Has Ph-like ALL Superseded Ph+ ALL as the Least Favorable Subtype?

Abstract

Philadelphia chromosome-like acute lymphoblastic leukemia (Ph-like ALL) is a subset of high-risk B-ALL associated with high relapse risk and inferior clinical outcomes across the pediatric-to-adult age spectrum. Ph-like ALL is characterized by frequent IKZF1 alterations and a kinase-activated gene expression profile similar to that of Philadelphia chromosome-positive (Ph+) ALL, yet lacks the canonical BCR-ABL1 rearrangement. Advances in high-throughput sequencing technologies during the past decade have unraveled the genomic landscape of Ph-like ALL, revealing a diverse array of kinase-activating translocations and mutations that may be amenable to targeted therapies that have set a remarkable precision medicine paradigm for patients with Ph+ ALL. Collaborative scientific efforts to identify and characterise Ph-like ALL during the past decade has directly informed current precision medicine trials investigating the therapeutic potential of tyrosine kinase inhibitor-based therapies for children, adolescents, and adults with Ph-like ALL, although the most optimal treatment paradigm for this high-risk group of patients has yet to be established. Herein, we describe the epidemiology, clinical features, and biology of Ph-like ALL, highlight challenges in implementing pragmatic and cost-effective diagnostic algorithms in the clinic, and describe the milieu of treatment strategies under active investigation that strive to decrease relapse risk and improve long-term survival for patients with Ph-like ALL as has been successfully achieved for those with Ph+ ALL.

Definition of Ph-like ALL

Ph-like ALL was identified in 2009 independently by two research groups using different gene expression profiling classifiers. Researchers in the Children’s Oncology Group (COG), University of New Mexico (UNM), and St Jude Children’s Research Hospital (SJCRH) via the National Cancer Institute (NCI) TARGET initiative (https://ocg.cancer.gov/programs/target/projects/acute-lymphoblastic-leukemia) utilized Affymetrix gene expression microarray data to identify 257 gene probe sets that defined a distinct gene expression signature of both Ph+ ALL and Ph-like ALL (1), while den Boer and colleagues at Erasmus Medical Center and the Dutch Children’s Oncology Group used hierarchical clustering of 110 probe sets to predict 6 major pediatric ALL subtypes (ETV6-RUNX1, high hyperdiploid, TCF3-PBX1, KMT2A-rearranged, BCR-ABL1, and TALL) (2). Despite sharing only 9 common probe sets of 7 genes (CCND2, SH3BP5, ABL1, SOCS2, DUSP6, LST1 and EGFL7), both gene classifiers identified a subset of high-risk B-ALL patients with poor survival who had frequent deletions of B-cell development and transcription factor genes, such as IKZF1 (3). These assays and additional advances in RNA-sequencing were further able to define new genetic alterations deregulating tyrosine kinase or cytokine receptor genes in Ph-like ALL, including CRLF2, JAK2, ABL1, PDGFRB, and others (4). The COG/UNM/SJCRH group subsequently developed a clinically-validated Ph-like ALL screening assay, which measures the expression of 8 or 15 genes in a 384-well low-density microarray (LDA) microfluidic card to detect the Ph-like ALL signature. This clinical assay is now capable of results return within 24–48 hours of sample submission (5). The LDA assay is currently used by the COG and other consortia to screen patients with newly-diagnosed high-risk B-ALL for the Ph-like signature and to allocate those with LDA positivity for downstream molecular testing to identify specific Ph-like-associated genetic alterations (6). The most clinically relevant endpoint in patients with Ph-like ALL remains identification of specific translocations and mutations that activate oncogenic kinase signaling and may be therapeutically targeted.

Biology and genomic landscape of Ph-like ALL

In their 2014 landmark paper, Roberts and colleagues described the genomic landscape of Ph-like ALL via a comprehensive genomic analysis of 154 high-risk B-ALL specimens from children and adolescents/young adults (AYAs) (7). The unifying molecular hallmark of Ph-like ALL resides in the heterogeneous spectrum of genetic alterations activating cytokine receptor genes and kinase-signaling pathways (7). These alterations can be subdivided in 4 distinct genomically-defined subsets based upon their underlying kinase-activating lesions: (1) alterations in JAK/STAT pathway genes (predominantly CRLF2, JAK2, EPOR, IL7R, SH2B3), (2) ABL class alterations (ABL1, ABL2, CSF1R, LYN, PDGFRA, and PDGFRB), (3) uncommon Ras pathway mutations (NRAS, KRAS, NF1, PTPN11, CBL), and (4) rare kinase fusions (NTRK3, PTK2B, BLNK) and are described in greater detail below (Table 1, Figure 1) (7, 8). While CRLF2 (cytokine receptor-like factor 2) rearrangements also occur with lower frequency in children with standard-risk (SR) B-ALL (9–11) and in >50% of trisomy 21/Down syndrome (DS)-associated B-ALL (12, 13) (and may or may not have the associated Ph-like gene expression signature), other Ph-like ALL-associated kinase fusions have only extremely rarely been discovered in patients with SR disease.

Table 1.

Current landscape of Ph-like ALL kinase rearrangements, therapeutic targets, and clinical trials.

Ph-like genetic subgroups	3’ kinase genes	5’ fusion partner genes	Kinase inhibitors	Clinical trials
JAK/STAT pathway alterations	CRLF2	CSF2RA, IGH, P2RY8	ruxolitinib	NCT02420717 (MDACC) NCT02723994 (COG AALL1521) NCT03117751 (SJCRH Total XVII) NCT03571321 (University of Chicago)
	JAK2	ATF7IP, BCR, EBF1, ETV6, GOLGA5, HMBOX1, OFD1, PAX5, PCM1, PPFIBP1, RFX3, SMU1, SNX29, SSBP2, STRN3, TERF2, TPR, USP25, ZBTB46, ZNF274, ZNF340	ruxolitinib
	EPOR	IGH, IGK, IGL, LAIR1, THADA	ruxolitinib
	TSLP	IQGAP2	ruxolitinib
	IL2RB	MYH9	ruxolitinib
ABL class alterations	ABL1	CENPC, ETV6, FOXP1, LSM14A, NUP153, NUP214, RANBP2, RCSD1, SFPQ, SHIP1, SNX1, SNX2, SPTNA1, ZMIZ1	dasatinib, imatinib, others	NCT01406756 (COG AALL1131) NCT02143414 (SWOG S1318) NCT02420717 (MDACC) NCT03007147 (COG AALL1631) NCT03117751 (SJCRH Total XVII)
	ABL2	PAG1, RCSD1, ZC3HAV1	dasatinib, imatinib
	CSF1R	MEF2D, SSBP2, TBL1XR1	dasatinib, imatinib
	PDGFRA	FIP1L1	dasatinib, imatinib
	PDGFRB	ATF7IP, EBF1, ETV6, NUMA1, SNX29, SSBP2, TERF2, TNIP1, ZEB2, ZMYND8, ZNF608	dasatinib, imatinib
	LYN	GATAD2A, NCOR1	dasatinib, imatinib
Other kinases	NTRK3	ETV6	entrectinib larotrectinib	NCT03066661 NCT03834961
	PTK2B	KDM6A, STAG2, TMEM2	FAK inhibitors
	FGFR1	BCR	ponatinib
	FLT3	ZMYM2	FLT3 inhibitors
	TYK2	MYB, SMARCA4, ZNF340	JAK1/3 inhibitor
	BLNK	DNTT
	CBL	KANK1
	DGKH	ZFAND3

COG = Children’s Oncology Group, SJCRH = St Jude Children’s Research Hospital, MDACC = MD Anderson Cancer Center, SWOG = Southwestern Oncology Group.

Figure 1. Frequency of Ph-like ALL and genetic subtypes by age group.

NCI HR B-ALL = National Cancer Institute high risk B-acute lymphoblastic leukemia, NCI SR B-ALL = National Cancer Institute standard risk B-acute lymphoblastic leukemia.

Deletions of IKZF1 and other lymphoid transcription factor genes occur commonly in Ph-like ALL, as has been similarly seen in Ph+ ALL (7, 14–17). IKZF1 encodes the zinc-finger DNA-binding Ikaros, a transcription factor essential for B-cell lymphoid development. Its alteration results in acquired stem cell-like properties, aberrant bone marrow stromal adhesion, and chemotherapy resistance (18–20). The most common IKZF1 alteration is intragenic focal deletion of exons 4–7, which results in the dominant-negative Ik6 isoform (20). In one study, IKZF1 alterations were detected in 68% of Ph-like ALL compared to 16% of non-Ph-like ALL cases (7). Inferior clinical outcomes in patients with IKZF1-deleted Ph-like ALL have also been reported (7, 21). Recent studies from European consortia have further described inferior outcomes of patients with the new IKZF1^plus molecular profile, which is defined by deletion of IKZF1 co-occurring with one or more deletions in CDKN2A (heterozygous or homozygous), CDKN2B (homozygous only), PAX5, or the pseudoautosomal region 1 (PAR1) region of the sex chromosomes where CRLF2 is located and in the absence of ERG deletion (22–25). Stanulla and colleagues reported that the IKZF1^plus signature conferred the highest hazard ratio for relapse in multivariate analysis and could be incorporated in clinical decision algorithms to refine risk stratification in addition to MRD response (22). The IKZF1^plus subgroup in these studies likely contains a high proportion of Ph-like ALL patients, as both populations share adverse clinical and biologic features of higher white blood cell (WBC) at diagnosis, poor prednisone response, positive minimal residual disease (MRD) after induction therapy, and higher frequency of the germline GATA3 variant rs3824662 reported to predispose to developing Ph-like ALL (22–24, 26).

JAK/STAT pathway gene alterations

Approximately 50% of children and AYAs with Ph-like ALL harbor CRLF2 rearrangements (10, 27, 28), which leads to CRLF2 overexpression and increased surface protein expression of the thymic stromal lymphopoietin receptor (TSLPR; encoded by CRLF2) detectable by flow cytometry (29). CRLF2 alterations occur via two major mechanisms: (1) focal deletion of PAR1 on chromosomes Xp22/Yp11 resulting in the P2RY8-CRLF2 fusion or as (2) translocation to the immunoglobulin heavy chain enhancer region on chromosome 14, resulting in IGH-CRLF2 rearrangement (12, 30). Both rearrangements result in overexpression of full-length CRLF2, which heterodimerizes with the IL7R-alpha subunit to form the TSLPR involved in early B-cell development (31–33). P2RY8-CRLF2 fusions appear to occur more frequently in younger children and in those with DS-ALL, (13) whereas IGH-CRLF2 predominates among adolescents and young adults, particularly those of Hispanic/Latino or Native American ancestry (12, 30, 34, 35). Rarely, activating CRLF2 F232C point mutations, which typically coexist with CRLF2 rearrangements, also lead to CRLF2 deregulation (36). Half of CRLF2-rearranged cases harbor concomitant point mutations in JAK2 or, less commonly, JAK1. The most frequently occurring point mutation is R683G in the JAK2 pseudokinase domain. JAK1 V658F, which is analogous to the JAK2 V617F mutation seen in adult myeloproliferative neoplasms, occurs much less frequently in CRLF2-rearranged Ph-like ALL (7, 10). Sequence mutations in IL7R and SH2B3 have also been identified in a small number of CRLF2-rearranged cases that lack concomitant JAK mutations (7, 37). Less commonly (~10% of patients), some children with detected CRLF2 rearrangements leading to CRLF2 overexpression (usually P2RY8-CRLF2 fusions in children with SR B-ALL) do not have the Ph-like expression signature (37). CRLF2 overexpression in the absence of specific rearrangement detection has been reported in a small percentage of patients (38).

Kinase fusions involving JAK2 or EPOR rearrangements, each representing approximately 5–10% of Ph-like ALL cases, are also associated with activation of JAK/STAT signaling (7, 10). Some groups have reported an increased prevalence of JAK2 fusions among the young adult population compared to children, ~15% vs 5%, respectively (5, 39). JAK2 is a promiscuous 3’ gene in Ph-like ALL with at least 19 different 5’ partner genes reported to date. All JAK2 fusions are in-frame and disrupt the JAK2 pseudokinase domain, which releases the normal auto-inhibition of the kinase domain and results in constitutive JAK/STAT signaling activation (4, 7, 8). Several 5’ partner genes in EPOR rearrangements have been described, each involving the juxtaposition of the EPOR gene to the enhancer region of immunoglobulin heavy (IGH) (most common), κ (IGK), or l (IGL) loci and leading to deregulated expression of a truncated form of EPOR that has been shown to drive Ph-like leukemogenesis and activated JAK/STAT signaling (40). Insertion and truncation of EPOR into the upstream region of LAIR1 or the THADA loci have also been reported in a very small number of patients (40). As with JAK2 fusions, prevalence of EPOR rearrangements rises with increasing age with peak prevalence among young adults (7, 40).

Additional mechanisms leading to JAK/STAT pathway activation beyond the aforementioned kinase or cytokine receptor-activating rearrangements implicate a diverse range of sequence mutations and copy number variations in genes such as JAK1, JAK3, IL7R, SH2B3, IL2RB, and TYK2 (7). These lesions collectively comprise 14% of children compared to 5% and 7.3% of adolescents and older adults respectively (4, 7). Although they lack a kinase-activating rearrangement, these cases often harbor chromosomal rearrangements expressing fusion oncoproteins involving transcription factor genes (EBF1, PAX5, ETV6) and/or epigenetic regulators (CREBBP, SETD2, ASXL1) that merit further study (4).

ABL class alterations

The second most clinically relevant subgroup of Ph-like ALL is ABL class alterations, which account for approximately 10% of cases and are mutually exclusive from the CRLF2 rearrangements and other JAK pathway-activating alterations (7, 27). Prevalence peaks among children with NCI HIGH-RISK B-ALL at 17%, then decreases to about 10% in young and older adults (7, 27, 37). ABL class rearrangements involve 3’ ABL1, ABL2, CSF1R, LYN, PDGFRA, or PDGFRB fusing with multiple 5’ partner genes. Preclinical and clinical studies have shown that ABL class fusion leukemias are sensitive to tyrosine kinase inhibitors, such as imatinib, dasatinib, or ponatinib (7, 8, 10, 41).

Ras pathway mutations

Approximately 4% of Ph-like ALL patients have activating mutations in Ras pathway genes, including KRAS, NRAS, NF1, PTPN11, and CBL (7). Ras pathway mutations are usually subclonal and can occur as the sole detected anomaly or in conjunction with sentinel Ph-like translocations (e.g., CRLF2, ABL class, JAK2, or EPOR fusions) (7, 28). Ras-activating mutations are also commonly found in hyperdiploid, hypodiploid, KTM2A-rearranged, and relapsed ALL and are also often subclonal (42–44). It is not currently known whether these mutations are pathogenic drivers in childhood ALL.

Rare kinase fusions

Other rare fusions involving NTRK3, BLNK, DGKH, FGFR1, PTK2B, FLT3, or TYK2 collectively account for 3% of Ph-like ALL cases (7, 37). The ETV6-NTRK3 fusion, which has been reported in other malignancies such as infantile fibrosarcoma and secretory breast carcinoma, can induce an aggressive ALL with in vitro and in vivo sensitivity to TRK inhibitors (45). Clinical responses to larotrectinib, a pan-TRK inhibitor, in patients with relapsed/refractory B-ALL harboring ETV6-NTRK3 have been described in single case reports (46, 47). Functional modeling of other kinase fusions is important to determine their oncogenic role and identify novel therapeutic targets. As an example, FGFR1 fusions may be targetable with ponatinib (48) or pazopanib (49).

Epidemiology and clinical picture of Ph-like ALL

The prevalence of Ph-like ALL rises with increasing age and varies by ethnicity, gender, and NCI-defined ALL risk group. A recent meta-analysis of 15 studies reported that the pooled prevalence of Ph-like ALL across the age spectrum was 15.4% (50). By age group, the Ph-like subtype comprised 15.6% of B-ALL cases in children 1 to 10 years old, 26.2% in adolescents aged 11 to 20 years old, 25.8% in young adults 21 to 40 years old, and 16.9% in adults older than 40 years (7, 27, 28, 50, 51). Among children and adolescents with B-ALL, Ph-like ALL accounts for 13.6% of NCI SR cases and 22.4% of NCI HIGH-RISK cases (11, 37). In comparison to Ph+ ALL, Ph-like ALL is 3 times more common in the pediatric age group (5). Males are more commonly affected than females across the age spectrum with a male-to-female ratio of 1.5:1 among children and adults (7, 27). Ph-like ALL also has a predilection for patients of Hispanic/Latino or Native American ethnicity, especially among those with CRLF2 rearrangements. This phenomenon has been attributed in part to the increased prevalence of germline Ph-like ALL risk variant in GATA3 (rs3824662) in Hispanic individuals with Native American genetic ancestry (26, 52, 53). Furthermore, Ph-like ALL patients frequently have adverse clinical features with significantly higher rates of hyperleukocytosis at diagnosis, end-induction MRD positivity, and increased risk of treatment failure and relapse (7, 27, 34, 37, 51, 54). The Ph-like ALL gene expression profile itself may independently confer an adverse risk, as shown in some multivariate analyses (27, 51, 55). Patients harboring PDGFRB, JAK2, or EPOR fusions are notoriously associated with more aggressive disease course and frequent high rates of end-induction MRD or even induction failure (7, 56–60).

The inferior survival of patients with Ph-like ALL (compared to non-Ph-like ALL) occurs across the age spectrum, and differential outcomes may further exist within the based upon induction chemotherapy response. Children with NCI high-risk B-ALL treated on the COG AALL0232 phase 3 trial (NCT00075725) with a retrospectively-identified Ph-like expression signature had a 5-year event-free survival (EFS) of 63% versus 86% (p<0.0001) of those with non-Ph-like ALL (55). Importantly, this inferior outcome was detected for patients with Ph-like ALL regardless of the randomized treatment arm assigned, which was in contrast to patients with non-Ph-like ALL with superiority of high-dose methotrexate versus dose-escalating Capizzi-style methotrexate in the first interim maintenance phase (55). More recent analyses of children with NCI SR Ph-like ALL treated on the COG AALL0331 phase 3 trial (NCT00103285) (61) showed statistically inferior outcomes versus those with SR non-Ph-like ALL with 7-year EFS 82.4% and 90.7% (p=0.0022). However, these patients appear to be salvageable at relapse with no difference in overall survival (OS) (93.2% vs 95.8%, p=0.14) between the two groups (11). Recent comprehensive analysis of all children with identified CRLF2-overexpressing ALL treated on COG SR B-ALL trials AALL0331 and AALL0932 (n=77) or high-risk B-ALL trials AALL0232 or AALL1131 (n=244) confirmed reasonable clinical outcomes for children with Ph-like SR B-ALL with 5-year EFS 87.2% and OS 94.5%, although EFS was inferior to that of children with non-Ph-like ALL (38). However, outcomes were quite poor for patients with high-risk B-ALL with 5-year EFS and OS of 56.3% and 75.4%, respectively. Finally, a recent comprehensive retrospective analysis from the Ponte di Legno group comprised of 14 pediatric oncology cooperative groups studied 122 children with Ph-like ABL-class ALL treated with chemotherapy alone (without TKI) and reported high rates of end-induction MRD and poor outcomes despite intensive chemotherapy and often HSCT in first remission. This cohort serves to establish baseline outcomes of children with ABL-class fusions in the pre-TKI era with reported 5-year cumulative incidence of relapse, EFS, and OS of 31.0%, 59.1%, and 76.1%, respectively (60).

Data from the CALGB 10403 phase 2 trial (NCT00558519) have shown marked improvement in outcome for adolescents and young adults (AYA) less than 40 years of age when using a similarly intensive pediatric-inspired regimen, although outcomes for Ph-like ALL AYA patients in that trial remained unfavorable with estimated 3-year EFS of 42% compared to 69% (p=0.008) for those with non-Ph-like ALL (62). Clinical outcomes of patients with Ph-like ALL also worsen with increasing age. Five recent studies focused upon defining the incidence and characteristics of Ph-like ALL occurring in young and older adults with B-ALL. Among 49 adults with Ph-like ALL treated at the MD Anderson Cancer Center (MDACC), the 5-year OS was 23% for Ph-like ALL vs 59% (p=0.006) for patients with non-Ph-like ALL (51). Another study of 194 Ph-like ALL patients from 21 to 86 years old reported 5-year EFS for young adults (21–39 years old), adults (40–59 years), and older adults (60–86 years) of 40.4%, 29.8%, and 18.9%, respectively (27). Among 88 Ph-negative B-ALL adult patients enrolled on the GIMEMA LAL1913 frontline protocol, 28 (31.8%) patients were identified as Ph-like harboring various kinase fusions. These patients had a significantly lower CR (74.1% vs. 91.5%; p=0.044) and EFS (33.5% vs. 66.2%; p=0.005) compared to non-Ph-like patients despite being treated with a pediatric-inspired and MRD-stratified protocol (63). Two additional North American and German studies of adult Ph-like ALL cohorts confirmed the poor outcomes of this population (28, 34). Ph-like ALL patients with concomitant IKZF1 alterations may have further inferior outcomes compared to those without IKZF1 alterations, although these analyses have been limited by small patient numbers (7, 21) (Table 2).

Table 2.

Prevalence, clinical features, and treatment outcomes of pediatric and adult patients with Ph-like ALL reported by study groups.

Study group	Study period	Patient population	Prevalence of Ph-like ALL n (%)	Median age at diagnosis (CI)	Median WBC at diagnosis (CI)	Male,%	Ph-like gene signature detection methods	Kinase alteration detection methods	Clinical trials	Treatment outcome	References
Ponte di Legno ALL Working Group	2000–2018	Pediatric B-ALL	ABL-class (n=122)	9.7 (mean age)	97.7 (mean WBC)	62	LDA	FISH RT-PCR RNA-seq	Frontline ALL trials without TKI	5-year EFS: 59% 5-year OS: 76.1%	den Boer et al Lancet Haematol, 2021
UK ALL	2012–2018	Pediatric ALL with slow response to induction (n=126)	ABL-class (n=21)	9	35	71	N/A	FISH SNP RNA fusion panel	UKALL2011 Chemotherapy (n=8) Imatinib + chemotherapy (n=13)	4-year EFS: 83.9% (TKI) 4-year OS: 83.9% (TKI) 4-year EFS : 37.5% (no TKI) 4-year OS : 75% (no TKI)	Moorman et al BJH 2020
GIMEMA	2014–2018	Adult Ph-negative B-ALL (n=88)	n=28, 31,8%	42 (18–65)	3.34 (0.23–347)	68	LDA	FISH MLPA NGS	GIMEMA LAL1913	2-year EFS: 33.5% 2-year OS: 48.5%	Chiaretti et al Haematologica 2020
MDACC	2010–2012	Adult R/R B-ALL (n=53)	n=12, 22.6%	36 (20–57)	4.2 (1.0–9.3)	67	PAM	WGS RNA-seq	Inotuzumab monotherapy Phase 2 trials	1-year EFS: 33% 1-year OS: 33%	Jabbour et al Blood 2019
AIEOP-BFM	2000–2018	B-ALL (n=3854)	ABL-class (n=46)	NA	NA	63	HC	FISH RT-PCR CGH RNA-seq	Chemotherapy (n=33) Chemotherapy + TKI (n=13)	5-year EFS: 62.9% (TKI) 5-year OS: 75.5% (TKI) 5-year EFS: 47.7% (no TKI) 5-year OS: 70.9% (no TKI)	Cario et al Haematologica 2019
FRALLE CAALL GRAALL EWALL EORTC	NA	Pediatric and adults with B-ALL	ABL-class (n=24)	24 (5–72)	30 (4–570)	67	HC	FISH RT-MLPA RQ-PCR BP-PCR RNA-seq	Chemo + TKI (n=24)	3-year EFS: 55% 3-year OS: 77%	Tanasi et al Blood 2019
CALGB	2007–2012	AYA B-ALL (n=131)	n=41, 31.3%	NA	NA	NA	LDA	LDA for CRLF2 aberration	Pediatric-based chemotherapy Regimen CALGB 10403	3-year EFS: 42% 3-year OS: 63%	Stock et al Blood 2019
COG	2006–2008	SR B-ALL (n=1023)	n=139, 13.6%	NA	15.9 ± 13.7	50	LDA	FISH RT-PCR Sanger sequencing RNA-seq	COG AALL0331 phase 3 for SR B-ALL	7-year EFS: 82% 7-year OS: 93%	Roberts et al Blood 2018
CALGB ECOG MDACC NILG PMCC SWOG City of Hope	NA	Adult B-ALL (n=798)	n=194, 24.3%	40 (21–86)	56.6 (0.2–434)	61	PAMLDA	FISH PCR Sanger sequencing RNA-seq	Multiple trials	5-year EFS: 22.5% 5-year OS:23.8%	Roberts et al J Clin Oncol 2017
GMALL	1999–2005	Adult B-ALL (n=207)	n=26, 13%	31 (16–59)	NA	74	PAM	FISH Q-PCR Sanger sequencing	Multiple trials	5-year DFS: 19% 5-year OS: 22%	Herold et al Haematologica 2017
MDACC	2000–2012	Adult B-ALL (n=148)	n=49, 33.1%	33.5 (15–71)	17 (1–603)	66	PAM LDA	FISH Targeted NGS	Chemotherapy (hyper-CVAD or augmented BFM)	5-year OS: 23%	Jain et al Blood 2017
University Pennsylvania University of Michigan	NA	Adult B-ALL (n=89)	n=18, 20.2%	43 (19–63)	71.1 ± 31.8 (median: 28.6)	69	LDA	FISH PCR RT-PCR Sanger sequencing Archer FusionPlex	Multiple trials	Median survival: 1.6 years	Tasian et al Leukemia 2017
JACLS TCCSG CCLSG KYCCSG	NA	Pediatric B-ALL (n=373)	n=29, 8%	8.8 (1.9–16)	94.3 (0.6–420)	76	GSEA	Multiplex RT-PCR RNA-seq	Multiple trials	5-year EFS: 48.6% 5-year OS: 73.5%	Imamura et al Blood Cancer J 2016
HOVON	1993–2009	Adult B-ALL (n=127)	n=21, 17%	25 (16–59)	NA	NA	HC	RT-PCR	Multiple trials	3-year EFS: ~25%	Boer et al Haematologica 2015
COG SJCRH ECOG CALGB MDACC	2000–2007	Children, AYA with B-ALL (n=1725)	n=264, 15.3%	NA	106	cHR: 62 A: 73 YA: 82	PAM	FISH RT-PCR WES RNA-seq WGS	Multiple trials	5-year EFS: 58 % (cHR), 41% (A), 24% (YA) 5-year OS: 73% (cHR), 66% (ado), 26% (YA)	Roberts et al NEJM 2014
SJCRH	2000–2007	Pediatric B-ALL (n=344)	n=40, 11.6%	5.3 (1.31–8.6)	7.1 (1.7–258.3)	68	PAM	FISH RT-PCR Sanger sequencing RNA-seq WGS	MRD-based risk intensification regimen	5-year EFS: 90% 5-year OS: 92.5%	Roberts et al J Clin Oncol 2014
DCOG COALL	1994–2015	Pediatric B-ALL (n=572)	n=94, 16%	7	44.5	49	HC	FISH CRLF2 mRNA expression by microarray	Multiple trials	5-year CIR: 32%	van der Veer et al Blood 2013
COG	2002–2011	HR B-ALL (n=572)	n=81, 14%	NA	NA	NA	PAM	Kinome sequencing	COG AALL0232 phase 3 trial for HR B-ALL	5-year EFS: 62.6%	Loh et al Blood 2013
COG	2000–2003	Pediatric B-ALL (n=221)	n=68, 31%	NA	NA	NA	PAM	SNP PCR	COG P9906	5-year EFS: 26%	Mullighan et al NEJM 2009
DCOG COALL	1990–2004	Pediatric B-ALL (n=154)	n=30, 19%	NA	NA	NA	HC	RT-PCR	Multiple trials	5-year DFS: 60% (COALL) 5-year DFS: 57% (DCOG)	Den Boer et al Lancet Oncol 2009

AIEOP-BFM = Associazione Italiana di Ematologia e Oncologia Pediatrica and Berlin-Frankfurt-Münster cooperative groups, AYA = adolescents and young adults, BP-PCR = breakpoint-specific multiplex polymerase chain reaction, CAALL = French Protocol for the Treatment of ALL in Children and Adolescents, CALGB = Cancer and Leukemia Group B, CCLSG = Children’s Cancer and Leukemia Study Group, CGH = comparative genomic hybridization, cHR = childhood high-risk, CIR = cumulative incidence of relapse, COALL = Childhood Oncology Acute Lymphoblastic Leukemia, COG = Children’s Oncology Group, CRLF2 = cytokine receptor like factor 2, DCOG = Dutch Childhood Oncology Group, DFS = disease-free survival, EFS = event-free survival, ECOG = Eastern Cooperative Oncology Group, EORTC = European Organisation for Research and Treatment of Cancer, EWALL = European Working Group on Adult ALL, FISH = fluorescence in situ hybridization, FRALLE = French Group for Childhood ALL, GMALL = German Multicenter Study Group for Adult ALL, GRAALL = French-Swiss-Belgian Group for Research on Adult Acute Lymphoblastic Leukemia, GSEA = gene-set enrichment analysis, HC = hierarchical clustering, HOVON = Hemato-Oncologie voor Volwassenen Nederland, JACLS = Japanese Association of Childhood Leukemia Study, KYCCSG = Kyushu-Yamaguchi Children’s Cancer Study Group, LDA = low-density array, MDACC = MD Anderson Cancer Center, MLPA = multiplex ligation-dependent probe amplification, MRD = minimal residual disease, NILG = Northern Italy Leukemia Group, NA = not available, OS = overall survival, PAM = prediction analysis of microarrays, PMCC = Princess Margaret Cancer Centre, RNA-seq = RNA sequencing, RQ-PCR = real-time quantitative polymerase chain reaction, R/R B-ALL = relapsed/refractory B-acute lymphoblastic leukemia, RT-PCR = reverse transcription-polymerase chain reaction, SJCRH = St Jude Children’s Research Hospital, SNP = single nucleotide polymorphism, SR = standard-risk, SWOG = Southwestern Oncology Group, TCCSG = Tokyo Children’s Cancer Study Group, TKI = tyrosine kinase inhibitor, WES = whole-exome sequencing, WGS = whole genome sequencing.

Clinical outcomes among Ph-like ALL patients may also vary according to their underlying kinase-activating alterations, although larger prospective studies are required to confirm these observations. Roberts and colleagues reported that there are significant differences in 5-year EFS among different Ph-like genomic subgroups, with JAK2/EPOR-rearranged cases (26.1%) and CRLF2-rearranged JAK2-mutant cases (38.8%) having dismal prognosis, in contrast to cases harboring other JAK/STAT (68.3%) or Ras pathway mutations (85.7%) having a more favorable outcome (p<0.0001)(7). The recent Ponte di Legno ALL working group study also highlights the variability in MRD response and clinical outcomes in their cohort of ABL-class fusions depending on the specific fusion subtype. Notably, children with PDGFRB and ABL2 fusions appear more likely to have MRD levels greater than 10⁻² at the end of induction and inferior EFS compared to their counterparts with ABL1 or the uncommon CSF1R rearrangements (60).

Diagnostic modalities and clinical workflow algorithms for Ph-like ALL

The heterogeneous spectrum of kinase-activating alterations, cryptic nature of these genetic aberrations by conventional cytogenetic analysis, and complexity of required testing has rendered Ph-like ALL quite challenging to diagnose via clinical laboratory assays. Several cooperative oncology consortia have adopted different screening strategies largely based on the patient population size to be tested, development of more rapid next-generation sequencing platforms, and availability of therapeutic clinical trials (6).

Clinical diagnosis of Ph-like ALL has involved assessment of the pathognomonic gene expression signature (used by some, but not all, groups) and detection of targetable kinase-activating alterations. Gene expression profiling, the utilized discovery modality in initial European and North American studies (1, 2), is not readily available in the clinic and has now been largely replaced by the TaqMan LDA microfluidic card measuring the expression of 8 or 15-gene panels (IGJ, SPATS2L, MUC4, CRLF2, CA6, NRXN3, BMPR1B, GPR110, CHN2, SEMA6A, PON2, SLC2A5, S100Z, TP53INP1, IFITM1) now used by the COG to determine the Ph-like ALL signature (6, 55, 64). An integrated score between 0 and 1 is generated from the 8 or 15-gene assay with a score ≥0.5 considered positive for the Ph-like gene signature (37). Higher LDA score (e.g., >0.7) typically suggests an underlying kinase fusion (6). This LDA-based approach has provided a rapid and cost-effective screening modality for some groups to identify patients with probable Ph-like ALL who require further detailed genomic characterization to identify targetable kinase-activating alterations and to minimize unnecessary downstream testing for LDA-negative patients. Aside from direct detection of CRLF2 overexpression and P2RY8-CRLF2 fusions on the LDA card, this modality is intended only as a screening tool and does not detect other Ph-like kinase-activating lesions. The LDA card is also capable of identifying and ‘ruling out’ patients with Ph+ ALL and ETV6-RUNX1 ALL who do not require additional Ph-like testing (65).

To enable detection of kinase-activating alterations (the most clinically relevant endpoint in the Ph-like ALL diagnostic work-up), several commercial, research-level, and clinical next-generation sequencing (NGS) platforms have been developed and are replacing prior multiplexed clinical RT-PCR panels (6, 66, 67). The ArcherDx FusionPlex Heme panel uses anchored multiplex PCR-based enrichment with the ability to detect novel fusions involving 87 genes associated with hematologic malignancies (6). The FoundationOne Heme panel is a targeted combined DNA and RNA sequencing method capable of fusion and mutation detection in >400 cancer-related genes (68). Of particular interest, transcriptomic/RNA sequencing (RNA-seq) represents a powerful tool for comprehensive fusion and mutation detection in addition to identifying the Ph-like GEP by hierarchical clustering (6) and is becoming more clinically available. Indeed, RNA-seq is the only single platform capable of fulfilling the two essential aspects of Ph-like ALL’s diagnosis, but is currently not considered time- or cost-effective for routine analysis of all patients. Clinical fusion and NGS assays are relatively more cost-effective, but still require a relatively long turnaround time (between 2–4 weeks) prior to clinical result reporting.

More rapid testing assays using conventional methods such as fluorescence in situ hybridization (FISH) and flow cytometry still retain clinical relevance for the diagnosis of Ph-like ALL. For example, dual color break-apart FISH probes are now commercially available for the canonical Ph-like 3’ kinase genes ABL1, ABL2, PDGFRB (will also detect CSF1R disruption at 5q32), CRLF2, JAK2, and EPOR with results typically delivered within 48 hours. Since the vast majority of these kinase fusions are exclusively seen with the Ph-like phenotype, an abnormal FISH result might be sufficient to start tyrosine kinase inhibitor (TKI) therapy while waiting for confirmatory molecular testing. The COG has expanded its routine FISH panel to include ABL class FISH probes in its HIGH-RISK B-ALL AALL1732 phase 3 trial (NCT03959085) and its Ph+ ALL AALL1631 phase 3 trial (NCT03007147) in order to facilitate earlier introduction of TKI during induction for patients with identified ABL class alterations. Increased flow cytometric staining of TSLPR also suggests underlying CRLF2 rearrangement, which is known to occur in over half of Ph-like ALL cases (29). Flow cytometry availability in most institutions’ diagnostic laboratories and rapid result return within 24 hours makes TSLPR immunophenotyping a compelling additional Ph-like screening assay.

In summary, successful identification of patients with Ph-like ALL will most likely benefit from a combined approach of cytogenetic, FISH, and molecular analysis via fusion and NGS testing given the known genetic heterogeneity of this leukemia subtype and ongoing new discovery. Pragmatic clinical screening algorithms will need to be personalized on one’s available resources, patient volume, and clinical goals. A suggested diagnostic algorithm for Ph-like ALL is shown in Figure 2 (6).

Figure 2. Suggested clinical screening algorithm for Ph-like ALL.

CRLF2 = cytokine receptor-like factor 2, LDA = low-density array, FISH = fluorescence in situ hybridization, NGS = next-generation sequencing, PCR = polymerase chain reaction, RNA = ribonucleic acid, RT-PCR = reverse-transcription polymerase chain reaction, TSLPR = thymic stromal lymphopoietin receptor. Adapted from Harvey & Tasian Blood Adv 2020.

Precision medicine trials in Ph-like ALL

Targeted therapies

The recently characterized genomic landscape of Ph-like ALL has uncovered a myriad of kinase-activating alterations that expand the treatment paradigm of molecularly-targeted therapies in ALL and leverage the success story of TKI incorporation for children and adults with Ph+ ALL (15, 69–72) (Table 3). Despite their heterogeneity, Ph-like ALL-associated genetic alterations commonly converge to activate JAK/STAT, ABL, or Ras signaling pathways. Extensive in vitro and in vivo data have provided compelling evidence to incorporate relevant ABL/PDGFR- or JAK-directed TKIs in combination with chemotherapy to improve these patients’ poor outcomes. The efficacy of such precision medicine approaches is currently being prospectively assessed in several clinical trials for children and adults with Ph-like ALL harboring ABL class alterations or JAK/STAT pathway gene lesions.

Table 3.

Outcomes of children and adolescents/young adults with Ph+ ALL treated with chemotherapy and tyrosine kinase inhibitors treated on cooperative group clinical trials.

	COG AALL0031	EsPhALL2004	EsPhALL2010	COG AALL0622	COG AALL1122	CCCG-ALL-2015
Phase	3	2	2	2	2	3
TKI	imatinib 340 mg/m2	imatinib 300 mg/m2	imatinib 300 mg/m2	dasatinib 60 mg/m2	dasatinib 60 mg/m2	imatinib 300 mg/m2 vs dasatinib 80 mg/m2
Years	2002–2006	2004–2009	2010–2014	2008–2012	2012–2014	2015–2018
Patients (n)	91	160	155	60	106	97 (imatinib) 92 (dasatinib)
HSCT in CR1	25%	83%	38%	32%	14%	2%
5y EFS	71% (cohort 5)	60%	57%	60%	55%	4y EFS49% (imatinib) 4y EFS: 71% (dasatinib) p=0.005
5y OS	81% (cohort 5)	72%	72%	86%	82%	4y EFS: 69% (imatinib) 4y EFS: 88% (dasatinib) p=0.04

CCCG = Chinese Children’s Cancer Group, COG = Children’s Oncology Group, CR1, complete first remission, EFS = event-free survival, EsPhALL = European Ph+ ALL working group, HSCT = hematopoietic stem cell transplantation (allogeneic), OS = overall survival, 4y = 4-year, 5y = 5-year.

ABL class alterations phenocopy BCR-ABL1 and exhibit exquisite sensitivity to imatinib and dasatinib in preclinical models (7, 8, 73). There is increasing anecdotal evidence demonstrating that the addition of imatinib or dasatinib monotherapy or in combination with chemotherapy can induce remission in patients with relapsed/refractory ALL with ABL class alterations (7, 56–58, 74, 75). Durable remissions and favorable outcomes have been reported in recent series of Ph-like ABL-class patients treated with imatinib or dasatinib in combination with chemotherapy compared to historical controls (76–78). Based on the increasing clinical experience in Ph-like ALL and the robust demonstration of safety and efficacy in children with Ph+ ALL, the COG AALL1131 phase 3 trial (NCT02883049) was amended in 2016 to add a dedicated treatment arm adding dasatinib to post-induction chemotherapy for Ph-like ALL patients with identified ABL class alterations. AALL1131 was closed to accrual in August 2019, and complete clinical results from this study are not yet available. The international phase 3 EsPhALL2017/COG AALL1631 phase 3 trial (NCT03007147), which is randomizing two different chemotherapy backbones in combination with imatinib for children with Ph+ ALL, will extend eligibility to include patients with Ph-like ABL class fusions in 2021. The St. Jude Children’s Research Hospital (SJCRH) Total Therapy XVII ALL (NCT03117751) protocol was activated in early 2017 and incorporates dasatinib beginning in induction therapy for those identified with an ABL class fusion by RNA-seq within 2 weeks of diagnosis (79). An MDACC phase 2 trial for adults with relapsed/refractory Ph-like ALL and ABL class fusions (NCT02420717) also combined dasatinib with the intensive hyper-CVAD chemotherapy backbone. Results have not yet been reported for these studies.

The largest class of Ph-like kinase-activating alterations constitutes those deregulating JAK/STAT signaling, making this pathway a major potential therapeutic vulnerability in Ph-like ALL, although JAK inhibitors have been less well studied in patients with ALL to date. Preclinical studies of engineered Ba/F3 cells and patient-derived xenograft (PDX) models harboring a diverse range of JAK/STAT pathway lesions (CRLF2 rearrangements with or without JAK mutations, JAK2 fusions, EPOR fusions, sequence mutations of IL7R and/or SH2B3) have shown in vitro and in vivo activity to different JAK inhibitors (7, 8, 29, 59, 73, 80–84). Potent, but sometimes differential, preclinical activity of the JAK1/2 inhibitor ruxolitinib has been reported in Ph-like ALL models with CRLF2 fusions or JAK2 fusions, which may be influenced by the level of JAK pathway oncogene addiction or, potentially, by paradoxical JAK2 hyperphosphorylation with prolonged treatment (51, 80, 84). Subsequent preclinical studies further demonstrated enhanced activity of combinatorial treatment with JAK and PI3K/mTOR inhibitors in Ph-like ALL cell lines and PDX models (81, 82). Investigating dual pathway inhibition seems quite relevant for Ph-like ALL, as upregulation of alternative signaling pathways is a known mechanism of resistance to single-targeting agents (85–87). To this end, a recent preclinical study showed that CRLF2-rearranged Ph-like ALL cells mediate a complex “BCR-like” signaling characterized by activated SRC family kinase and downstream signaling in the absence of surface μ-heavy chain expression, which may mediate resistance to ruxolitinib monotherapy, but could be overcome by multi-TKI and/or dexamethasone combination (87).

Currently, ruxolitinib is being assessed prospectively in several clinical trials for patients with JAK/STAT pathway-mutant Ph-like ALL. The COG AALL1521 phase 2 trial (NCT02723994) is investigating the efficacy of incorporating ruxolitinib with post-induction chemotherapy for children, adolescents, and young adults with HIGH-RISK Ph-like ALL harboring JAK/STAT pathway lesions (88). In this study, patients are stratified into 4 different cohorts based on their underlying Ph-like genetic lesions and by end-induction MRD status to delineate potential differential efficacy for each subset. A soon-to-open phase 1 trial will also assess the safety and tolerability of ruxolitinib addition to chemotherapy specifically in a Ph-like ALL AYA population ages 18–39 years (NCT03571321). The two previously discussed SJRCH and MDACC trials also have a ruxolitinib treatment arm in combination with chemotherapy for patients with de novo or relapsed JAK-mutant Ph-like ALL, respectively.

The enriched prevalence of IKZF1 deletions in Ph-like ALL opens up another potential therapeutic avenue for this HIGH-RISK patient population, although it is not yet known how these alterations might best be targeted. As above, IKZF1 alterations are known to mediate aberrant stromal adhesion and therapy resistance in murine models of Ph+ ALL, and it is plausible that such effects could be reversed by retinoic acid compounds or focal adhesion kinase (FAK) inhibitors when combined with TKIs (18, 19).

Targeted inhibitors of apoptotic proteins are promising therapeutic candidates for children and adults with relapsed/refractory acute leukemias and are under early-phase clinical investigation in pediatric trials. Venetoclax, an orally-administered selective inhibitor of the anti-apoptotic BCL-2 protein, has shown encouraging efficacy when combined with chemotherapy in 25 children with relapsed/refractory ALL with an overall response rate (ORR) of 56%, including 4 complete remission (CR), 4 CR with incomplete marrow recovery (CRi), and 1 CR without platelet recovery (CRp) (89). The combination of venetoclax with dasatinib or ponatinib appears highly synergistic in Ph+ ALL cell lines and PDX models with demonstrated inhibition of LYN signaling and prevention of downstream MCL-1 upregulation (90). A recent study demonstrated that dual BCL-2 and MCL-1 inhibition exhibits potent anti-leukemic activity in Ph+ ALL and CRLF2-rearranged Ph-like ALL PDX models (91), providing further rationale for potential investigation of BH3-mimetic inhibitors in clinical trials for patients with Ph+ and Ph-like ALL.

Hematopoietic stem cell transplantation

The role of hematopoietic stem cell transplantation (HSCT) in the care of patients with Ph-like ALL remains unclear in the TKI era (92). Earlier data demonstrated definitive improvement in EFS and OS of children with Ph+ ALL with imatinib or dasatinib addition to chemotherapy, which also eliminated need for HSCT in most patients (15, 69, 93, 94). Mirroring this Ph+ ALL experience, it is plausible that TKI addition to chemotherapy could be similarly successful for patients with Ph-like ABL-class ALL.

A single-center study recently reported comparable outcomes between children with Ph-like ALL and non-Ph-like ALL (5-year EFS 90.0% vs 88.4%, p=0.41, respectively) using MRD-directed therapy intensification for relevant patients (95). Consequently, a significant higher proportion of patients with Ph-like ALL underwent HSCT in first remission due to end-induction MRD positivity (95), which is known to occur in two-thirds of children with Ph-like ALL (55). These results demonstrate the therapeutic efficacy of HSCT in patients with Ph-like ALL and suggest that MRD is also an important outcome predictor in this patient population. Conversely, HSCT in first complete remission did not improve the EFS and OS of children with Ph-like ABL-class ALL, and HSCT-related mortality was particularly high (17%) (60). Another single-center study reported that adult patients with Ph-like ALL fare comparably poorly even when they achieve post-remission MRD negativity (median OS for MRD+ group 23.0 months vs MRD-group 26.2 months; p=0.318). Decisions for HSCT in first CR for adult patients may be challenged in the current era where pediatric-inspired chemotherapy regimens and access to frontline immunotherapy trials foster high hopes for inducing remission, deepening MRD response, and improving long-term survival (62, 96–99).

Antibody-based and cellular immunotherapy

Major advances in immunotherapy during the past decade have revolutionized the landscape of relapsed leukemia therapy. The CD19xCD3 bispecific T-cell engager antibody blinatumomab, anti-CD22 antibody-drug conjugate inotuzumab ozogamicin, and CD19-redirected chimeric antigen receptor (CAR)-modified T-cell immunotherapy tisagenlecleucel have consecutively received FDA approval for patients with relapsed/refractory B-ALL based on several paradigm-shifting trials (98, 100–102). Although the above trials did not specifically screen for the Ph-like ALL subtype, it is presumed that a reasonable proportion of these relapsed/refractory, heavily pretreated patients were Ph-like given their known high rates of chemoresistance and relapse. The randomized TOWER phase 3 trial (NCT02013167) showed that treatment with blinatumomab resulted in significantly higher remission rates and longer survival compared to standard chemotherapy amongst adults with relapsed/refractory Ph-negative B-ALL (98). Subsequent results from the BLAST phase 2 trial (NCT01207388) demonstrated that the majority (78%) of MRD-positive adult B-ALL patients in first or later CR achieved a complete MRD response following 1 cycle of blinatumomab, which was associated with better outcomes than MRD non-responders (100). Moreover, single-agent blinatumomab had strong anti-leukemic activity among adults with relapsed Ph+ ALL (103) and was associated with favorable treatment outcomes when compared to an external cohort receiving standard chemotherapy in a propensity score analysis (104). Favorable safety profiles and anti-leukemic activity with blinatumomab monotherapy were also observed in a heavily pre-treated relapsed/refractory pediatric B-ALL cohort (105, 106).

A recent retrospective analysis of 42 adults with R/R B-ALL who had available material for genomic analysis and were treated with blinatumomab monotherapy also showed that Ph-like ALL patients had a high response rate to blinatumomab (16/23; 70%), especially those with Ph-like ALL harboring CRLF2 rearrangements (12/16; 75%) and non-CRLF2 rearrangements (4/7; 57%) (107). Small cases series have also reported anecdotal safety and efficacy of combining blinatumomab and dasatinib in small numbers of patients with relapsed/refractory Ph+ ALL or Ph-like ABL-class ALL (108–111). The SWOG S1318 phase 2 trial (NCT02143414) is currently comparing the efficacy of blinatumomab with combination chemotherapy versus blinatumomab and dasatinib in older adults with ABL-driven leukemias. The GIMEMA D-ALBA LAL2116 phase 2 trial (NCT02744768) is also currently assessing rates of molecular remission after 2 cycles of blinatumomab and dasatinib consolidation therapy in adult patients with newly-diagnosed Ph+ ALL. In an interim analysis with very early follow-up, 60% of study patients achieved a molecular response at the primary endpoint of approximately five months (three months of dasatinib and glucocorticoid chemotherapy followed by two cycles of blinatumomab). Additional improvement to 81% of molecular response after four cycles of blinatumomab was also reported with 1-year disease-free and overall survival of 88% and 95%, respectively, and 24 of the 63 enrolled patients received allogeneic HSCT (112). Longer-term follow-up is needed to determine if these early favorable outcomes will be sustained. Concerns have been raised with respect to potential antagonism of blinatumomab-dasatinib combination, as dasatinib has been shown to inhibit T cell function and could potentially abrogate the desired anti-leukemic activity of blinatumomab that requires endogenous T cell engagement (109, 113, 114). Correlative functional assays that comprehensively assess potential immunomodulatory effects of dasatinib upon immune cells will provide critical data for future trial design.

In a retrospective analysis of 53 adult patients with relapsed/refractory B-ALL treated with inotuzumab as salvage therapy, 12 patients identified as having the Ph-like subtype had an ORR of 58%, including 3 with CR (25%) and 4 with CR (33%) and partial hematologic recovery (CRh) (115). Five of the 7 (71%) Ph-like patients with inotuzumab-induced CR achieved MRD negativity (115). Inotuzumab at the FDA-approved fractionated adult dosing of 1.8 mg/m² per cycle also induced impressive CR response rates among heavily-pretreated children with relapsed/refractory B-ALL (116, 117). In a retrospective study of 51 pediatric R/R B-ALL patients who received inotuzumab via a compassionate use program, the overall CR rate was 67%; 3 of 4 Ph-like ALL achieved CR/CRi, one of whom was MRD-negative (117).

Two recent studies showed that CD123 is more commonly expressed in adults with Ph+ ALL (118) and Ph-like ALL (119) compared to Ph-negative B-ALL and that CD123 may represent an additional therapeutic target for these high-risk leukemias. Anti-CD123 targeted immunotherapies (e.g., IMGN-632, flotetuzumab) have shown promising activity in preclinical models of acute myeloid leukemia (AML) (120–122) and are under current clinical investigation in early-phase trials for patients with relapsed/refractory AML.

Finally, anecdotal reports of CD19 chimeric antigen receptor (CAR) T cell-induced remission in patients with relapsed/refractory Ph+ and Ph-like ALL have been reported (102). The current COG AALL1721 phase 2 trial (NCT03876769) assessing the efficacy of tisagenlecleucel in patients with newly-diagnosed HIGH-RISK B-ALL with end-consolidation MRD positivity excludes patients treated with kinase inhibitors, however. A planned phase 1 trial based upon promising preclinical data (123) will specifically investigate the clinical safety and preliminary efficacy of TSLPR-redirected CAR T cell immunotherapy in children, adolescents, and young adults with CRLF2-rearranged/overexpressing leukemias, including Ph-like ALL and Down Syndrome-associated ALL.

Conclusions and future perspective

Ph-like ALL is now known to be a prevalent subtype of B-ALL defined by its kinase-activated gene expression signature and associated genetic alterations. Children and adults with Ph-like ALL have historically experienced high relapse rates and inferior clinical outcomes despite best-available conventional chemotherapy, and their outcomes are currently far inferior to those of patients with Ph+ ALL with contemporary TKI-based therapy. Compelling evidence now exists from an extensive preclinical body of work for incorporation of relevant TKIs in combination with chemotherapy also for patients with Ph-like ALL, although results from clinical trials testing these strategies are not yet known. Although Ph-like ALL is thrice as common than Ph+ ALL in children and adolescents, its genetic heterogeneity with >70 fusions identified to date (6) represents a limiting factor in designing appropriately statistically-powered, randomized controlled trials to assess potential TKI efficacy within the major ABL class and CRLF2/JAK pathway-mutant subsets. As some patients with Ph-like ALL (particularly those with EBF1-PDGFRB or JAK2 fusions) are at high risk of induction chemotherapy failure, future efforts must focus upon swift identification of these genetic alterations and TKI addition early in induction therapy. Such strategy has proven successful in children with Ph+ ALL with superior CR rates and MRD negativity with TKI addition mid-induction versus at the beginning of consolidation (15, 69). Several immunotherapy modalities have now also demonstrated exciting efficacy in children with relapsed/refractory B-ALL. Ongoing and future clinical trials may help to elucidate the potential of such approaches (as monotherapy or combined with TKIs) more specifically in patients with Ph-like ALL.

International collaboration in designing the next generation of Ph-like ALL studies will expedite systematic study of novel treatment strategies that may improve clinical outcomes for this high risk patient population (124). Development and implementation of standardized clinical screening strategies among cooperative groups for rapid identification of patients with Ph-like ALL will also be essential for successful and efficient conduction of clinical trials. In parallel, future investigations in Ph-like ALL should also focus on investigating potential mechanisms of TKI resistance, as has been observed in patients with BCR-ABL1-driven chronic myeloid leukemia or Ph+ ALL with emergence of drug-resistant kinase domain mutations after long-term imatinib exposure (125–129). Similar mutations have also been identified via in vitro saturation mutagenesis screens of Ph-like ALL with EBF1-PDGFRB fusions, and resistance mutations likely facilitating clinical relapse in patients with Ph-like ALL have now been reported (130, 131).

In summary, Ph-like ALL illustrates a paradigm of genomic discovery translation into targeted therapeutic approaches and presents an exciting opportunity for new precision medicine opportunities that aim to decrease relapse and improve long-term survival for patients with these high-risk leukemias across the age spectrum.