytical Subsystem are further characterized in that: a) the Data Reduction Module (DRM) is configured to isolate B-lineage cells to generate a reduced data file; b) the Data Cleaning Module (DCM) is configured to apply an automated algorithm to remove anomalies from the reduced data file to generate a cleaned data file; c) the Data Clustering Module (DClM) is configured to apply an automated, unsupervised clustering algorithm to the cleaned data file to generate the plurality of clusters and associated cluster statistics; and d) the Cluster Analysis Module (CAM) is configured to process the cluster statistics to identify the at least one cluster corresponding phenotypically to MRD cells based on multiparameter expression profiles and calculate a quantity of said MRD cells. In some embodiments, the DCM is configured to apply an automated algorithm selected from the group consisting of flowClean, flowAI, and PeacoQC. In further embodiments, the automated algorithm identifies and removes events likely associated with technical artifacts. In some embodiments, the events likely associated with technical artifacts are selected from the group consisting of air bubbles, cell aggregates that passed the doublet discrimination gate, transient changes in flow rate, or electronic noise. In some embodiments, the DClM is configured to apply an automated, unsupervised clustering algorithm selected from the group consisting of FlowSOM, Phenograph, X-shift, flowMeans, and flowClust. In some embodiments, the Cluster Analysis Module is ClusterExplorer. In some embodiments, utilizing algorithms within DRM, the data can be graphically displayed on the user interface. For example, a bivariate plot showing forward scatter (FSC) versus side scatter (SSC) can be generated. On this plot, the operator may apply an initial gate, to isolate the total leukocyte population while excluding non-cellular debris and platelets. The events within the leukocyte gate can then be visualized on a subsequent bivariate plot, perhaps plotting FSC height (FSC-H) versus FSC area (FSC-A), to discriminate between single cells and doublet events. A second gate selects the single cell population. These single leukocyte events are further displayed on another bivariate plot, typically plotting a scatter parameter (e.g., FSC 602) against a key lineage marker, such as the B-cell specific marker CD19. A third gate is applied to select CD19-positive cells, which encompass the B-lineage cells, including the target leukemic MRD cells in B-ALL. In some embodiments, the Data Cleaning Module (DCM) removes residual data anomalies that might have occurred during the acquisition from patient sample. In some embodiments, data anomalies are associated with technical artifacts, such as those caused by air bubbles, cell aggregates that passed the doublet discrimination gate, transient changes in flow rate, or electronic noise. In some embodiments, the Data Clustering Module (DClM) executes the core task of grouping the high-dimensional data for each of the cleaned cells into distinct clusters. In some embodiments each of the distinct clusters represents a phenotypically homogeneous cell subpopulation. In some embodiments, the CAM analyzes the defined clusters to identify the specific cluster(s) exhibiting the phenotypic characteristics of the leukemic MRD cells relevant to the patient's B- ALL diagnosis. In some embodiments, the DRM generates a FCSdr file. In some embodiments, the DCM generates a FCScd file. In some embodiments, the CAM generates a CS file. In some embodiments, the sample from a patient comprises bone marrow cells. In some embodiments, the sample from a patient comprises peripheral blood cells. In some embodiments, the sample from a patient comprises at least 4x106cells. In another aspect, the invention relates to a computer program comprising instructions which, when the program is executed by a computer, cause the computer to carry out a method as described herein. In some embodiments, the method is a computer-implemented method for analysing minimal residual disease (MRD) associated with paediatric B-cell precursor acute lymphoblastic leukaemia (B-ALL) in a sample from a patient using multiparameter flow cytometry (MPFC) data, said method comprising: i) isolating a subset of data corresponding to a target cell lineage; ii) applying an automated algorithm to identify and remove data anomalies from the data subset to produce a cleaned data subset; iii) applying an automated, unsupervised clustering algorithm to group events in the cleaned data subset into a plurality of clusters based on the multiparameter data; and iv) analysing statistical properties of the plurality of clusters to identify at least one cluster corresponding phenotypically to MRD cells. In some embodiments, the patient is a human. In some embodiments, the patient has previously been diagnosed with pediatric B-cell precursor acute lymphoblastic leukemia. In some embodiments, the patient has previously been treated for pediatric B-cell precursor acute lymphoblastic leukemia. In some embodiments, step iii) further comprises generating graphical representations of the clusters. In some embodiments, step i) is performed by a data reduction module (DRM). In some embodiments, step ii) is performed by a data cleaning module (DCM). In some embodiments, step iii) is performed by a data clustering module (DClM). In some embodiments, step iv) is performed by a cluster analysis module (CAM). In some embodiments, the computer-implemented method comprises: i) isolating a subset of data corresponding to a target cell lineage using a Data Reduction Module (DRM), ii) applying an automated algorithm to identify and remove data anomalies from the data subset to produce a cleaned data subset using a Data Cleaning Module (DCM), iii) applying an automated, unsupervised clustering algorithm to group events in the cleaned data subset into a plurality of clusters based on the multiparameter data using a Data Clustering Module (DClM); and iv) analyzing statistical properties of the plurality of clusters to identify at least one cluster corresponding phenotypically to MRD cells using a Cluster Analysis Module (CAM), wherein the DRM, the DCM, the DClM and the CAM comprise an Analytical Subsystem. In some embodiments, the computer-implemented method comprises: i) transmitting a file to the Analytical Subsystem or a file repository; ii) within the Data Reduction Module (DRM) of the Analytical Subsystem, retrieving the file and isolating a subset of data corresponding to a target cell lineage; iii) within a Data Cleaning Module (DCM) of the Analytical Subsystem, retrieving the reduced data file and applying an automated algorithm to identify and remove data anomalies from the data subset to produce a cleaned data subset; iv) within a Data Cleaning Module (DCM) of the Analytical Subsystem, retrieving the reduced data file and applying an automated, unsupervised clustering algorithm to group events in the cleaned data subset into a plurality of clusters based on the multiparameter data; and v) within a Cluster Analysis Module (CAM) of the Analytical Subsystem, retrieving the cluster statistics and analysing statistical properties of the plurality of clusters to identify at least one cluster corresponding phenotypically to MRD cells. In some embodiments, the DRM isolates a subset of data corresponding to a target cell lineage by applying a data reduction process. In further embodiments, the data reduction process comprises identifying a total leukocyte population within a reference gate and isolating a B-cell population within a subsequent gate, thereby generating a reduced data file containing data for the B-cell population and storing a count of the total leukocyte population. In some embodiments, the CAM determines a quantity of cells within the identified at least one cluster corresponding to MRD cells and calculates a percentage of MRD cells by dividing the quantity of cells determined by the DRM by the count of the total leukocyte population stored by the DRM and multiplying by 100. In some embodiments, the computer-implemented method comprises: i) transmitting a file to the Analytical Subsystem or a file repository; ii) within the Data Reduction Module (DRM) of the Analytical Subsystem, retrieving the file and applying a data reduction process, comprising identifying a total leukocyte population within a reference gate and isolating a B-cell population within a subsequent gate, thereby generating a reduced data file containing data for the B-cell population and storing a count of the total leukocyte population; iii) within a Data Cleaning Module (DCM) of the Analytical Subsystem, retrieving the reduced data file and applying an automated data cleaning algorithm to remove data anomalies, thereby generating a cleaned data file; iv) within a Data Clustering Module (DClM) of the Analytical Subsystem, retrieving the cleaned data file and applying an automated, unsupervised clustering algorithm to partition events within the cleaned data file into a plurality of clusters based on their multiparameter data; v) within the Cluster Analysis Module (CAM), analysing the cluster statistics and optionally visualizing cluster expression profiles to identify at least one specific cluster exhibiting a multiparameter phenotype characteristic of leukemic MRD cells for B- ALL; and vi) within the Cluster Analysis Module (CAM), (1) determining a quantity of cells within the identified at least one specific cluster corresponding to MRD cells; and calculating a percentage of MRD cells by dividing the quantity of cells identified in step (1) by the count of the total leukocyte population stored in step (ii) and multiplying by 100. In some embodiments, following step iv), the DClM further generates graphical representations of the plurality of clusters and generates cluster statistics characterizing each cluster and stores the cluster statistics. In some embodiments, the file transmitted to the Analytical Subsystem or the file repository is an FCS file. In some embodiments, the data reduction process performed by the DRM comprises displaying data from the file on the user interface, thereby enabling an operator to perform steps including: a) defining Gate-1 on a first bivariate plot to encompass the leukocyte population and storing the count of events within Gate-1 for subsequent MRD percentage calculation; and b) defining Gate-3 on a subsequent bivariate plot to encompass the B-cell population, wherein the data corresponding to events within Gate-3 constitutes the reduced data file utilized for subsequent clustering and identification of leukemic MRD cells. In some embodiments, the automated data cleaning algorithm applied by the DCM, in addition to generating the cleaned data file, causes the display on the user interface of at least one graphical representation comparing data characteristics before cleaning and after cleaning, optionally accompanied by corresponding event counts, thereby enabling an operator to visually assess the effectiveness of the cleaning process and determine if adjustments to preceding steps, such as gating within the DRM, are warranted. In some embodiments, the graphical representations generated by the DClM and displayed on the user interface comprise at least one of a cluster tree, a cluster map, and a heatmap, wherein: a) the cluster tree visually represents hierarchical relationships among the plurality of clusters, potentially illustrating groupings into metaclusters and indicating relative phenotypic distances between clusters; and b) the heatmap displays the plurality of clusters against the measured parameters, utilizing a color scale to represent the expression level of each parameter within each cluster, thereby providing a comprehensive visualization of the immunophenotype associated with the cell population within each respective cluster. In further embodiments, the heatmap displays the plurality of clusters as rows and / or the heatmap displays measured parameters as columns and / or the expression level is median or mean intensity. In some embodiments, the analysis performed by the CAM using the cluster statistics comprises displaying cluster information on the user interface via a graphical representation, wherein: a) the graphical representation displays axes corresponding to the measured parameters and signal intensity; b) each cluster of the plurality of clusters is represented as a profile line interconnecting points corresponding to the characteristic signal intensity value for that cluster for each respective parameter; and c) the method further comprises enabling the application of one or more analytical gates based on signal intensity thresholds for selected parameters, wherein logical combinations of said gates using operators such as AND, OR, or NOT can be employed to computationally filter or select clusters matching a predefined MRD phenotype. In some embodiments, utilizing algorithms within DRM, the data can be graphically displayed on the user interface. For example, a bivariate plot showing forward scatter (FSC) versus side scatter (SSC) can be generated. On this plot, the operator may apply an initial gate, to isolate the total leukocyte population while excluding non-cellular debris and platelets. The events within the leukocyte gate can then be visualized on a subsequent bivariate plot, perhaps plotting FSC height (FSC-H) versus FSC area (FSC-A), to discriminate between single cells and doublet events. A second gate selects the single cell population. These single leukocyte events are further displayed on another bivariate plot, typically plotting a scatter parameter (e.g., FSC 602) against a key lineage marker, such as the B-cell specific marker CD19. A third gate is applied to select CD19-positive cells, which encompass the B-lineage cells, including the target leukemic MRD cells in B-ALL. In some embodiments, the Data Cleaning Module (DCM) removes residual data anomalies that might have occurred during the acquisition from patient sample. In some embodiments, data anomalies are associated with technical artifacts, such as those caused by air bubbles, cell aggregates that passed the doublet discrimination gate, transient changes in flow rate, or electronic noise. In some embodiments, the Data Clustering Module (DClM) executes the core task of grouping the high-dimensional data for each of the cleaned cells into distinct clusters. In some embodiments each of the distinct clusters represents a phenotypically homogeneous cell subpopulation. In some embodiments, the CAM analyzes the defined clusters to identify the specific cluster(s) exhibiting the phenotypic characteristics of the leukemic MRD cells relevant to the patient's B- ALL diagnosis. In some embodiments, the DRM generates a FCSdr file. In some embodiments, the DCM generates a FCScd file. In some embodiments, the DCM is configured to apply an automated algorithm selected from the group consisting of flowClean, flowAI, and PeacoQC. In further embodiments, the automated algorithm identifies and removes events likely associated with technical artifacts. In some embodiments, the events likely associated with technical artifacts are selected from the group consisting of air bubbles, cell aggregates that passed the doublet discrimination gate, transient changes in flow rate, or electronic noise. In some embodiments, the DClM is configured to apply an automated, unsupervised clustering algorithm selected from the group consisting of FlowSOM, Phenograph, X-shift, flowMeans, and flowClust. In some embodiments, the CAM generates a CS file. In some embodiments, the Cluster Analysis Module is ClusterExplorer. In another aspect the invention relates to a computer-readable medium comprising instructions which, when executed by a computer, cause the computer to carry out a method described herein. In some embodiments, the method is a computer-implemented method for analysing minimal residual disease (MRD) associated with paediatric B-cell precursor acute lymphoblastic leukaemia (B-ALL) in a sample from a patient using multiparameter flow cytometry (MPFC) data, said method comprising: i) isolating a subset of data corresponding to a target cell lineage; ii) applying an automated algorithm to identify and remove data anomalies from the data subset to produce a cleaned data subset; iii) applying an automated, unsupervised clustering algorithm to group events in the cleaned data subset into a plurality of clusters based on the multiparameter data; and iv) analysing statistical properties of the plurality of clusters to identify at least one cluster corresponding phenotypically to MRD cells. In some embodiments, the patient is a human. In some embodiments, the patient has previously been diagnosed with pediatric B-cell precursor acute lymphoblastic leukemia. In some embodiments, the patient has previously been treated for pediatric B-cell precursor acute lymphoblastic leukemia. In some embodiments, step iii) further comprises generating graphical representations of the clusters. In some embodiments, step i) is performed by a data reduction module (DRM). In some embodiments, step ii) is performed by a data cleaning module (DCM). In some embodiments, step iii) is performed by a data clustering module (DClM). In some embodiments, step iv) is performed by a cluster analysis module (CAM). In some embodiments, the method comprises: i) isolating a subset of data corresponding to a target cell lineage using a Data Reduction Module (DRM), ii) applying an automated algorithm to identify and remove data anomalies from the data subset to produce a cleaned data subset using a Data Cleaning Module (DCM), iii) applying an automated, unsupervised clustering algorithm to group events in the cleaned data subset into a plurality of clusters based on the multiparameter data using a Data Clustering Module (DClM); and iv) analyzing statistical properties of the plurality of clusters to identify at least one cluster corresponding phenotypically to MRD cells using a Cluster Analysis Module (CAM), wherein the DRM, the DCM, the DClM and the CAM comprise an Analytical Subsystem. In some embodiments, the method comprises: i) transmitting a file to the Analytical Subsystem or a file repository; ii) within the Data Reduction Module (DRM) of the Analytical Subsystem, retrieving the file and isolating a subset of data corresponding to a target cell lineage; iii) within a Data Cleaning Module (DCM) of the Analytical Subsystem, retrieving the reduced data file and applying an automated algorithm to identify and remove data anomalies from the data subset to produce a cleaned data subset; iv) within a Data Cleaning Module (DCM) of the Analytical Subsystem, retrieving the reduced data file and applying an automated, unsupervised clustering algorithm to group events in the cleaned data subset into a plurality of clusters based on the multiparameter data; and v) within a Cluster Analysis Module (CAM) of the Analytical Subsystem, retrieving the cluster statistics and analysing statistical properties of the plurality of clusters to identify at least one cluster corresponding phenotypically to MRD cells. In some embodiments, the DRM isolates a subset of data corresponding to a target cell lineage by applying a data reduction process. In further embodiments, the data reduction process comprises identifying a total leukocyte population within a reference gate and isolating a B-cell population within a subsequent gate, thereby generating a reduced data file containing data for the B-cell population and storing a count of the total leukocyte population. In some embodiments, the CAM determines a quantity of cells within the identified at least one cluster corresponding to MRD cells and calculates a percentage of MRD cells by dividing the quantity of cells determined by the DRM by the count of the total leukocyte population stored by the DRM and multiplying by 100. In some embodiments, the method comprises: i) transmitting a file to the Analytical Subsystem or a file repository; ii) within the Data Reduction Module (DRM) of the Analytical Subsystem, retrieving the file and applying a data reduction process, comprising identifying a total leukocyte population within a reference gate and isolating a B-cell population within a subsequent gate, thereby generating a reduced data file containing data for the B-cell population and storing a count of the total leukocyte population; iii) within a Data Cleaning Module (DCM) of the Analytical Subsystem, retrieving the reduced data file and applying an automated data cleaning algorithm to remove data anomalies, thereby generating a cleaned data file; iv) within a Data Clustering Module (DClM) of the Analytical Subsystem, retrieving the cleaned data file and applying an automated, unsupervised clustering algorithm to partition events within the cleaned data file into a plurality of clusters based on their multiparameter data; v) within the Cluster Analysis Module (CAM), analysing the cluster statistics and optionally visualizing cluster expression profiles to identify at least one specific cluster exhibiting a multiparameter phenotype characteristic of leukemic MRD cells for B- ALL; and vi) within the Cluster Analysis Module (CAM), (1) determining a quantity of cells within the identified at least one specific cluster corresponding to MRD cells; and calculating a percentage of MRD cells by dividing the quantity of cells identified in step (1) by the count of the total leukocyte population stored in step (ii) and multiplying by 100. In some embodiments, following step iv), the DClM further generates graphical representations of the plurality of clusters and generates cluster statistics characterizing each cluster and stores the cluster statistics. In some embodiments, the file transmitted to the Analytical Subsystem or the file repository is an FCS file. In some embodiments, the data reduction process performed by the DRM comprises displaying data from the file on the user interface, thereby enabling an operator to perform steps including: a) defining Gate-1 on a first bivariate plot to encompass the leukocyte population and storing the count of events within Gate-1 for subsequent MRD percentage calculation; and b) defining Gate-3 on a subsequent bivariate plot to encompass the B-cell population, wherein the data corresponding to events within Gate-3 constitutes the reduced data file utilized for subsequent clustering and identification of leukemic MRD cells. In some embodiments, the automated data cleaning algorithm applied by the DCM, in addition to generating the cleaned data file, causes the display on the user interface of at least one graphical representation comparing data characteristics before cleaning and after cleaning, optionally accompanied by corresponding event counts, thereby enabling an operator to visually assess the effectiveness of the cleaning process and determine if adjustments to preceding steps, such as gating within the DRM, are warranted. In some embodiments, the graphical representations generated by the DClM and displayed on the user interface comprise at least one of a cluster tree, a cluster map, and a heatmap, wherein: a) the cluster tree visually represents hierarchical relationships among the plurality of clusters, potentially illustrating groupings into metaclusters and indicating relative phenotypic distances between clusters; and b) the heatmap displays the plurality of clusters against the measured parameters, utilizing a color scale to represent the expression level of each parameter within each cluster, thereby providing a comprehensive visualization of the immunophenotype associated with the cell population within each respective cluster. In further embodiments, the heatmap displays the plurality of clusters as rows and / or the heatmap displays measured parameters as columns and / or the expression level is median or mean intensity. In some embodiments, the analysis performed by the CAM using the cluster statistics comprises displaying cluster information on the user interface via a graphical representation, wherein: a) the graphical representation displays axes corresponding to the measured parameters and signal intensity; b) each cluster of the plurality of clusters is represented as a profile line interconnecting points corresponding to the characteristic signal intensity value for that cluster for each respective parameter; and c) the method further comprises enabling the application of one or more analytical gates based on signal intensity thresholds for selected parameters, wherein logical combinations of said gates using operators such as AND, OR, or NOT can be employed to computationally filter or select clusters matching a predefined MRD phenotype. In some embodiments, utilizing algorithms within DRM, the data can be graphically displayed on the user interface. For example, a bivariate plot showing forward scatter (FSC) versus side scatter (SSC) can be generated. On this plot, the operator may apply an initial gate, to isolate the total leukocyte population while excluding non-cellular debris and platelets. The events within the leukocyte gate can then be visualized on a subsequent bivariate plot, perhaps plotting FSC height (FSC-H) versus FSC area (FSC-A), to discriminate between single cells and doublet events. A second gate selects the single cell population. These single leukocyte events are further displayed on another bivariate plot, typically plotting a scatter parameter (e.g., FSC 602) against a key lineage marker, such as the B-cell specific marker CD19. A third gate is applied to select CD19-positive cells, which encompass the B-lineage cells, including the target leukemic MRD cells in B-ALL. In some embodiments, the Data Cleaning Module (DCM) removes residual data anomalies that might have occurred during the acquisition from patient sample. In some embodiments, data anomalies are associated with technical artifacts, such as those caused by air bubbles, cell aggregates that passed the doublet discrimination gate, transient changes in flow rate, or electronic noise. In some embodiments, the Data Clustering Module (DClM) executes the core task of grouping the high-dimensional data for each of the cleaned cells into distinct clusters. In some embodiments each of the distinct clusters represents a phenotypically homogeneous cell subpopulation. In some embodiments, the CAM analyzes the defined clusters to identify the specific cluster(s) exhibiting the phenotypic characteristics of the leukemic MRD cells relevant to the patient's B- ALL diagnosis. In some embodiments, the DRM generates a FCSdr file. In some embodiments, the DCM generates a FCScd file. In some embodiments, the DCM is configured to apply an automated algorithm selected from the group consisting of flowClean, flowAI, and PeacoQC. In further embodiments, the automated algorithm identifies and removes events likely associated with technical artifacts. In some embodiments, the events likely associated with technical artifacts are selected from the group consisting of air bubbles, cell aggregates that passed the doublet discrimination gate, transient changes in flow rate, or electronic noise. In some embodiments, the DClM is configured to apply an automated, unsupervised clustering algorithm selected from the group consisting of FlowSOM, Phenograph, X-shift, flowMeans, and flowClust. In some embodiments, the CAM generates a CS file. In some embodiments, the Cluster Analysis Module is ClusterExplorer. Definitions The term “non-leukemic B-ALL cells” as used herein, unless otherwise indicated, refers to a cell or population of cells with a phenotype present in a subject not diagnosed as having MRD associated with B-ALL or diagnosed as not having MRD associated with B-ALL. In one embodiment, non-leukemic B-ALL cells are from a subject not having B-ALL. The term “treating” or “treatment”, as used herein, unless otherwise indicated, means reversing, alleviating, inhibiting the progress of, or slowing or delaying the progression or reoccurrence of, the disorder or condition to which such term applies, or one or more symptoms of such disorder or condition. The term “treating” also refers to prophylactic or preventative measures that prevent and / or slow the development of a targeted pathologic condition or disorder. Thus, those in need of treatment include those already with the disorder; those prone to have the disorder; and those in whom the disorder is to be prevented. The term “minimal residual disease”, as used herein, unless otherwise indicated, refers to the presence of a subpopulation of leukemic cells persisting within the bone marrow following therapy or treatment. Said cells can lead to subsequent disease relapse. The term “sensitivity”, as used herein, unless otherwise indicated, refers to the ability of the invention as descried herein to correctly identify true positives e.g. to correctly identify MRD in cell populations containing a population of MRD cells. As described herein, sensitivity may be represented as a percentage. In instances wherein sensitivity is represented as a percentage value, the percentage value indicates the percentage of instances in which the invention is capable of detecting true positive cell populations e.g. the percentage of patients in which the invention is capable of correctly identifying MRD in a patient containing a population of MRD cells. As used herein, sensitivity may also be referred to by the limit of detection. The “limit of detection”, as described herein, unless otherwise stated, refers to the lowest concentration of a substance that can be reliably detected e.g. the lowest concentration of MRD cells that can be reliably detected in a cell population. In some embodiments, the systems and methods of the invention are capable of achieving a limit of detection of at least 10⁻⁴. In some embodiments, the systems and methods of the invention are capable of achieving a limit of detection of at least 10⁻⁵. In some embodiments, the systems and methods ans systems of the invention are capble of achieving a sensitivity of at least 60%, at least 70%, at least 80%, a least 90% or at least 100%. In some embodiments, the systems and methods of the invention are capable of achieving a sensitivity of at least 80%. In some embodiments, the systems and methods of the invention are capable of achieving a sensitivity of at least 90%. In some embodiments, the systems and methods of the invention are capable of achieving a sensitivity of at least 95%. In some embodiments, the systems and methods of the invention are capable of achieving a sensitivity of 100%. In some embodiments the systems and methods of the invention are capable of achieving improved sensitivity as compared to systems and methods utilizing the BD FACSDiva™ software (BD Biosciences) associated with the FACSAria™ III instrument and following a conventional manual gating strategy, employing a predefined gating template conceptually illustrated in Figure 10. In some embodiments, the systems and methods of the invention are capable of achieving a sensitivity at least 5%, 10%, 15%, 20%, 25% or 27% higher than the specificity of a system or method utilizing the BD FACSDiva™ software (BD Biosciences) associated with the FACSAria™ III instrument and following a conventional manual gating strategy, employing a predefined gating template conceptually illustrated in Figure 10. The term “specificity”, as used herein, unless otherwise indicated, refers to the ability of the invention as described herein to correctly identify true negatives e.g. to correctly not identify MRD cell populations in cell populations which do not contain a popualtion of MRD cells. As descibed herein, specificity may be represented as a percenatge. In insatnces wherein specificty is represented as a percentage value, the percentage value indicates the percentage of instances in which the invention is capable of detecting true negative cell populations e.g. the percenatge of patients thar the invention is capable of correctly identifying as MRD-negative (i.e containing no MRD cells). In some embodiments the systems and methods of the invention are capable of achieving an improved specificity as compared to systems and utilizing the BD FACSDiva™ software (BD Biosciences) associated with the FACSAria™ III instrument and following a conventional manual gating strategy, employing a predefined gating template conceptually illustrated in Figure 10. In some embodiments, the systems and methods of the invention are capable of achieving a specificity at least 5% or 10% higher than the specificity of a system or method utilizing the BD FACSDiva™ software (BD Biosciences) associated with the FACSAria™ III instrument and following a conventional manual gating strategy, employing a predefined gating template conceptually illustrated in Figure 10. In some embodiments, the systems and methods of the invention are capable of achieving a sensitivity at least 5%, 10%, 15%, or 16% higher than the specificity of a system or method utilizing the BD FACSDiva™ software (BD Biosciences) associated with the FACSAria™ III instrument and following a conventional manual gating strategy, employing a predefined gating template conceptually illustrated in Figure 10. In some embodiments, the systems and methods of the invention are capable of achieving a specificty at least 16% higher than the specificity of a system or method utilizing the BD FACSDiva™ software (BD Biosciences) associated with the FACSAria™ III instrument and following a conventional manual gating strategy, employing a predefined gating template conceptually illustrated in Figure 10. System for Measuring and Analyzing MRD in B-ALL via Multiparameter Flow Cytometry Figure 1 presents a conceptual diagram illustrating a system 100 for measuring MRD in B-ALL using MPFC, in accordance with an embodiment of the invention. System 100 may be implemented in various configurations, potentially comprising distinct subsystems. In one exemplary embodiment, system 100 comprises three primary subsystems: an Acquisition Subsystem 101, a Control and File Generation Subsystem 103, and an Analytical Subsystem 106. The system may further include a file repository 105 and potentially interact with an external server 107. Subsystems 103 and 106, along with file repository 105, may be interconnected via a communication network 112, which can include wired or wireless networks, local area networks (LANs), intranets, or similar communication infrastructure. Analytical Subsystem 106 may optionally connect to an external server 107 via a wider network, such as the internet 113. Acquisition Subsystem 101 is configured to perform the MPFC measurement of a patient sample 102. The digital data generated during this measurement process is transmitted (e.g., via a direct connection or network interface) to Control and File Generation Subsystem 103. Subsystem 103 receives the raw flow cytometry data stream and processes it to generate a standardized raw data file 104, typically in the Flow Cytometry Standard (FCS) format. The generated FCS file 104 can be stored locally within Subsystem 103, transmitted via the network 112 to the file repository 105 for archival, or forwarded directly to Analytical Subsystem 106 for processing. Analytical Subsystem 106 is responsible for executing the computational analysis of the flow cytometry data contained within the FCS file 104. In an embodiment, Subsystem 106 comprises four functional modules: a Data Reduction Module (DRM) 108, a Data Cleaning Module (DCM) 109, a Data Clustering Module (DClM) 110, and a Cluster Analysis Module (CAM) 111. Detailed exemplary implementations related to System 100 and its modules are further illustrated in Figures 2–5 and discussed subsequently. The principal components of Acquisition Subsystem 101 are depicted conceptually in Figure 2, according to an embodiment. Subsystem 101 fundamentally includes a multiparameter flow cytometer 201, which incorporates detectors and signal processing elements that effectively function as an event-to-raw-digital-data converter 202. Subsystem 101 may also include a network interface 203 for data transmission. The primary function of Subsystem 101 is to interrogate the patient sample 102, measure the specified parameters for each cell (event), convert these measurements into digital signals, and transmit the resulting raw data stream. The components of Control and File Generation Subsystem 103 are illustrated in Figure 3, consistent with an embodiment. Subsystem 103 typically controls the operation of the flow cytometer within Subsystem 101, acquires the raw digital data stream, and formats this data into the standard FCS file 104. These functions are commonly executed by a dedicated computer system, which may comprise components such as a central processing unit (CPU) 301, optionally a graphics processing unit (GPU) 302 for enhanced display capabilities, a user interface including a display monitor 303, one or more input devices 304, primary memory (e.g., RAM) 305, and secondary memory (e.g., hard drive, SSD) 306 for software and data storage. The CPU 301 executes software instructions that orchestrate the tasks performed by Subsystem 103. It processes incoming data from the flow cytometer (101) and manages data display, potentially utilizing the GPU 302, on the user interface monitor 303. The user interface 303 presents operational status information from the flow cytometer 101 and visualizations of the incoming cell measurement data. Via the user interface 303 and input device(s) 304, an operator can manage the flow cytometer's operation, including configuring instrument settings (e.g., parameter definitions, detector voltages, compensation), controlling sample flow rates, setting acquisition thresholds, and initiating or terminating data acquisition. According to embodiments of the present invention, the input device 304 may encompass various standard interaction tools, including, but not limited to, a mouse, trackball, touch interface, touchpad, joystick, voice-activated control system, or touchscreen, facilitating user interaction with the user interface 303. Upon completion of the flow cytometry measurement acquisition process within Subsystem 101, Control and File Generation Subsystem 103 transforms the acquired raw digital data stream into the standardized FCS file format, represented as FCS file 104. This file 104 is typically stored temporarily in primary memory 305 (e.g., Random Access Memory, RAM) during processing and subsequently stored more permanently in secondary memory 306. Collectively, primary memory 305 and secondary memory 306 within Subsystem 103 store the operating system, one or more software applications (including the instrument control and data acquisition software), data utilized or generated by these applications, and the output data files such as FCS file 104. Secondary memory 306 may comprise various non-volatile storage media, including, for example, a hard disk drive (HDD), a solid-state drive (SSD), removable storage devices (e.g., USB flash drives), network-attached storage, or other suitable persistent storage. Removable storage interfaces might support devices such as optical disk drives or flash memory readers. The specific software executing within Subsystem 103 often varies depending on the manufacturer of the flow cytometry instrument (201). Through a network interface (such as network interface 203 or a similar interface within Subsystem 103), the generated FCS files 104 can be transmitted either to the file repository 105 for archival or directly to Analytical Subsystem 106 for processing. File repository 105 serves as a centralized storage location for FCS files 104 and may be implemented using internal or external storage devices, network-attached storage (NAS) systems, dedicated file servers, cloud-based storage solutions, or comparable data storage infrastructure. FCS files 104 stored in the repository 105 are subsequently retrieved by Analytical Subsystem 106 when required for analysis. The components constituting Analytical Subsystem 106 are conceptually illustrated in Figure 4, according to an embodiment of the invention, alongside a brief description of their functions. Subsystem 106 is responsible for performing the analysis of the multiparameter flow cytometry data contained in FCS file 104. This analysis proceeds through a defined sequence of steps executed by distinct functional modules: the Data Reduction Module (DRM) 108, the Data Cleaning Module (DCM) 109, the Data Clustering Module (DClM) 110, and the Cluster Analysis Module (CAM) 111. The execution of these modules ultimately yields a result indicating the measured level of MRD in the context of B-ALL. These analytical functions may be implemented on a dedicated computer system, potentially comprising hardware components analogous to those in Subsystem 103, such as a CPU 401, a GPU 402 (particularly beneficial for computationally intensive tasks), a user interface with a monitor 303, input device(s) 304, primary memory 305, and secondary memory 306. The DRM 108 executes instructions designed to reduce the dataset size by selectively isolating events corresponding to a target lineage, such as B-cells in the context of B-ALL MRD. The output of this module is typically a new data file, designated herein as FCSdr file 403 (Flow Cytometry Standard, data reduced), which contains only the selected events. This file 403 is stored in primary memory 305 and / or secondary memory 306. The DCM 109 executes instructions implementing automated algorithms for cleaning the flow cytometry data, aiming to remove artifacts or anomalous events (e.g., doublets, debris, events resulting from flow instability). The output is a cleaned data file, designated herein as FCScd file 404 (Flow Cytometry Standard, cleaned data), which is stored in primary memory 305 and / or secondary memory 306. The DClM 110 executes instructions for performing automated, unsupervised clustering of the cleaned, high-dimensional flow cytometry data. This module generates output files representing the clustering results, designated herein as a Cluster Graphics (CG) file 405 (containing information for graphical representation) and a Cluster Statistics (CS) file 406 (containing quantitative information about each cluster). These files 405 and 406 are stored in primary memory 305 and / or secondary memory 306. The CAM 111 executes instructions to analyze the clusters defined by DClM 110, enabling the identification and characterization of the MRD population. This may involve presenting correlations between cluster membership and parameter expression (901), potentially generating further analysis graphics (AG) file 407 and analysis statistics (AS) file 408, which are stored in primary memory 305 and / or secondary memory 306. The computational tasks performed by these modules (DRM 108, DCM 109, DClM 110, CAM 111) are executed by the CPU 401 within Subsystem 106. CPU 401 may be a specialized processor or a general-purpose processor, potentially including a multiprocessor or multi-core configuration (not explicitly shown) capable of parallel execution to handle the significant computational demands of processing large datasets. The GPU 402, if present, is particularly advantageous for accelerating complex mathematical operations common in clustering and dimensionality reduction algorithms applied to multiparameter data, as well as for rendering complex two- or three-dimensional graphical representations of the data and results. A GPU 402 suitable for these tasks should possess sufficient computational performance and dedicated video memory capacity. The roles and characteristics of primary memory 305, secondary memory 306, user interface 303, and input device 304 within Subsystem 106 are analogous to their counterparts described previously for Subsystem 103. For data exchange, Analytical Subsystem 106 is typically connected to Control and File Generation Subsystem 103 and / or the file repository 105 via the network 112. Additionally, Subsystem 106 may possess connectivity to external resources, such as an external server 107 via the internet 113, for purposes such as downloading updated analysis algorithms, accessing public bioinformatics repositories (e.g., the Bioconductor Project mentioned later), or other external interactions. Method for Measuring and Analyzing MRD in B-ALL via Multiparameter Flow Cytometry The method for measuring and analyzing MRD in B-ALL using MPFC, according to an embodiment of the invention, is schematically depicted as a block diagram in Figure 5. The overall method generally comprises the following principal steps: a) Preparation of the patient sample 102; b) Measurement of the prepared sample using MPFC (Acquisition); c) Initial data processing, potentially including lineage gating, and generation of a standard FCS file 104 (or a reduced FCSdr file 403); and d) Detailed analysis of the acquired data through four sequential sub-steps corresponding to the modules within Analytical Subsystem 106. In the first step (a), a biological sample 102 is obtained from the patient, typically peripheral blood or bone marrow aspirate. The cells within sample 102 are processed and treated with a panel of reagents, primarily comprising antibodies specific for various cellular markers relevant to distinguishing leukemic B-cell precursors from normal hematopoietic cells. These antibodies are conjugated to distinct fluorochromes, enabling their detection via fluorescence emission upon appropriate laser excitation. In the second step (b), the prepared sample 102 is introduced into the multiparameter flow cytometer 201 within Acquisition Subsystem 101. Within the flow cytometer, cells are hydrodynamically focused to pass individually through one or more focused laser beams operating at specific wavelengths. Laser excitation induces fluorescence emission from the fluorochromes bound to the cells. This emitted light, along with scattered laser light (FSC, SSC), is collected and optically separated (e.g., by dichroic mirrors and bandpass filters) into distinct wavelength bands corresponding to each measured parameter. The light intensity within each band is detected by sensitive photodetectors (e.g., photomultiplier tubes or avalanche photodiodes). Modern commercial flow cytometers 201 typically measure between 12 and 50 distinct parameters simultaneously. The analog signal generated by each detector for each passing cell (event) is converted into a digital value by signal processing electronics, effectively functioning as the digital converter 202. This process assigns a set of quantitative parameter values (representing fluorescence intensity or light scatter characteristics) to each individual cell event. The resulting stream of digital data is transmitted from Subsystem 101 to Subsystem 103, potentially via the network interface 203. In the third step (c), the digital data stream originating from the flow cytometer 101 is received by Control and File Generation Subsystem 103. The CPU 301 within Subsystem 103 processes this incoming data. The data may be visualized in real-time or near real-time on the user interface 303 (potentially utilizing GPU 302 for rendering) in various graphical formats, such as histograms, bivariate plots (dot plots), contour plots, or density plots. Statistical information may also be displayed. Using the input device 304, an operator can interact with these visualizations, for example, by defining electronic gates (e.g., gate 604 in Figure 6) on histograms or bivariate plots to select specific cell populations of interest. In the context of B-ALL MRD, this step might involve gating to isolate the total B-cell population or exclude debris and unwanted cell types. The data corresponding to the selected cell population(s) is then compiled and saved in the standard FCS file format as file 104 (or potentially directly as a reduced file like FCSdr 403 if initial gating is performed here). This file is stored in primary memory 305 and / or secondary memory 306. Subsequently, the FCS file (104 or 403) can be transmitted via network 112 to the file repository 105 or directly to Analytical Subsystem 106 for further analysis. In the fourth step (d), the detailed analysis commences when Analytical Subsystem 106 retrieves the relevant FCS file (104 or 403) either from the file repository 105 or directly from Subsystem 103 (e.g., via network interface 203 or equivalent). This analytical phase comprises four sequential sub-steps, executed by the interconnected functional modules: DRM 108, DCM 109, DClM 110, and CAM 111. In the first analytical sub-step, executed by the DRM 108, the input FCS file (104) may be processed. If initial broad gating was not performed in Subsystem 103, DRM 108 can facilitate this. Data is typically displayed graphically on the user interface 303, for instance, as bivariate plots as exemplified in Figure 6. The functionality of DRM 108 might leverage standard algorithms found in conventional flow cytometry analysis software. Embodiments allow for interactive analysis, enabling the operator to select graph types, define axes parameters, and apply gates (604) via the user interface 303. A key function here, if not done previously, is the selection of the B-cell population (e.g., population 614) to separate its associated events from other events (606), thereby significantly reducing the volume of data processed in subsequent steps. The resulting subset of data, containing primarily B-lineage cells, is saved, potentially in primary memory 305 and / or secondary memory 306, as the FCSdr file 403 for input into the next module. The second analytical sub-step, performed by the DCM 109, focuses on cleaning the data within the FCSdr file 403 to remove technical artifacts. Such anomalies (701) can arise during sample acquisition due to factors like air bubbles entering the flow stream, cell aggregates passing through the laser, transient changes in sample flow rate, or electronic noise. DCM 109 retrieves the FCSdr file 403 from memory (305, 306) within Subsystem 106 and applies one or more automated data cleaning algorithms. Examples of such algorithms include, but are not limited to, flowClean, flowAI, or PeacoQC, many of which are publicly available, for instance, through resources like the Bioconductor Project, potentially accessible via an external server 107. The output of this cleaning process is the FCScd file 404, containing the data presumed to be free from major technical artifacts. This cleaned file 404 is saved in primary memory 305 and / or secondary memory 306 of Subsystem 106, and the method proceeds to the third analytical sub- step. The third analytical sub-step involves grouping the multidimensional flow cytometry data points (events) within the cleaned FCScd file 404 into distinct clusters, where each cluster ideally corresponds to a biologically meaningful cell subpopulation. This task is performed by the DClM 110. DClM 110 retrieves the FCScd file 404 from memory (305, 306) and applies one or more unsupervised clustering algorithms. Suitable algorithms include, but are not limited to, FlowSOM, Phenograph, X-shift, flowMeans, or flowClust. Many such algorithms are also available through resources like the Bioconductor Project (potentially accessed via external server 107). The DClM 110 analyzes the data within FCScd 404 and automatically partitions the events into multiple clusters without requiring manual gating or predefined templates. Each resulting cluster represents a population of cells exhibiting similar multiparameter characteristics. The number of clusters generated can be substantial; for example, algorithms like FlowSOM may initially define over 200 clusters depending on parameter settings. Optionally, DClM 110 may subsequently apply dimensionality reduction techniques to facilitate visualization of the high-dimensional cluster structure in a lower-dimensional space (typically two dimensions). Algorithms such as t-distributed Stochastic Neighbor Embedding (t-SNE), UMAP, or methods integrated within clustering tools like FlowSOM, or others like PHATE, can be employed for this purpose (potentially sourced from Bioconductor via server 107). The resulting two-dimensional graphical representations, such as a cluster tree (801), cluster map (802, e.g., a UMAP or t-SNE plot), or heatmap (803) (as exemplified in Figure 8), can be saved as the CG file 405 in secondary memory 306. Concurrently, detailed statistical information characterizing each identified cluster (e.g., cell count, mean / median parameter intensities per cluster) is generated and saved as the CS file 406 in primary memory 305 and / or secondary memory 306. Following the completion of clustering and optional visualization, the method proceeds to the fourth and final analytical sub-step. The fourth analytical sub-step, executed by the CAM 111, involves the analysis and interpretation of the clusters generated in the preceding step to specifically identify the cluster(s) corresponding to the target MRD cells in B-ALL and to quantify their abundance (e.g., count and percentage relative to total relevant cells). CAM 111 retrieves the CS file 406 (containing cluster statistics) from memory (305, 306) and may utilize cluster analysis tools or algorithms (e.g., potentially components related to tools like ClusterExplorer, possibly available via Bioconductor from server 107). Leveraging the statistical data from CS file 406, CAM 111 can present the clusters and their characteristics on the user interface 303, for example, in a specialized graphical view 900 (Figure 9). This view might allow interactive exploration where individual clusters or groups of clusters can be selected. Gates (905) might be applied based on marker expression profiles (904) across clusters. Logical combinations (AND, OR, NOT operators) of marker criteria (905) across multiple markers (904) can be used to precisely define and isolate the cluster(s) exhibiting the specific immunophenotype characteristic of the MRD population in the context of B-ALL for that patient sample. The selected MRD cluster(s) can be highlighted on related visualizations, such as the cluster tree 801, and quantitative results, like the percentage contribution of each cluster (906), can be displayed, for instance, in a bar graph. Final statistical data pertaining to the identified MRD cluster(s), derived from the CS file 406, is typically presented in a tabular format (e.g., a spreadsheet view), clearly indicating the absolute count and percentage of the leukemic MRD cells relative to the appropriate denominator population (e.g., total leukocytes or total B-cells). This determination concludes the MRD measurement process according to this embodiment of the method. Advantages of the Invention Compared to known technical solutions, the system and method of the present invention offer significant advantages, including: a) Independence from Reference Samples and Supervised Machine Learning: The core analytical process does not require pre-defined 'normal' reference samples for comparison, nor does it rely on supervised machine learning models trained on specific datasets, thereby enhancing objectivity and reducing dependency on curated training data. b) Broad Compatibility: The system and method are designed for compatibility with data acquired using various antibody panels targeting relevant leukocyte markers and generated by diverse multiparameter flow cytometry instruments. c) Enhanced Applicability and Standardization: Resulting from the independence and compatibility noted above, the invention facilitates broader implementation across different clinical or research laboratories and holds potential for greater standardization of high-sensitivity MRD analysis. d) Objective Automated Clustering: Utilization of fully automated, unsupervised clustering algorithms for cell population identification minimizes operator-dependent subjectivity inherent in manual gating or supervised classification approaches. e) High-Dimensional Data Analysis Capability: Demonstrated capability to effectively analyze high-parameter datasets, such as those derived from 16-parameter flow cytometry, addressing the increasing dimensionality and complexity of contemporary MRD assays. f) Superior Sensitivity: Achievement of enhanced sensitivity for MRD detection in B-ALL, demonstrated down to a level of 10⁻⁵ (0.001%), representing a significant improvement in detection limits compared to many conventional methods and meeting or exceeding current clinical requirements.

[0002] EXAMPLES Exemplary embodiments of the system and method for measuring and analyzing Minimal Residual Disease (MRD) in B-cell acute lymphoblastic leukemia (B-ALL) utilizing multiparameter flow cytometry (MPFC) are further detailed with reference to Figures 1–10. The invention is illustrated by the following non-limiting examples. The following examples are offered by way of illustration, and not by way of limitation. Example 1: Embodiment for Measuring and Analyzing MRD in B-ALL Using 16-Parameter Flow Cytometry The embodiment described in Example 1 commences with the preparation of a patient sample 102 diagnosed with B-ALL, intended for MRD assessment using 16-parameter flow cytometry. An adequate volume of peripheral blood or, more commonly, bone marrow aspirate is collected from the patient. The sample 102 is processed following established protocols, typically adhering to manufacturer recommendations, employing a panel of antibodies directed against specific human leukocyte markers. These antibodies are conjugated with distinct fluorochromes to enable their detection. In this specific embodiment, an antibody panel with the composition detailed in Table 1 is utilized. Following sample preparation and staining, the sample 102 is acquired using Acquisition Subsystem 101, which comprises, in this example, a 16-parameter capable flow cytometer 201. The instrument configuration and acquisition parameters are managed via Control and File Generation Subsystem 103, typically utilizing software provided by the instrument manufacturer. A sufficient number of events, clinically recommended to be at least 4 × 10⁶ total events for high-sensitivity MRD detection, are acquired from sample 102. Data is collected for 14 fluorescence parameters (corresponding to the markers in Table 1) along with two physical light scatter parameters: forward scatter (FSC) and side scatter (SSC). In this specific illustrative run, a total of 4,573,827 events (cells) were acquired. For each event and each parameter, the detected signal is digitized by the converter 202 integrated within the flow cytometer, and the resulting digital data stream is transmitted to Subsystem 103, potentially via network interface 203. TABLE 1: Composition of an Exemplary Antibody Panel Against Human Leukocytes for Measuring MRD in B-ALL Using 16-Parameter Flow Cytometry The data received from Subsystem 101 is processed by the central processing unit (CPU) 301 within Subsystem 103. This data can be visualized on the user interface 303 (rendered potentially by the graphics processing unit (GPU) 302) as histograms and bivariate plots for quality control and potential preliminary gating. Using the input device 304, the operator can monitor and potentially adjust flow cytometer 201 settings, define acquisition thresholds, and, if applicable, apply initial gates to the data displayed on histograms and bivariate plots. The acquired data, potentially encompassing all events or a preliminarily gated subset, is saved in the standard FCS file format (104) within the primary memory 305 and / or secondary memory 306 of Subsystem 103. This FCS file 104 may also be transmitted via the local network 112 to the file repository 105 for archival. The detailed data analysis is performed by Analytical Subsystem 106. Subsystem 106 retrieves the relevant FCS file 104 (or a preliminarily gated version) via the local network 112 from either Subsystem 103 or the file repository 105. This file is then provided as input to the Data Reduction Module (DRM) 108. Utilizing algorithms within the DRM 108, the data can be graphically displayed on the user interface 303. For instance, a bivariate plot 601 (Figure 6) showing forward scatter (FSC) 602 versus side scatter (SSC) 603 can be generated. On this plot, the operator may apply an initial gate, designated Gate-1501, to isolate the total leukocyte population 605 while excluding non-cellular debris and platelets 606. In this specific example, the leukocyte population 605 within Gate-1 comprised 3,963,892 cells; this value serves as the denominator for subsequent MRD percentage calculation. The events within the leukocyte gate 501 can then be visualized on a subsequent bivariate plot 607, perhaps plotting FSC height (FSC-H) 608 versus FSC area (FSC-A) 609, to discriminate between single cells 610 and doublet events 611. A second gate, Gate-2502, applied to plot 607, selects the single cell population 610. These single leukocyte events are further displayed on another bivariate plot 612, typically plotting a scatter parameter (e.g., FSC 602) against a key lineage marker, such as the B-cell specific marker CD19 (613). On plot 612, a third gate, Gate-3503, is applied to select CD19-positive cells, which encompass the B-lineage cells, including the target leukemic MRD cells in B-ALL. In this example, Gate-3 contained 537,374 cells, representing 13.6% of the initial leukocyte population (605). This gating strategy achieves an 86.4% reduction in the number of events requiring further complex analysis, thereby significantly reducing the computational burden for the subsequent clustering step performed by the Data Clustering Module (DClM) 110. The data corresponding to the events selected by Gate- 3503 (i.e., single CD19+ B-lineage cells) is saved as the reduced data file, FCSdr file 403, within the primary memory 305 and secondary memory 306 of Subsystem 106. In the subsequent stage of analysis, the Data Cleaning Module (DCM) 109 within Subsystem 106 processes the FCSdr file 403 to remove residual data anomalies 701 that might have occurred during the acquisition from patient sample 102. The DCM 109 retrieves the data from FCSdr file 403 and applies an automated flow cytometry data cleaning algorithm. In this embodiment, the flowClean algorithm is employed as an example, although other algorithms (e.g., flowAI, PeacoQC) could be utilized. This algorithm identifies and removes events likely associated with technical artifacts 701, such as those caused by air bubbles, cell aggregates that passed the doublet discrimination gate, transient changes in flow rate, or electronic noise. For quality control, the DCM 109 may display diagnostic graphs on the user interface 303. For instance, bivariate plots showing measurement time 702 on the X-axis versus a scatter parameter (e.g., SSC 603) on the Y-axis can visualize flow stability. Raw data might be shown in plot 507, while the cleaned data is shown in plot 508 (Figure 7). Corresponding event counts 509 for plots 507 (before cleaning) and 508 (after cleaning) can be displayed. In this example, counts were 537,374 and 501,433 events / cells, respectively, indicating that 35,941 events / cells (approximately 6.7%) were identified as anomalous and removed. The cleaning process can potentially be applied iteratively if deemed necessary. The resulting cleaned event data is saved as the FCScd file 404 in the primary memory 305 and secondary memory 306 of Subsystem 106. Following data cleaning, the Data Clustering Module (DClM) 110 executes the core task of grouping the high-dimensional (16-parameter in this example) data for each of the 501,433 cleaned cells into distinct clusters, each ideally representing a phenotypically homogeneous cell subpopulation. The DClM 110 retrieves the FCScd file 404 from memory (305 or 306) within Subsystem 106. It applies an automated, unsupervised clustering algorithm. In this embodiment, the FlowSOM algorithm is used as an illustrative example; however, other suitable automated clustering algorithms (e.g., Phenograph, X-shift) could alternatively be employed. The DClM 110 processes the data and, optionally, utilizes dimensionality reduction techniques to visualize the resulting clusters derived from the multidimensional space as two-dimensional graphical representations on the user interface 303. In this specific example run using FlowSOM, 100 individual clusters (nodes) and 8 aggregated metaclusters were identified. These results can be visualized, for instance, as a cluster tree 801 (Figure 8), which illustrates the hierarchical relationships and phenotypic similarity between the individual clusters. Alternative graphical representations may include a cluster map 802 (e.g., a self-organizing map grid or a t-SNE / UMAP projection colored by cluster) and a heatmap 803. In the heatmap 803, rows typically correspond to identified clusters, columns correspond to the measured parameters (markers), and the color intensity within each cell represents the median or mean expression level of that marker within that specific cluster. The graphical outputs visualizing the cluster structure (e.g., tree 801, map 802, heatmap 803) are saved, potentially within the CG file 405, in the secondary memory 306 of Subsystem 106. Simultaneously, detailed statistical information characterizing each cluster (e.g., cell count, parameter expression statistics) is compiled and saved as the CS file 406 in primary memory 305 and / or secondary memory 306. In the final analytical stage, the Cluster Analysis Module (CAM) 111 analyzes the defined clusters to identify the specific cluster(s) exhibiting the phenotypic characteristics of the leukemic MRD cells relevant to the patient's B-ALL diagnosis. The CAM 111 retrieves the CS file 406 containing the cluster statistics from memory (305 or 306) within Subsystem 106. In this embodiment, the CAM 111 processes the statistical data from CS file 406, potentially utilizing algorithms associated with tools like ClusterExplorer. The module presents the cluster information on the user interface, for example, via a specialized graph 901 (Figure 9). In graph 901, the X-axis 902 might list all measured parameters (fluorescence and scatter channels, potentially annotated with corresponding leukocyte markers), while the Y- axis 903 represents signal intensity (e.g., fluorescence intensity on a logarithmic or transformed scale). Each identified cluster can be represented as a line profile 904, connecting the median or mean expression level for that cluster across all parameters listed on the X-axis 902. Using interactive tools or automated criteria based on known MRD phenotypes, specific clusters can be investigated. For instance, gates (905 in Figure 9 concept) might be applied to filter clusters based on expression levels of key markers. In this example, by examining the expression profiles, Cluster No.7 (904) was identified as exhibiting the phenotypic characteristics consistent with the patient's known leukemic MRD population in B-ALL. This identification can be cross-referenced with other visualizations; for example, Cluster No.7 might appear as a distinct, potentially small, population 804 on the cluster tree 801, phenotypically separated from clusters representing normal B-cell precursors or mature B-cells. According to the statistical data retrieved from the CS file 406, Cluster No.7 contained 503 cells. The percentage of MRD is then calculated using the cell count from the identified MRD cluster (Cluster No.7) and the total relevant cell count established earlier (leukocytes in Gate-1605): % MRD cells = (Number of cells in the identified MRD cluster / Number of cells in the reference gate (e.g., Gate-1605)) × 100 Applying the values from this example: % MRD cells = (503 / 3,963,892) × 100 ≈ 0.0127% Rounding appropriately, the analysis concludes that the patient’s bone marrow sample contains approximately 0.013% leukemic MRD cells relative to the total leukocyte population, confirming a positive MRD result for B-ALL at this level. Example 2: Determination of the Limit of Detection (LOD) of MRD Using the System and Method with 16-Parameter Flow Cytometry To experimentally determine the Limit of Detection (LOD) for MRD cells using the described system 100 and method, a dilution series was prepared. Five samples were created containing known, progressively lower concentrations of leukemic cells: approximately 1%, 0.1%, 0.01%, 0.001%, and 0.0001%. This was achieved by spiking varying numbers of known leukemic cells into a constant background matrix of bone marrow mononuclear cells confirmed to be devoid of leukemic cells. The source of leukemic cells was a diagnostic bone marrow sample from a patient with B-cell precursor acute lymphoblastic leukemia (BCP-ALL), which contained approximately 80% leukemic blasts. The background matrix consisted of pooled bone marrow mononuclear cells obtained from pediatric patients diagnosed with Ewing sarcoma, where the absence of leukemic cells had been previously confirmed. The following procedure was implemented: Five standard flow cytometry tubes were prepared. Into each tube, approximately 4.5 × 10⁶ non- leukemic bone marrow background cells were aliquoted. Subsequently, specific quantities of the BCP-ALL leukemic cells were added to each tube to achieve the target dilutions: ^ Tube No.1: 45,000 leukemic cells added → Target concentration ≈ 1% (10⁻²) ^ Tube No.2: 4,500 leukemic cells added → Target concentration ≈ 0.1% (10⁻³) ^ Tube No.3: 450 leukemic cells added → Target concentration ≈ 0.01% (10⁻⁴) ^ Tube No.4: 45 leukemic cells added → Target concentration ≈ 0.001% (10⁻⁵) ^ Tube No.5: 5 leukemic cells added → Target concentration ≈ 0.0001% (10⁻⁶) (Note: Precise target percentage depends slightly on the exact final volume / total cell count) The cells in each of the five prepared samples were then stained using the same 16-parameter flow cytometry antibody panel detailed in Table 1. Following staining, each sample was processed and analyzed using the system 100 and the analytical method described in Example 1. The results obtained from the analysis of these five samples are summarized in Table 2. As indicated by the results (and implied qualitative description in the original text), in Tubes No. 1 through No.4, the analysis successfully identified and quantified a population of leukemic cells consistent with the spiked amount. The measured percentage of detected leukemic cells closely corresponded to the target concentrations introduced into each tube down to the 0.001% level. In Tube No.5, containing only approximately 5 target leukemic cells in the background matrix, the analysis did not reliably detect a distinct leukemic cell cluster. TABLE 2: Results of MRD Detection Sensitivity Experiment Based on these experimental findings, it can be concluded that the limit of detection (LOD) or sensitivity of MRD measurement achievable with the described system 100 and method, utilizing the specified 16-parameter panel and analytical workflow, is approximately 0.001% (corresponding to a frequency of 1 leukemic cell in 100,000 total cells, or 10⁻⁵). This represents a significant improvement, potentially an order of magnitude higher sensitivity, compared to the commonly accepted LOD of 0.01% (10⁻⁴) often associated with less complex panels or traditional analysis methods. Example 3: Comparative Determination of Sensitivity and Specificity of the System and Method for Measuring and Analyzing MRD in B-ALL via 16-Parameter Flow Cytometry The objective of this example was to compare the performance of Minimal Residual Disease (MRD) assessment in B-cell acute lymphoblastic leukemia (B-ALL) using multiparameter flow cytometry data analyzed via a conventional gating method versus analysis using System 100 of the present invention. Bone marrow aspirate samples were prospectively collected from 77 pediatric patients diagnosed with B-ALL who were clinically determined to be in remission. These samples were processed in parallel using the 16-color flow cytometry antibody panel detailed previously in Table 1. Sample acquisition was performed using a BD FACSAria™ III flow cytometer (BD Biosciences). The resulting Flow Cytometry Standard (FCS) data files from all 77 samples were subsequently analyzed by a single experienced operator employing two distinct methodologies: 1. Classical Analysis Method: Utilizing the BD FACSDiva™ software (BD Biosciences) associated with the FACSAria™ III instrument. This analysis followed a conventional manual gating strategy, employing a predefined gating template conceptually illustrated in Figure 10 (representing the complexity rather than a literal template depiction). This approach involved the sequential visual inspection and manual delineation of gates on a series of bivariate plots (potentially requiring analysis of a minimum of 45 and up to 60 or more plots per sample) to progressively isolate and quantify potential MRD populations. 2. System 100 Analysis Method: Utilizing the system and method of the present invention as detailed in Example 1. This involved processing the same FCS files through the automated workflow comprising data reduction (DRM 108), data cleaning (DCM 109), data clustering (DClM 110), and cluster analysis (CAM 111). The MRD quantification results obtained from both analytical methods for all 77 samples were statistically compared to determine diagnostic sensitivity and specificity relative to established clinical outcomes or reference standards (details of the reference standard are omitted here but assumed for the ROC analysis). Receiver Operating Characteristic (ROC) curve analysis was performed using the dedicated module within the MedCalc® statistical software package (MedCalc Software Ltd; cf. https: / / www.medcalc.org / manual / roccurves.php). The comparative outcomes of the ROC analysis are conceptually illustrated in Figure 10. Analysis using the classical BD FACSDiva™ method (1001) yielded a diagnostic sensitivity of 72.2% and a specificity of 66.7% (p = 0.001, indicating statistical significance relative to random chance). In contrast, analysis of the same dataset using System 100 (1002) of the present invention demonstrated significantly improved performance, achieving a sensitivity of 100% and a specificity of 83.3% (p < 0.001). These results provide quantitative evidence supporting the superior technical performance and diagnostic accuracy of System 100 (1002) compared to the conventional manual gating approach (1001) for the assessment of MRD in pediatric B-ALL using high-parameter flow cytometry data. Scope of the Invention The subject matter disclosed herein may be embodied in various forms and should not be construed as being limited solely to the specific embodiments described. These embodiments are presented to ensure that the disclosure is thorough and complete, thereby fully conveying the scope and nature of the invention to those skilled in the relevant art. The principles and features described may be applied to other variations and applications without departing from the spirit and scope of the invention. The invention is not to be limited in scope by the specific embodiments described herein. Indeed, various modifications of the invention in addition to those described will become apparent to those skilled in the art from the foregoing description and accompanying figures. Such modifications are intended to fall within the scope of the appended claims. All references (e.g., publications or patents or patent applications) cited herein are incorporated herein by reference in their entirety and for all purposes to the same extent as if each individual reference (e.g., publication or patent or patent application) was specifically and individually indicated to be incorporated by reference in its entirety for all purposes. Other embodiments are within the following claims. REFERENCES Patent references 1. Guldberg, S.M., et al., Computational Methods for Single-Cell Proteomics. Annu Rev Biomed Data Sci, 2023.6: p.47-71. 2. Montante, S. and R.R. Brinkman, Flow cytometry data analysis: Recent tools and algorithms. Int J Lab Hematol, 2019.41 Suppl 1: p.56-62. 3. EP2347352B1, Interactive tree plot for flow cytometry data. 4. US 2003078703 А1 Cytometry analysis system and method using database-driven network of cytometers.2003. 5. US 5605805 A2 Automatic lineage assignment of acute leukemias by flow cytometry 1997. 6. EP1785899B1, Methods for identifying discrete populations (e.g. clusters) of data within a flow cytometer multi-dimensional data set. 7. US8214157B2, Method and apparatus for representing multidimensional data. 8. WO2006089190A2, System, method, and article for detecting abnormal cells using multi- dimensional analysis. 9. CN103942415A, Automatic data analysis method of flow cytometer Non patent references 10. Jafari, K., et al., Visualization of Cell Composition and Maturation in the Bone Marrow Using 10-Color Flow Cytometry and Radar Plots. Cytometry B Clin Cytom, 2018.94(2): p.219- 229. 11. Karai, B., et al., A Novel Method for the Evaluation of Bone Marrow Samples from Patients with Pediatric B-Cell Acute Lymphoblastic Leukemia-Multidimensional Flow Cytometry. Cancers (Basel), 2021.13(20). 12. Shopsowitz, K., et al., MAGIC-DR: An interpretable machine-learning guided approach for acute myeloid leukemia measurable residual disease analysis. Cytometry B Clin Cytom, 2024. 13. Shopsowitz, K.E., et al., Machine learning optimized multiparameter radar plots for B-cell acute lymphoblastic leukemia minimal residual disease analysis. Cytometry B Clin Cytom, 2022. 102(5): p.342-352. 14. Reiter, M., et al., Automated Flow Cytometric MRD Assessment in Childhood Acute B- Lymphoblastic Leukemia Using Supervised Machine Learning. Cytometry Part A, 2019.95(9): p. 966-975. Exemplary clauses Clause 1: A system (100) for measuring and analyzing minimal residual disease (MRD) in pediatric B-cell precursor acute lymphoblastic leukemia (B-ALL) using multiparameter flow cytometry (MPFC) data, the system comprising: a) an Acquisition Subsystem (101) comprising an MPFC instrument (201) configured to measure cellular parameters of a patient sample (102) and signal processing electronics (202) configured to convert measured signals into a digital data format; b) a Control and File Generation Subsystem (103) operatively coupled to the Acquisition Subsystem (101), configured to receive the digital data format and transform said digital data into a Flow Cytometry Standard (FCS) file format (104); c) an Analytical Subsystem (106) operatively coupled to receive the FCS file format (104), configured to analyze the MPFC data contained therein, said Analytical Subsystem (106) comprising four interconnected functional modules configured to operate sequentially: i. a Data Reduction Module (DRM) (108) configured to computationally isolate a subset of data corresponding to a target cell lineage; ii. a Data Cleaning Module (DCM) (109) configured to apply an automated algorithm to identify and remove data anomalies (701) from the data subset; iii. a Data Clustering Module (DClM) (110) configured to apply an automated, unsupervised clustering algorithm to group events in the cleaned data subset into a plurality of clusters based on the multiparameter data, and optionally generate graphical representations (801, 802, 803) of the clusters; and iv. a Cluster Analysis Module (CAM) (111) configured to analyze statistical properties of the plurality of clusters to identify at least one cluster corresponding phenotypically to MRD cells; d) a file repository (105) operatively coupled for storing data files, including the FCS file format (104); e) optionally, an external server (107) operatively accessible for providing analysis algorithms or related resources; and f) at least one communication network (112, 113) operatively connecting the Control and File Generation Subsystem (103), the Analytical Subsystem (106), and the file repository (105) and the external server (107). Clause 2: The system (100) of clause 1, wherein the Control and File Generation Subsystem (103) and the Analytical Subsystem (106) are implemented on one or more computing devices, each comprising at least one central processing unit (CPU) (301, 401), optionally at least one graphics processing unit (GPU) (302, 402), primary memory (305), secondary memory (306), a network interface (203 or equivalent), a user interface (303), and at least one input device (304). Clause 3: The system (100) of clause 1, wherein the interconnected functional modules of the Analytical Subsystem (106) are further characterized in that: a) the Data Reduction Module (DRM) (108) is configured to isolate B-lineage cells (614) from the FCS file format (104) to generate a reduced data file (403); b) the Data Cleaning Module (DCM) (109) is configured to apply an automated algorithm selected from the group consisting of flowClean, flowAI, and PeacoQC to remove anomalies (701) from the reduced data file (403) to generate a cleaned data file (404); c) the Data Clustering Module (DClM) (110) is configured to apply an automated, unsupervised clustering algorithm selected from the group consisting of FlowSOM, Phenograph, X-shift, flowMeans, and flowClust to the cleaned data file (404) to generate the plurality of clusters and associated cluster statistics (406); and d) the Cluster Analysis Module (CAM) (111) is configured to process the cluster statistics (406) to identify the at least one cluster (904) corresponding phenotypically to MRD cells based on multiparameter expression profiles and calculate a quantity of said MRD cells. Clause 4: A method for measuring and analyzing minimal residual disease (MRD) in pediatric B-cell precursor acute lymphoblastic leukemia (B-ALL) using multiparameter flow cytometry (MPFC) data via an analysis system, the method comprising the steps of: a) treating a patient sample (102) with a panel of fluorochrome-conjugated antibodies specific for a plurality of cellular markers relevant to distinguishing leukemic B-ALL cells from normal cells; b) measuring the treated patient sample (102) using an MPFC instrument (201) to generate measured signals for each of a plurality of events across multiple parameters, and converting said measured signals into a digital data format (202); c) transmitting the digital data format to a Control and File Generation Subsystem (103); d) within the Control and File Generation Subsystem (103), processing the received digital data format, optionally displaying graphical representations of the data on a user interface (303), and transforming the digital data format into a Flow Cytometry Standard (FCS) file (104) stored in memory (305, 306); e) transmitting the FCS file (104) to an Analytical Subsystem (106) or a file repository (105); f) within a Data Reduction Module (DRM) (108) of the Analytical Subsystem (106), retrieving the FCS file (104) and applying a data reduction process, comprising identifying a total leukocyte population (605) within a reference gate (501) and isolating a B-cell population (614) within a subsequent gate (503), thereby generating a reduced data file (FCSdr 403) containing data for the B-cell population (614) and storing a count of the total leukocyte population (605); g) within a Data Cleaning Module (DCM) (109) of the Analytical Subsystem (106), retrieving the reduced data file (FCSdr 403) and applying an automated data cleaning algorithm to remove data anomalies (701), thereby generating a cleaned data file (FCScd 404); h) within a Data Clustering Module (DClM) (110) of the Analytical Subsystem (106), retrieving the cleaned data file (FCScd 404) and applying an automated, unsupervised clustering algorithm to partition events within the cleaned data file (404) into a plurality of clusters based on their multiparameter data; i) within the Data Clustering Module (DClM) (110), optionally generating graphical representations (801, 802, 803) of the plurality of clusters and generating cluster statistics characterizing each cluster, storing the cluster statistics as a CS file (406); j) within a Cluster Analysis Module (CAM) (111) of the Analytical Subsystem (106), retrieving the cluster statistics (CS file 406); k) within the Cluster Analysis Module (CAM) (111), analyzing the cluster statistics (406) and optionally visualizing cluster expression profiles (901) to identify at least one specific cluster (904) exhibiting a multiparameter phenotype characteristic of leukemic MRD cells for B-ALL; l) within the Cluster Analysis Module (CAM) (111), determining a quantity of cells within the identified at least one specific cluster (904) corresponding to MRD cells; and m) calculating a percentage of MRD cells by dividing the quantity of cells determined in step (l) by the count of the total leukocyte population (605) stored in step (f) and multiplying by 100. Clause 5: The method of clause 4, wherein the data reduction process performed by the DRM (108) comprises displaying data from the FCS file (104) on the user interface (303), thereby enabling an operator to perform steps including: a) defining Gate-1 (501) on a first bivariate plot (601) to encompass the leukocyte population (605) and storing the count of events within Gate-1 (501) for subsequent MRD percentage calculation; and b) defining Gate-3 (503) on a subsequent bivariate plot (612) to encompass the B-cell population (614), wherein the data corresponding to events within Gate-3 (503) constitutes the reduced data file (FCSdr 403) utilized for subsequent clustering and identification of leukemic MRD cells. Clause 6: The method of clause 4, wherein the automated data cleaning algorithm applied by the DCM (109), in addition to generating the cleaned data file (FCScd 404), causes the display on the user interface (303) of at least one graphical representation comparing data characteristics before cleaning (derived from FCSdr 403, e.g., plot 507) and after cleaning (derived from FCScd 404, e.g., plot 508), optionally accompanied by corresponding event counts (509), thereby enabling an operator to visually assess the effectiveness of the cleaning process and determine if adjustments to preceding steps, such as gating within the DRM (108), are warranted. Clause 7: The method of clause 4, wherein the graphical representations generated by the DClM (110) and displayed on the user interface (303) comprise at least one of a cluster tree (801), a cluster map (802), and a heatmap (803), wherein: a) the cluster tree (801) visually represents hierarchical relationships among the plurality of clusters, potentially illustrating groupings into metaclusters and indicating relative phenotypic distances between clusters; and b) the heatmap (803) displays the plurality of clusters (e.g., as rows) against the measured parameters (e.g., as columns), utilizing a color scale to represent the expression level (e.g., median or mean intensity) of each parameter within each cluster, thereby providing a comprehensive visualization of the immunophenotype associated with the cell population within each respective cluster. Clause 8:The method of clause 4, wherein the analysis performed by the CAM (111) using the cluster statistics (CS file 406) comprises displaying cluster information on the user interface (303) via a graphical representation (901), wherein: a) the graphical representation (901) displays axes corresponding to the measured parameters (902) and signal intensity (903); b) each cluster of the plurality of clusters is represented as a profile line (904) interconnecting points corresponding to the characteristic signal intensity value for that cluster for each respective parameter; and c) the method further comprises enabling the application of one or more analytical gates (905) based on signal intensity thresholds for selected parameters, wherein logical combinations of said gates (905) using operators such as AND, OR, or NOT can be employed to computationally filter or select clusters matching a predefined MRD phenotype.

Claims

Claims 1. A computer-implemented method for analysing minimal residual disease (MRD) associated with paediatric B-cell precursor acute lymphoblastic leukaemia (B-ALL) in a sample from a patient using multiparameter flow cytometry (MPFC) data, said method comprising: i) isolating a subset of data corresponding to a target cell lineage; ii) applying an automated algorithm to identify and remove data anomalies from the data subset to produce a cleaned data subset; iii) applying an automated, unsupervised clustering algorithm to group events in the cleaned data subset into a plurality of clusters based on the multiparameter data; and iv) analysing statistical properties of the plurality of clusters to identify at least one cluster corresponding phenotypically to MRD cells.

2. The computer-implemented method of claim 1, wherein step iii) further comprises generating graphical representations of the clusters.

3. The computer-implemented method of claim 1 or 2, wherein step i) is performed by a data reduction module (DRM).

4. The computer-implemented method of claims 1-3, wherein step ii) is performed by a data cleaning module (DCM).

5. The computer-implemented method of claims 1-4, wherein step iii) is performed by a data clustering module (DClM).

6. The computer-implemented method of claims 1-5, wherein step iv) is performed by a cluster analysis module (CAM).

7. The computer-implemented method of any preceding claim, the method comprising: i) isolating a subset of data corresponding to a target cell lineage using a Data Reduction Module (DRM), ii) applying an automated algorithm to identify and remove data anomalies from the data subset to produce a cleaned data subset using a Data Cleaning Module (DCM), iii) applying an automated, unsupervised clustering algorithm to group events in the cleaned data subset into a plurality of clusters based on the multiparameter data using a Data Clustering Module (DClM); and iv) analysing statistical properties of the plurality of clusters to identify at least one cluster corresponding phenotypically to MRD cells using a Cluster Analysis Module (CAM), wherein the DRM, the DCM, the DClM and the CAM comprise an Analytical Subsystem.

8. The computer-implemented method of claim 7, the method comprising:i) transmitting a file to the Analytical Subsystem or a file repository; ii) within the Data Reduction Module (DRM) of the Analytical Subsystem, retrieving the file and isolating a subset of data corresponding to a target cell lineage; iii) within a Data Cleaning Module (DCM) of the Analytical Subsystem, retrieving the reduced data file and applying an automated algorithm to identify and remove data anomalies from the data subset to produce a cleaned data subset; v) within a Data Cleaning Module (DCM) of the Analytical Subsystem, retrieving the reduced data file and applying an automated, unsupervised clustering algorithm to group events in the cleaned data subset into a plurality of clusters based on the multiparameter data; and iv) within a Cluster Analysis Module (CAM) of the Analytical Subsystem, retrieving the cluster statistics and analysing statistical properties of the plurality of clusters to identify at least one cluster corresponding phenotypically to MRD cells.

9. The computer-implemented method of claims 3-8, wherein the DRM isolates a subset of data corresponding to a target cell lineage by applying a data reduction process.

10. The computer-implemented method of claim 9, wherein the data reduction process comprises identifying a total leukocyte population within a reference gate and isolating a B-cell population within a subsequent gate, thereby generating a reduced data file containing data for the B-cell population and storing a count of the total leukocyte population.

11. The computer-implemented method of claims 6-10, wherein the CAM determines a quantity of cells within the identified at least one cluster corresponding to MRD cells and calculates a percentage of MRD cells by dividing the quantity of cells determined by the DRM by the count of the total leukocyte population stored by the DRM and multiplying by 100.

12. The computer-implemented method according to any one of claims 7-11, the method comprising; i) transmitting a file to the Analytical Subsystem or a file repository; ii) within the Data Reduction Module (DRM) of the Analytical Subsystem, retrieving the file and applying a data reduction process, comprising identifying a total leukocyte population within a reference gate and isolating a B-cell population within a subsequent gate, thereby generating a reduced data file containing data for the B-cell population and storing a count of the total leukocyte population;iii) within a Data Cleaning Module (DCM) of the Analytical Subsystem, retrieving the reduced data file and applying an automated data cleaning algorithm to remove data anomalies, thereby generating a cleaned data file; iv) within a Data Clustering Module (DClM) of the Analytical Subsystem, retrieving the cleaned data file and applying an automated, unsupervised clustering algorithm to partition events within the cleaned data file into a plurality of clusters based on their multiparameter data; v) within the Cluster Analysis Module (CAM), analysing the cluster statistics and optionally visualizing cluster expression profiles to identify at least one specific cluster exhibiting a multiparameter phenotype characteristic of leukemic MRD cells for B-ALL; and vi) within the Cluster Analysis Module (CAM), (1) determining a quantity of cells within the identified at least one specific cluster corresponding to MRD cells; and calculating a percentage of MRD cells by dividing the quantity of cells identified in step (1) by the count of the total leukocyte population stored in step (ii) and multiplying by 100.

13. The computer-implemented method according to claim 12, wherein following step iv), the DClM further generates graphical representations of the plurality of clusters and generates cluster statistics characterizing each cluster and stores the cluster statistics.

14. The computer-implemented method of any one of claims 4-13, wherein the file transmitted to the Analytical Subsystem or the file repository is an FCS file.

15. The computer-implemented method of claims 3-14, wherein the data reduction process performed by the DRM comprises displaying data from the file on the user interface, thereby enabling an operator to perform steps including: a) defining Gate-1 on a first bivariate plot to encompass the leukocyte population and storing the count of events within Gate-1 for subsequent MRD percentage calculation; and b) defining Gate-3 on a subsequent bivariate plot to encompass the B-cell population, wherein the data corresponding to events within Gate-3 constitutes the reduced data file utilized for subsequent clustering and identification of leukemic MRD cells.

16. The computer-implemented method of claims 4-15, wherein the automated data cleaning algorithm applied by the DCM, in addition to generating the cleaned datafile, causes the display on the user interface of at least one graphical representation comparing data characteristics before cleaning and after cleaning, optionally accompanied by corresponding event counts, thereby enabling an operator to visually assess the effectiveness of the cleaning process and determine if adjustments to preceding steps, such as gating within the DRM, are warranted.

17. The computer-implemented method of any of claims 5-16, wherein the graphical representations generated by the DClM and displayed on the user interface comprise at least one of a cluster tree, a cluster map, and a heatmap, wherein: a) the cluster tree visually represents hierarchical relationships among the plurality of clusters, potentially illustrating groupings into metaclusters and indicating relative phenotypic distances between clusters; and b) the heatmap displays the plurality of clusters against the measured parameters, utilizing a color scale to represent the expression level of each parameter within each cluster, thereby providing a comprehensive visualization of the immunophenotype associated with the cell population within each respective cluster.

18. The computer-implemented method of claim 17, wherein the heatmap displays the plurality of clusters as rows.

19. The computer-implemented method of claims 17 or 18, wherein the heatmap displays measured parameters as columns.

20. The computer-implemented method of claims 17-19, wherein the expression level is median or mean intensity.

21. The computer-implemented method of any of claims 5-20, wherein the analysis performed by the CAM using the cluster statistics comprises displaying cluster information on the user interface via a graphical representation, wherein: a) the graphical representation displays axes corresponding to the measured parameters and signal intensity; b) each cluster of the plurality of clusters is represented as a profile line interconnecting points corresponding to the characteristic signal intensity value for that cluster for each respective parameter; and c) the method further comprises enabling the application of one or more analytical gates based on signal intensity thresholds for selected parameters, wherein logical combinations of said gates using operators such as AND, OR, or NOT can be employed to computationally filter or select clusters matching a predefined MRD phenotype.

22. The computer-implemented method of any one of claims 3-21, wherein the DRM generates a FCSdr file.

23. The computer-implemented method of any one of claims 4-22, wherein the DCM generates a FCScd file.

24. The computer-implemented method of any one of claims 4-23, wherein the DCM is configured to apply an automated algorithm selected from the group consisting of flowClean, flowAI, and PeacoQC.

25. The computer-implemented method of any one of claims 5-24, wherein the DClM is configured to apply an automated, unsupervised clustering algorithm selected from the group consisting of FlowSOM, Phenograph, X-shift, flowMeans, and flowClust.

26. The computer-implemented method of any one of claims 6-25, wherein the CAM generates a CS file.

27. The computer-implemented method of any one of claims 6-62, wherein the Cluster Analysis Module is ClusterExplorer.

28. A method for analysing minimal residual disease (MRD) associated with paediatric B- cell precursor acute lymphoblastic leukaemia (B-ALL) in a sample from a patient using multiparameter flow cytometry (MPFC) data, the method comprising: i) treating the sample from the patient with a panel of fluorochrome-conjugated antibodies specific for a plurality of cellular markers relevant to distinguishing leukemic B-ALL cells from non-leukemic B-ALL cells; ii) measuring the treated sample from the patient using an MPFC instrument to generate measured signals for each of a plurality of events across multiple parameters, and converting said measured signals into a digital data format; and iii) using a processor, performing the computer-implemented method of any one of claims 1-27.

29. The method of claim 28, wherein the sample from a patient comprises bone marrow cells.

30. The method of claim 28, wherein the sample from a patient comprises peripheral blood cells.

31. The method of claims 28-30, wherein the sample from a patient comprises at least 4x106cells.

32. The method of claims 28-31, wherein the panel of fluorochrome-conjugated antibodies specific for a plurality of cellular markers comprises a combination of antibodies directed against markers, wherein the combination of antibodies comprises:i) antibodies targeting markers CD45, CD20, CD34, CD38, CD10, CD58, CD66c, CD73, CD81, CD123, CD304, CD44, CD86, CD99 and CD371, and ii) antibodies targeting markers CD19 and / or CD22, wherein the antibodies are conjugated with fluorochromes.

33. The method according to claim 32, wherein the panel of fluorochrome-conjugated antibodies specific for a plurality of cellular markers further comprises an antibody targeting CD36.

34. The method of claims 32 or 33, wherein the panel of fluorochrome-conjugated antibodies comprises the panel of Table 1.

35. A system comprising a memory and a processor configured to: i) isolate a subset of data corresponding to a target cell lineage; ii) apply an automated algorithm to identify and remove data anomalies from the data subset to produce a cleaned data subset; iii) apply an automated, unsupervised clustering algorithm to group events in the cleaned data subset into a plurality of clusters based on the multiparameter data, and optionally generate graphical representations of the clusters; and iv) analyse statistical properties of the plurality of clusters to identify at least one cluster corresponding phenotypically to MRD cells.

36. The system of claim 35 for analysing minimal residual disease (MRD) associated with paediatric B-cell precursor acute lymphoblastic leukaemia (B-ALL) in a sample from a patient using multiparameter flow cytometry (MPFC) data, the system comprising a memory and a processor, wherein the system is configured to: i) isolate a subset of data corresponding to a target cell lineage; ii) apply an automated algorithm to identify and remove data anomalies from the data subset to produce a cleaned data subset; iii) apply an automated, unsupervised clustering algorithm to group events in the cleaned data subset into a plurality of clusters based on the multiparameter data; and, iv) analyse statistical properties of the plurality of clusters to identify at least one cluster corresponding phenotypically to MRD cells.

37. The system according to claims 35 or 36, wherein following step iii), the system is further configured to generate graphical representations of the clusters.

38. The system of claims 35-37, wherein the system comprises a data reduction module (DRM) configured to computationally isolate a subset of data corresponding to a target cell lineage.

39. The system of claims 35-38, wherein the system comprises a data cleaning module (DCM) configured to apply an automated algorithm to identify and remove data anomalies from the data subset.

40. The system of claims 35-39, wherein the system comprises a data clustering module (DClM) configured to apply an automated, unsupervised clustering algorithm to group events in the cleaned data subset into a plurality of clusters based on the multiparameter data.

41. The system of claims 35-40, wherein the system comprises a Cluster Analysis Module (CAM) configured to analyze statistical properties of the plurality of clusters to identify at least one cluster corresponding phenotypically to MRD cells.

42. The system of claims 35-41, wherein the system comprises: i) a data reduction module (DRM) configured to computationally isolate a subset of data corresponding to a target cell lineage; ii) a data cleaning module (DCM) configured to apply an automated algorithm to identify and remove data anomalies from the data subset; iii) a data clustering module (DClM) configured to apply an automated, unsupervised clustering algorithm to group events in the cleaned data subset into a plurality of clusters based on the multiparameter data; and iv) a Cluster Analysis Module (CAM) configured to analyze statistical properties of the plurality of clusters to identify at least one cluster corresponding phenotypically to MRD cells.

43. The system of any one of claims 41-42, wherein the CAM is further configured to generate graphical representations of the clusters.

44. The system of any one of claims 42-43, the system comprising: a) an Acquisition Subsystem comprising an MPFC instrument configured to measure cellular parameters of the sample from the patient and signal processing electronics configured to convert measured signals into a digital data format; b) a Control and File Generation Subsystem operatively coupled to the Acquisition Subsystem, configured to receive the digital data format and transform said digital data into a file format; c) an Analytical Subsystem operatively coupled to receive the file format, configured to analyze the MPFC data contained therein, said Analytical Subsystem comprising: i. a Data Reduction Module (DRM) configured to computationally isolate a subset of data corresponding to a target cell lineage;ii. a Data Cleaning Module (DCM) configured to apply an automated algorithm to identify and remove data anomalies from the data subset; iii. a Data Clustering Module (DClM) configured to apply an automated, unsupervised clustering algorithm to group events in the cleaned data subset into a plurality of clusters based on the multiparameter data; and iv. a Cluster Analysis Module (CAM) configured to analyze statistical properties of the plurality of clusters to identify at least one cluster corresponding phenotypically to MRD cells; d) a file repository operatively coupled for storing data files; and e) at least one communication network operatively connecting the Control and File Generation Subsystem, the Analytical Subsystem, and the file repository and the external server.

45. The system of claim 44, wherein the CAM is further configured to generate graphical representations of the clusters.

46. The system of claim 44 or 45, wherein the system further comprises an external server operatively accessible for providing analysis algorithms or related resources.

47. The system of claims 44-46, wherein the analytical subsystem comprises four interconnected functional modules configured to operate sequentially.

48. The system of claims 44-47, wherein the Control and File Generation Subsystem and the Analytical Subsystem are implemented on one or more computing devices, each comprising at least one central processing unit (CPU), primary memory, secondary memory, a network interface, a user interface, and at least one input device.

49. The system according to claim 48, wherein the one or more computing devices further comprises at least one graphics processing unit (GPU).

50. The system of claims 47-49, wherein the interconnected functional modules of the Analytical Subsystem are further characterized in that: a) the Data Reduction Module (DRM) is configured to isolate B-lineage cells to generate a reduced data file; b) the Data Cleaning Module (DCM) is configured to apply an automated algorithm to remove anomalies from the reduced data file to generate a cleaned data file; c) the Data Clustering Module (DClM) is configured to apply an automated, unsupervised clustering algorithm to the cleaned data file to generate the plurality of clusters and associated cluster statistics; andd) the Cluster Analysis Module (CAM) is configured to process the cluster statistics to identify the at least one cluster corresponding phenotypically to MRD cells based on multiparameter expression profiles and calculate a quantity of said MRD cells.

51. The system of claims 39-50, wherein the DCM is configured to apply an automated algorithm selected from the group consisting of flowClean, flowAI, and PeacoQC.

52. The system of claims 40-51, wherein the DClM is configured to apply an automated, unsupervised clustering algorithm selected from the group consisting of FlowSOM, Phenograph, X-shift, flowMeans, and flowClust.

53. The system of any one of claims 41-52, wherein the Cluster Analysis Module is ClusterExplorer.

54. The system of any one of claims 38-53, wherein the DRM generates a FCSdr file.

55. The system of any one of claims 39-54, wherein the DCM generates a FCScd file.

56. The system of any one of claims 40-55, wherein the CAM generates a CS file.

57. The system of any one of claims 35-56, wherein the sample from a patient comprises bone marrow cells.

58. The system of any one of claims 35-56, wherein the sample from a patient comprises peripheral blood cells.

59. The system of claims 35 -58, wherein the sample from a patient comprises at least 4x106cells.

60. A computer program comprising instructions which, when the program is executed by a computer, cause the computer to carry out the method of any one of claims 1-27.

61. A computer-readable medium comprising instructions which, when executed by a computer, cause the computer to carry out the method of any one of claims 1-27.

Images