Integration of Next-Generation Sequencing in Diagnosing and Minimal Residual Disease Detection in Patients With Philadelphia Chromosome-Like Acute Lymphoblastic Leukemia

Philadelphia-like (Ph-like) acute lymphoblastic leukemia (ALL) is a high-risk subtype of B cell ALL. It accounts for 20% of all B cell ALL cases and is similar to BCR-ABL1 in gene expression profile but lacks BCR-ABL fusion. It is highly heterogeneous and is characterized by genetic alterations that activate kinase and cytokine receptor signaling. Most of these alterations are amenable to tyrosine kinase inhibitors. Ph-like ALL is prevalent in pediatric and young adults, more common in males, and frequently seen in patients with Hispanic ancestry. It is associated with inadequate response to induction therapy, high minimal residual disease (MRD) levels, and increased risk of relapse. Overall survival and event-free survival are also inferior in these patients as compared to non-Ph-like ALL. In the clinical practice, low-density array, real-time quantitative polymerase chain reaction (RQ-PCR), flow cytometry, fluorescence in situ hybridization are used to identify genetic alteration in these patients. With the advent of next-generation sequencing (NGS), our understanding of disease pathogenesis and precision medicine has been improved. In this review, we analyzed data from several studies that used NGS as one of the diagnostic methods to identify genomic lesions in this high-risk subtype of B cell ALL. Studies have shown that NGS is a vital technique to identify various genomic lesions at diagnosis and throughout the treatment that can be missed by the widely used current methods. NGS has improved our understanding of various genomic lesions associated with Ph-like ALL and has helped define disease pathogenesis, MRD evaluation, and stratify therapy to prevent over or under treatment. We are in the era of precision medicine. Therefore unbiased, comprehensive genomic characterization of Ph-like ALL is important to implicate treatment directed against these genomic lesions and improve outcomes in these patients. We also analyzed data from studies that compared NGS with multi-flow cytometry and RQ-PCR for the evaluation of MRD. In the future, more extensive prospective studies are required to confirm the prognostic usefulness of NGS.

precision, that can impact the clinical decision making. It is both sensitive and specific, generates more data with a smaller sample, it is faster, more efficient and its cost is rapidly decreasing. The concept of NGS involves series of massively parallel sequencing through various approaches such as targeted gene sequencing, whole-genome sequencing (WGS), which can reveal structural variations (SVs), wholeexome sequencing (WES) that is useful for detecting point mutations, transcriptome sequencing (RNA-seq) which is used to analyze the expression of mRNA or non-coding RNA and can also identify sequence mutation and fusion genes [6].
NGS has led to the identification of many newer molecular entities of ALL and has also provided a more profound understanding of the ones that are already known [7]. Both B cell ALL and T cell ALL are comprised of multiple subtypes defined by structural DNA alterations as an initiating lesion, with secondary somatic (tumor acquired) alterations and sequence mutation, which jointly contribute to leukemogenesis. Structural alterations include aneuploidy and chromosomal rearrangements that can result in the expression of chimeric fusion genes. Sequence mutations commonly alter lymphoid development, cytokine receptors, kinase, and RAS signaling, tumor suppression, and chromatin modification [5].
B cell ALL represents 75% of all cases of ALL and is comprised of various molecular subtypes [2]. In 2016, a new subtype of B-cell ALL was recognized by WHO classification of myeloid neoplasm and acute leukemia; it was called BCR-ABL1-like or Philadelphia chromosome-like (Ph-like) B cell ALL [8]. It was first detected by Mullighan and his colleagues from the Children's Oncology Group (COG) and St. Jude Children's Research Hospital (SJCRH), and den Boer and Colleagues from the Netherlands in 2009. Ph-like ALL has a gene expression profile similar to BCR-ABL1 but lacks BCR-ABL1 expression [9]. The Hallmark of Ph-like ALL is the high frequency of IKAROS family zinc finger one (IKZF1) alteration that is 70%-80% as compared to non-Ph-like ALL that is 15%. It is associated with high-risk clinical features, inadequate response to induction therapy, high frequency of persistent minimal residual disease (MRD), and poor outcome, with a five-year disease-free survival of about 60% [10][11][12][13][14].
Transcriptome sequencing studies have shown that Ph-like ALL have a significant genetic heterogeneity with > 70 discrete alterations that dysregulate several classes of cytokine receptors and tyrosine kinases [11,15]. In clinical practice, genetic analysis is conducted through the low throughput techniques like low-density arrays (LDA), fluorescence in situ hybridization (FISH), and Q-PCR. These techniques identify a limited number of detectable alterations but are still primarily used worldwide [7]. NGS has improved our understanding of disease pathogenesis and has also helped identify key biomarkers that are of diagnostic and prognostic importance. Ph-like ALL is a challenging subgroup due to its highly heterogeneous background. Standardized diagnostic tests and clinical management of this subgroup are also not well defined [15,16].
MRD is defined as a state in which leukemic cells remain after treatment. MRD measurement is the single most important factor in patients with ALL that can be used to stratify risk, determine prognosis, and to guide the treatment. Since it has been implemented to stratify treatment strategies, the cure rate has been improved in all age groups. Methods that are widely used in clinical practice to measure MRD are multiparametric flow cytometry and Q-PCR. Studies have shown that these methods can give falsepositive results, and measurement of MRD cannot rely solely on these methods. With the advent of NGS, more attention has been shown to NSG-MRD assay, as it is more sensitive and specific as compared to other methods [17,18].
It is essential to understand the benefits of NGS in clinical practice and to incorporate information gathered through NGS to improve risk stratification and new drug discoveries from the perspective of precision medicine. In this comprehensive review, we will highlight the benefits of NGS in disease definition, risk stratification, and treatment strategies in patients with Ph-like B cell precursor ALL. We will also briefly discuss the studies showing the benefits of NGS in disease definition and MRD monitoring.

Review Method
PubMed and PMC online databases were used exclusively for the collection of corresponding data. Results yielded 219 scientific papers based on keywords, and all were in English. After applying specific inclusion-exclusion criteria, 28 scientific papers were deemed relevant. All 28 articles met the quality specification and were peer-reviewed.

Inclusion and exclusion criteria
All selected scientific papers were written in English and included data collected and reviewed from 2010-2020. The studies that were excluded did not match the inclusion and exclusion criteria.

Result
Ph-like ALL is a high-risk subtype of B cell ALL. Current methods used to diagnose this high-risk subtype include LDA, Q-PCR, FISH, and flow cytometry. In the last few years, the application of NGS has proved to be an essential diagnostic tool in identifying genomic lesions with higher sensitivity and less bias. Of the four selected research papers, all used NGS as one of the diagnostic tools to identify various genomic lesions in Ph-like ALL. Robert et al. studied Ph-like ALL in children, adolescents, young adults, and older adults. Whereas, Reshmi et al., studied disease in children and young adults and Herold et al. studied Ph-like ALL in adult patients. Studies showed that NGS is useful in identifying the complete repertoire of the leukemic and the normal cells. It can identify new fusions and the splice variants of known lesions during the course of the disease. Despite heterogeneity of Ph-like ALL, most of the treatable genomic lesions can be identified by conventional methods like LDA, Q-PCR, and FISH, however, these methods can only identify known fusions [10,14,19,20]. Table 1   and higher sensitivity with NGS as compared to RQ-PCR without the need to identify patientspecific probes [23,24]. Faham et al. compared all three methods and confirmed that NGS could detect residual disease at levels below one in one million leukocytes (0.0001%), which represents one-two folds of the higher sensitivity of NGS than the gold standard methods [25].

Ph-like ALL
Ph-like ALL is a high-risk subtype and accounts for 20% of all B cell ALL cases. Its prevalence differs by age, gender, ethnicity and National Cancer Institute (NCI)-defined risk groups. It comprises of approximately 10% in children of NCI standard risk (age:1-9 years and white blood cell (WBC) count <50 x 10 9 /L), 13% in children of NCI high risk ( age: 10-15 and WBC count >50 x 10 9 /L), 21% of adolescents (16-20 years), 27% of young adults (21-39 years) and 20%-24% of adults over 40 years of age. It is more common in males with male to female ratio is 2:1 as compared to 1.6:1 in non-Ph-like ALL. It is highly prevalent in Hispanic patients due to presence of germline variant GATA3 in Hispanic and native American genetic ancestry, particularly common with Ph-like ALL with cytokine receptor like factor two-rearrangement (CRLF2-R). Phlike ALL is a frequently occurring subtype in pediatric and young adults, and is associated with poor response to induction chemotherapy, elevated WBC count, high levels of MRD. Overall survival (OS) and event free survival (EFS) across all age group is inferior as compared to non-Ph-like ALL. It is important to identify these high-risk patients who could potentially benefit from targeted therapy [10,14,19].
Ph-like ALL is similar to BCR-ABL1 ALL in gene expression profile but lacks BCR-ABL1 fusion protein expressed from the t(9;22) (q34; q11.2). Ph-like ALL with IKZF1 alteration has an inferior five-year EFS as compared to Ph-like ALL with no IKZF1 alteration. IKZF1 alterations are more common in patients with kinase mutation as compared to patients with sequence mutation [9][10][11][12][13][14]. Approximately half of the Ph-like ALL cases have CRLF2-R with concomitant Janus kinase one/Janus kinase two (JAK1/JAK2) mutation. Recent introduction of transcriptome sequencing and whole genome sequencing in patients without CRLF2-R, have identified kinase and cytokine receptor signaling alteration in Ph-like ALL patients. These include, erythropoietin receptor (EPOR), ABL-class fusion, RAS pathway mutations, uncommon kinase mutation and sequence mutations activating JAK-STAT pathway [26]. patients without CRLF2 rearrangement) with Ph-like ALL underwent one or more different types of NGS, like transcriptome sequencing, whole genome sequencing, and whole-exome sequencing. NGS was also performed in 160 patients with non-Ph-like ALL. Genomic analysis identified kinase signaling alterations in 91% of patients with Ph-like ALL, and they were divided into distinct subgroups. These included JAK-STAT mutations, ABL class fusion, RAS pathway mutations, other uncommon mutations, and no kinase alterations [10]. These are discussed later in this section.

Identification of Ph-like ALL
Reshmi et al. also studied patients in the age group 1-30 years [19]. Roberts et al. also studied Ph-like ALL in adult patients aged 21-86 years [14]. They retrospectively identified patients from multi-center clinical trial between 1999 and 2021. These trials included 3474 total participants, 909 patients were diagnosed with B cell ALL, 798 of whom were selected for this study due to the availability of suitable material for genomic analysis. Ph-like ALL was identified in 194 out of 798 patients. Out of 194, 180 patients were selected for genomic analysis. CRLF2 expression was identified in 99 out of 194 patients. Eight six of the remaining 95 patients underwent transcriptome sequencing. Kinase and cytokine receptor signaling alterations were identified in 88% of patients with Ph-like ALL. They were divided into subtypes that are discussed later in this section. Eleven new genetic alterations of B ALL were also identified during this study.
Herold et al. also studied Ph-like ALL in adult patient aged 15-65 years [20]. They identified 207 patients with B cell ALL from two clinical trial between 1999 and 2008. A total of 1475 patients were included in these trials. Ph-like ALL was diagnosed in 26 out of 207 patients using gene expression profiling. Targeted amplicon sequencing was performed on 131 genes that were recurrently mutated in 16 patients with Ph-like ALL and 23 patients with remaining BCP-ALL.
All studies concluded that Ph-like ALL has a large number of complex genomic lesions that can activate a limited number of signaling pathways, and many of these are amenable to tyrosine kinase inhibitors. Despite of complex nature of Ph-like ALL, majority of treatable genomic lesions can be rapidly and accurately identified by the conventional methods like LDA, FISH and Q-PCR. These methods are simple and cost effective, but can only identify known fusions. Newer fusions or splice variants of known fusions cannot be identified with conventional methods; therefore, patients with B cell ALL at screening or those identified as having Ph-like ALL should undergo unbiased RNA sequencing. It can identify all genomic lesions in a comprehensive and timely manner. It is not feasible to incorporate NGS for diagnosis of Ph-like ALL, particularly in the clinical trials involving large number of patients from multiple sites, but its declining cost and improvement in bioinformatic analysis in future can make NGS a powerful tool to identify genomic lesions and help in precision medicine.

Kinase Alterations in Ph-like ALL
Gene expression profiling has identified kinase and cytokine receptor signaling alteration in Ph-like ALL. They are divided into four subtypes: JAK-STAT mutations, ABl class fusion, RAS pathway mutation, and uncommon kinase mutation [10,14].
JAK-STAT signaling: It is the largest class of kinase activation lesion that can result in the activation of Janus kinase signal transducer and activator of transcription (JAK-STAT) signaling. Half of the mutation is a rearrangement of cytokine receptor-like factor two (CRLF2). CRLF2 is dysregulated by translocation of the immunoglobulin heavy chain (IGH-CRLF2) or focal deletion resulting in the formation of P2RY8-CRLF2. Both of these can result in the expression of CRLF2, which heterodimerizes with interleukin seven receptor alpha (IL7RA) to form the thymic stromal-derived lymphopoietic receptor (TSLPR) [28]. Less commonly, there is a sequence mutation of CRLF2 (F232C). CRLF2-R is common in Phlike ALL and patients with down syndrome, and it is age-dependent with IGH-CRLF2 common in older age and Hispanic ancestry, whereas P2RY8 is seen in young age. Most studies have shown that CRLF2 is associated with poor prognosis primarily when related to concomitant IKZF1 alteration. The majority of CRLF2 have concomitant alteration, particularly JAK1 or JAK2, and are associated with the poorest outcome.
Other genetic alterations that activate JAK-STAT signaling are Janus kinase two (JAK2) and erythropoietin receptor (EPOR). There are sequence mutations and DNA copy number alterations that activate the JAK-STAT signaling pathway but have no kinase or cytokine receptor gene rearrangement. These include Janus kinase one (JAK1), JAK2, interleukin seven receptor (IL7R), Src homology two adaptor protein three (SH2B3), interleukin two receptor beta (IL2RB), and tyrosine kinase two (TYK2). All these above-mentioned genetic lesions activate the JAK-STAT signaling pathway and can be treated with JAK inhibitors like ruxolitinib [9,10,14,28].
ABL class fusion: Kinases that are altered in this subtype of Ph-like ALL are Abelson murine leukemia one (ABL1), ABL2, colony-stimulating factor one receptor (CSF1R), platelet-derived growth factor receptor A (PDGFRA), and PDGFRB. These kinases can be inhibited by ABL1 inhibitors like imatinib and dasatinib [10,14].

3: Subtypes and therapeutic targets of Ph-like ALL
Abbreviations: JAK-STAT, Janus kinase signal transducer and activator of transcription; Ph-like ALL, Philadelphia chromosome-like acute lymphoblastic leukemia Figure 1 shows the frequency of genetic subtypes in patients with Ph-like ALL by age group.

FIGURE 1: Frequency of genetic subtypes in patients with Ph-like ALL by age group
Abbreviations: NCI: National Cancer Institute, HR: high risk (age:10-15 and WBC count >50 x 10 9 /L); Ph-like ALL: Philadelphia chromosome-like acute lymphoblastic leukemia

Measurement of MRD
MRD measurement is performed to evaluate treatment response and to define risk stratification. Undetectable MRD after induction therapy indicate good prognosis, whereas persistent MRD indicates a high risk of relapse. MRD analysis should be accurate, sensitive (≤0.01% or ≤10 -4 ) and specific. MRD is diagnosed by targeting molecular rearrangements like gene fusion, immunoglobulins (IgH), and T cell receptor (TCR) gene. Current methods used to measure MRD are multi-parametric flow cytometer (MFC), ASO-PCR, and NGS of the IgH VDJ and/or TCR gene rearrangement. Ph-like ALL is associated with persistent MRD levels and poor response to therapy [21][22][23][24][25]. It is crucial to accurately measure MRD in these high-risk subtype patients to identify MRD levels and treat accordingly. Several studies have been done to compare NGS, MFC, and RQ-PCR for the measurement of residual disease. NGS can detect all genomic rearrangements at diagnosis and during the course of disease and treatment as compared to Q-PCR, this can reduce the risk of false-negative MRD results, that occur due to clonal evolution during the course of disease and treatment [25].
Above studies, concluded that, undetectable pre-HCT IgH VDJ NGS-MRD have a very low risk of relapse, whereas detectable post-HCT IgH VDJ NGS-MRD have a high risk of relapse. NGS-MRD is superior to MFC-MRD in predicting overall survival, relapse and no relapse in pre-HCT and post-HCT patients. If the patient has no pre-HCT NGS-MRD, they are eligible for less intense treatment strategy. NGS is more sensitive than MFC in the detection of residual disease and can identify patients that might benefit from transplant. Due to increased sensitivity of NGS, peripheral blood samples can complement bone marrow samples, for the detection of residual disease, as data from Torra et al. showed that only 17% of cases were not detected by NGS PB samples, but these cases had low MRD levels in bone marrow. NGS is more specific than Q-PCR in predicting relapse in post-transplant patients, as data indicates that Q-PCR can give false-positive results that can lead to over treatment. NGS sensitivity is comparable with Q-PCR but without the need for patientspecific probes. Studies showed pre-treatment high concordance between these methods. There were some discordances post-treatment due to clonal evolution that are not detected by MFC or Q-PCR at an earlier time point as compared to NGS. Table 4 summarizes the advantages and disadvantages of NGS, RQ-PCR and MFC. Expensive, require expertise, lack of standardization, require highly efficient bioinformatics and large number of data is generated to be stored and analyzed.

Limitations
Our study had faced a few limitations as described below: 1. Most of the clinical studies included pediatric, adolescence and young adults with some studies focusing on older adults.
2. Despite having found a large number of sample sizes, it would be beneficial if NGS have been conducted on a larger percentage of the chosen sample size patients.
3. Every study utilized a different approach to NGS. 4. We selected studies conducted in the English language only that might have caused us to miss some more valuable reviews.

Conclusions
This review article aimed to analyze and assess the importance of NGS in genomic characterization of PHlike ALL and MRD monitoring. It was shown: (1) Ph-like ALL has a low survival in all age groups and is highly heterogenous subtype of B cell ALL. More accurate method like NGS is required for the genomic characterization of Ph-like ALL early in the disease, to improve outcomes in these patients. Analysis of these fusions should not rely only on conventional methods, as they can miss some and/or many targetable alterations.
(2) Ph-like ALL is associated with persistent MRD levels and high risk of relapse. Clinical studies have shown that NGS is highly accurate, sensitive, specific and can measure MRD levels early after the induction therapy that can be missed by other methods used. (3) Implementation of NGS in the clinical trials have broadened our understanding of disease pathogenesis, allowing careful patient management in regard to prognosis, treatment choice, MRD evaluation and new drug discoveries.
Our findings support that in the future, all patients with B cell ALL should be considered to undergo expression-based screening at diagnosis either with LDA, RQ-PCR or FISH to identify phenotype, followed by the search of targetable genotype and measurement of MRD levels during the treatment with any of the various approaches of NGS. We believe that in the future, we will see larger prospective studies implementing NGS across all age groups, that will completely describe the genomic landscape and will help refine targeted and individualized therapy in this high-risk subtype of B cell ALL.

Conflicts of interest:
In compliance with the ICMJE uniform disclosure form, all authors declare the following: Payment/services info: All authors have declared that no financial support was received from any organization for the submitted work. Financial relationships: All authors have declared that they have no financial relationships at present or within the previous three years with any organizations that might have an interest in the submitted work. Other relationships: All authors have declared that there are no other relationships or activities that could appear to have influenced the submitted work.