Acute lymphoblastic leukemia (ALL) is an aggressive cancer that is most common among children. ALL comprises multiple molecular subtypes and is broadly divided into B-cell acute lymphoblastic leukemia (B-ALL) and T-cell acute lymphoblastic leukemia, each harboring a distinct constellation of somatic genetic alterations. Recent advances in genetic testing provides a comprehensive understanding of these alterations and an opportunity to refine the current National Cancer Institute (NCI) criteria of age, white blood cell (WBC) count, and minimal residual disease (MRD). In this educational theme article, we present a summary of different genetic subtypes of B-ALL and discuss the impact of these subtypes on current risk criteria and treatment responses. The summary provided here is based on a recent review by Lee et al.1, published in Cancers, on the genetic subtypes of ALL and how these interact with clinical features, treatment responses, and outcomes.
Genetic subtypes of B-ALL
Using the next-generation sequencing technology RNA-Seq, investigators from the Malaysia-Singapore (Ma-Spore) ALL study group defined 20 genetic subtypes of B-ALL, each with their own distinct clinical characteristics, treatment responses, and outcomes. Based on this, Lee et al.1 grouped them into three genetic risk groups: favorable- (FRG), intermediate- (IRG), and high-risk genetics (HRG) (Table 1).
Table 1. B-ALL subtypes based on Malaysia-Singapore 2003/2010 trials*
B-ALL subtype
|
Proportion of B-ALL, %
|
MRD-negative at EOI, %
|
MRD-negative at EOC, %
|
5-year CIR, % (range)
|
5-year OS, % (range)
|
Favorable-risk genetics
|
ETV6-RUNX1
|
20
|
76
|
98
|
5.2 (1.3–13.0)
|
100.0
|
Hyperdiploidy
|
24
|
54
|
92
|
5.5 (1.7–12.6)
|
98.8 (91.8–99.8)
|
Intermediate-risk genetics
|
TCF3-PBX1
|
5
|
58
|
94
|
5.6 (0.3–23.1)
|
94.4 (66.6–99.2)
|
DUX4
|
14
|
22
|
82
|
8.9 (2.8–19.5)
|
97.8 (85.3–99.7)
|
ETV6-RUNX1-like
|
2
|
57
|
83
|
12.7 (0.5–45.3)
|
88.9 (43.3–98.4)
|
ZNF384
|
5
|
18
|
77
|
6.3 (0.4–25.5)
|
93.3 (61.3–99.0)
|
ZNF384-like
|
1
|
50
|
50
|
—
|
—
|
NUTM1
|
1
|
100
|
100
|
0.0
|
100.0
|
PAX5alt
|
10
|
39
|
89
|
18.1 (6.3–34.7)
|
92.8 (73.7–98.2)
|
PAX5-P80R
|
1
|
0
|
100
|
—
|
—
|
B-others
|
7
|
45
|
89
|
20.7 (7.3–39.0)
|
94.1 (65.0–99.1)
|
IGH-CEBPE
|
<1
|
0
|
100
|
—
|
—
|
High-risk genetics
|
Ph (BCR-ABL1)
|
2
|
44
|
83
|
37.5 (7.2–69.4)
|
75.0 (31.5–93.1)
|
Ph-like
(BCR-ABL-like)
|
2
|
12
|
60
|
37.5 (6.9–69.8)
|
75.0 (31.5–93.1)
|
MLL (KMT2A)
|
3
|
11
|
43
|
54.3 (16.7–81.2)
|
64.8 (25.3–87.2)
|
Hypodiploidy
|
1
|
0
|
0
|
50.0 (0.0–96.0)
|
50.0 (0.6–91.0)
|
Near-haploidy
|
1
|
100
|
100
|
50.0 (0.0–96.0)
|
50.0 (0.6–91.0)
|
MEF2D
|
1
|
100
|
100
|
0.0
|
100.0
|
HLF-rearranged
|
<1
|
0
|
NA
|
—
|
—
|
CRLF2
|
3
|
22
|
88
|
20.0 (2.6–49.2)
|
59.1 (16.0–86.0)
|
B-ALL, B-cell acute lymphoblastic leukemia; CIR, cumulative risk of relapse; EOC, end of consolidation; EOI, end of induction; MRD, minimal residual disease; OS, overall survival. *Adapted from Lee et al.1
|
Favorable-risk genetics group
- This group is defined as subtypes with <10% cumulative incidence of relapse (CIR) and deemed eligible for treatment de-intensification.
- The group comprises two major subtypes: ETV6-RUNX1 and hyperdiploidy (>50 chromosomes).
- In the Ma-Spore cohort, patients with FRG had good outcomes, with overall survival (OS) rates of 100% and 98.8% for the ETV6-RUNX1 and hyperdiploidy subtypes, respectively (Table 1), and a combined 5-year event free survival (EFS) rate of 94%.
- The subtypes differed in their MRD response: ETV6-RUNX1 showed more rapid MRD clearance than hyperdiploidy (76% vs 54% at end of induction [EOI]), however both subtypes had high rates of MRD negativity at end of consolidation (EOC; 98% vs 92%).
- The FRG group had improved 5-year EFS even for patients with low EOI MRD positivity (<0.1%), and therefore has been considered to qualify for treatment de-intensification in the ongoing MS2020 study; treatment intensification is indicated only for FRG with high EOI MRD (≥1%).
Intermediate-risk genetics group
- This is the most diverse group, comprising more than ten different genetic subtypes (Table 1).
- Most subtypes have a 5-year CIR of 10–20%; the exceptions being TCF3-PBX1, DUX4, and ZNF384-rearranged, which have a CIR <10% but continue to be classified as IRG.
- The remaining subtypes are grouped based on gene expression profile (GEP), and despite having similar GEPs, each subtype is molecularly diverse with different fusion partners.
- In the Ma-Spore trials, the EOI MRD remained the most significant prognostic factor for EFS (p = 7.2 × 10-4), with high-risk patients with EOI MRD associated with poorer outcomes (5-year EFS, 40%).
PBX1 fusions including TCF3-PBX1
- This fusion is generally associated with a high WBC count (median, 56,000/µL).
- PBX1 rarely fuses with another genetic partner and has a similar GEP as TCF3-PBX1.
- In the Ma-Spore trials, outcomes for those with TCF3-PBX1 considerably improved with more intensive chemotherapy.
- Patients achieved MRD negativity in 53% and 94% of cases by EOI and EOC, respectively.
- De-intensification may be an option for TCF3-PBX1 EOI MRD-negative patients; however, it poses potential risks due to previous poor outcomes associated with less intensive regimens.
ZNF384-rearranged
- This subtype is found in 1–6% and 5–5% of childhood and adult B-ALL cases, respectively.
- It is more frequent in older children, with a median age of 6.8 years and median WBC of 37,000/µL in the Ma-Spore trials.
- In the Ma-Spore trials, MRD response was slow with only 18% of cases MRD-negative at EOI. However, responses improved and 77% were MRD-negative by EOC.
- Interestingly, relapses for patients with ZNF384 rearranged with TCF3 or TAF15 ensued several years after treatment completion.
PAX5
- This comprises two genetic subtypes: PAX5-P80R and PAXalt.
- ALL with PAXalt has a higher WBC (>50,000/µL), which is increased in patients with more than one PAX5 aberration.
- In the Ma-Spore trials, 39% of patients with PAXalt were MRD-negative, but poor outcomes were observed in patients with both PAXalt and IKZF1del.
ETV6-RUNX1-like
- Although comparable to ETV6-RUNX1 in terms of median age, WBC, and GEP, the ETV6-RUNX1-like subtype is associated with poorer outcomes.
- IKZF1del frequently occurs in patients with ETV6-RUNX1-like, but whether IKZF1del drove adverse outcomes in the Ma-Spore cohort was not clear due to the small numbers.
- In this cohort, patients with the ETV6-RUNX1-like subtype showed good CIR and OS, and therefore may benefit from higher-intensity therapy.
DUX4
- This subtype is characterized by the rearrangement of the DUX4 gene to the immunoglobulin heavy chain locus, resulting in a distinctive GEP with exceedingly high expression of DUX4.
- ERG deletion occurs almost exclusively in the DUX4 subtype and is associated with better MRD responses and outcomes, while IKZF1del shows no adverse effect on DUX4.
- Although patients with DUX4 usually demonstrate favorable EFS and OS that exceeds 90%, whether de-intensifying therapy in this group is appropriate remains unclear.
- Polymerase chain reaction-based MRD may overestimate the potential of relapse in patients with DUX4 compared with flow cytometry.
High-risk genetics group
- The ALL subtypes BCR-ABL1, KMT2A-r, MEF2D, HLF-r, and CRLF2/BCR-ABL1-like are specifically more high-risk (Table 1).
- In the Ma-Spore trials, the 5-year CIR was high, with a range of 20–55%.
- Overall, MRD remained prognostically important in the HRG group of the Ma-Spore cohort (p = 0.015 for EFS; p = 0.28 for OS).
Philadelphia (BCR-ABL1)-positive
- Philadelphia chromosome (Ph) ALL is defined by the translocation t(9;22), and the incidence of Ph+ ALL increases with age.
- The 5-year survival has significantly improved for this subtype with the use of tyrosine kinase inhibitors, although treatment-related toxicity also increased in Ma Spore ALL 2010 cohort.
- In the ongoing Ma-Spore ALL 2020 cohort, patients with EOI MRD-negative Ph+ ALL will continue on tyrosine kinase inhibitors plus reduced intensity chemotherapy, while EOI MRD-positive patients will qualify for chimeric antigen receptor T-cell therapy or hematopoietic stell cell transplantation in first complete remission.
BCR-ABL-like (Ph-like) with or without CRLF2 rearrangements
- This subtype is characterized by diverse genetic alterations, with a similar transcriptional profile to Ph+ ALL but without BCR-ALB1 fusion.
- Genetic alterations can be targeted with ABL or JAK inhibitors.
- High prevalence in adolescent and young adult patients is associated with higher rates of treatment failure along with high EOI and EOC MRD, similar to Ph+ ALL.
- Patients with CRLF2 rearrangements, particularly CRLF2-PY2R8, have inferior outcomes (5-year EFS, 57% vs 83% for all other B-ALL) and significantly increased CIR (43% vs 14% for other B-ALL).
- CRLF2 is overexpressed and occurred in approximately 15% of adult and high-risk paediatric B-ALL.
- CRLF2-PY2R8 is associated with young age (median age, 4 years) and IGH-CRLF2 (median age, 8 years) with older age.
MEF2D
- The most common fusion is MEF2D-BCL9, and although MEF2D-rearranged ALL is a rare subtype, it is associated with inferior outcomes.
- Increased expression of MEF2D is associated with activation of histone deacetylase 9, leading to deacetylase and proteosome inhibitor treatment sensitivity.
KMT2A-rearranged
- The KMT2A-rearranged ALL subtype is associated with poor outcomes to treatment, with only modest improvements in the EFS rate to 40–50% from 20–40% historically.
- Relapse or survival in KMT2A-rearranged ALL is not dependent on the fusion partner and therefore recent risk stratifications do not consider them.
- The effectiveness of hematopoietic stem cell transplant and chimeric antigen receptor T-cell CD-19 therapy for patients with KMT2A-rearranged ALL remains unclear.
Low-hypodiploid and near-haploid
- Patients with low-hypodiploid and near-haploid ALL presents with low WBC counts at diagnosis, with median blood counts of <10 × 109/L.
- In addition, children with this subtype generally have an unfavorable MRD response and poor outcomes.
- Low-hypodiploid ALL in children is associated with a TP53 germline mutation, leading to poor outcomes irrespective of the genetic subtype.
HLF
- The translocation t(17;19)(q22;p13) results in the fusion gene TCF3-HLF, which is associated with treatment failure, relapse, and death within 2 years of diagnosis.
- Cells with TCF3-HLF are sensitive to the B-cell lymphoma 2 inhibitor venetoclax.
IKZF1 deletion and interactions with genetic subtypes
- Somatic IKZF1del occurs in ~15% of pediatric ALL cases and confers a significantly worse outcome.
- Clinically, IKZF1del is associated with older age at diagnosis, a higher WBC, and higher EOI MRD.
- It has also been determined that IKZF1del results in inferior outcomes in patients with relapsed ALL, specifically in IRG (except TCF3-PBX1 and DUX4), HRG, and EOI-MRD-positive patients.
- In the Ma-Spore trials, the prevalence of HRG increased in IKZF1del and IKZF1+ compared with IKZF1−.
- IKZF1 variation occurs in ~1% of patients with B-ALL, and recently germline IKZF1 had been characterized as a leukemia predisposition gene.
Impact of genetic subtypes on NCI criteria
- NCI high-risk is defined by infant or adolescent age (<1 or ≥10 years), or a high WBC count at diagnosis (≥50 x 109/L), and this group has a poor prognosis compared with NCI standard-risk patients.
- NCI standard-risk patients account for 60%, 52%, and 22% of the FRG, IRG, and HRG groups, respectively, while NCI high-risk patients account for 40%, 48%, and 78% of the FRG, IRG, and HRG groups, respectively.
- Infants with ALL (aged <1 year) are uniformly NCI high-risk and have the poorest overall treatment outcomes, especially infants younger than 6 months of age.
- In the Ma-Spore cohort, adolescent and young adults (>10 years of age) in the FRG group showed a similar 5-year EFS and OS compared with the 1–10-year age group.
Conclusion
This review summarized the genetic subtypes of B-ALL and their associated clinical characteristics. Identification of these genetic subtypes will improve the current risk stratification of B-ALL and help to determine the optimal intensity of treatment for individual patients. It has also been reported that there is less variability in the treatment responses and survival outcomes within the genetic subtype groupings compared with NCI groups. Overall, improved and accurate risk criteria will enable better treatment for patients with B-ALL with fewer side effects.