Genetic Testing in the Diagnosis and Biology of Acute Leukemia

2017 Society for Hematopathology/European Association for Haematopathology Workshop Report

Marian H. Harris, MD, PhD; David R. Czuchlewski, MD; Daniel A. Arber, MD; Magdalena Czader, MD, PhD


Am J Clin Pathol. 2019;152(3):322-346. 

In This Article

B Lymphoblastic Leukemia/Lymphoma

B lymphoblastic leukemia (B-ALL) is a neoplasm of B-cell precursors. Many genetic subtypes are recognized by the WHO, and two new provisional categories were introduced in the latest revision: B-ALL with intrachromosomal amplification of chromosome 21 (iAMP21) and B-ALL, BCR-ABL1-like.[1] In addition, the importance of the association between the low hypodiploid subtype of B-ALL with hypodiploidy and mutations in TP53, which are often germline, was recognized. Recent studies have continued to describe additional genetic groupings of B-ALL based on genomic profiling and expression analysis; future work will no doubt clarify the prognostic and therapeutic relevance of these groups.

In addition to karyotype, the CAP/ASH guidelines recommend testing cases of B-ALL for the BCR-ABL1 rearrangement and for KMT2A rearrangements, with additional testing for ETV6-RUNX1 rearrangement, iAMP21, and trisomies 4 and 10 in pediatric cases. Genetic testing for alterations in PAX5, JAK1, JAK2, and IKZF1 may also be performed.[2] The NCCN guidelines offer similar recommendations of karyotype and appropriate FISH, as well as RT-PCR for transcript size in BCR-ABL1 rearranged B-ALL. In cases that are BCR-ABL1 negative, the NCCN guidelines encourage testing for alterations associated with BCR-ABL1-like B-ALL, including fusions of ABL1, ABL2, CRLF2, CSF1R, EPOR, JAK2, or PDGFRB, and mutations of FLT3, IL7R, SH2B3, JAK1, JAK3, and JAK2.[49]

Overall, 23 cases of B-ALL were included in the workshop: 15 cases of BCR-ABL1-like B-ALL and nine cases of other B-ALL Table 5 and Table 6. Another case brought up a differential diagnosis between B-ALL, BCR-ABL1-like vs MPAL, B/myeloid, NOS. Among the nine cases of other B-ALL, six underwent DNA sequencing and none underwent RNA sequencing. A case of hypodiploid B-ALL used microarray to clarify chromosomal status (case 45) after sequencing revealed a TP53 mutation. The array revealed a likely hypodiploid origin of the leukemia indicating high-risk disease, and the patient was offered HSCT. Two cases of B-ALL with iAMP21 (cases 235 and 366) were both detected by standard FISH for the ETV6-RUNX1 rearrangement (which also provides the necessary copy number information to diagnose iAMP21), leading to intensified therapy for both patients. As was seen in these cases, B-ALL with iAMP21 can show additional genomic changes, including abnormalities of chromosome 7 (case 235) and rearrangements of CRLF2 (suggested by CRLF2 overexpression detected by flow cytometry in case 366).

BCR-ABL1-like B-ALL comprises 10% to 25% of B-ALL and is genetically diverse. It was initially described as a group of B-ALL with a gene expression profile similar to that of BCR-ABL1-positive B-ALL but without the BCR-ABL1 rearrangement.[50–52] BCR-ABL1-like B-ALL shows activation of kinase or cytokine receptor signaling pathways via a number of different genetic alterations. including CRLF2 rearrangements leading to CRLF2 overexpression, ABL-class fusions (rearrangements of ABL1, ABL2, CSF1R, PDGFRB, and other kinases with many different partners), and alterations leading to activation of the JAK/STAT pathway, including JAK2 and EPOR fusions as well mutations in genes within the JAK/STAT pathway.

BCR-ABL1-like B-ALL carries a poor prognosis, but many of the genetic alterations may be sensitive to targeted therapy. Case reports have described sensitivity of cases with ABL-class fusions to tyrosine kinase inhibitors such as imatinib and dasatinib, while cases with JAK/STAT pathway activation may be sensitive to JAK inhibitors such as ruxolitinib. These possible therapeutic options have led to high interest in the identification of these cases. Perhaps as a reflection of this interest, 15 cases of BCR-ABL1-like B-ALL were submitted to the workshop (Table 6). Twelve cases had CRLF2 rearrangements, which were identified using a variety of different approaches. Some CRLF2 rearrangements were identified following detection of CRLF2 overexpression via flow cytometry. In other cases, CRLF2 rearrangement was suspected based on karyotype or gene expression findings using a low-density array.[53] In yet other cases, the CRLF2 rearrangement was detected directly by RNA sequencing. Confirmatory testing was done by FISH, RT-PCR, or RNA sequencing. Of the three cases without a documented CRLF2 rearrangement, one showed high CRLF2 expression but no testing was performed to confirm a rearrangement, one had a JAK2 rearrangement (Image 5, case 123), and one had a PDGFRB rearrangement. At least five cases had JAK2 mutations detected by sequencing, all of which occurred at the same hotspot location (three cases with Arg683Gly and two with Arg683Ser); at least two of these patients received JAK inhibitors as part of their therapy. The case of BCR-ABL1-like B-ALL with a PDGFRB rearrangement (case 343) raises interesting questions about the relationship between this group of B-ALL and the group of myeloid/lymphoid neoplasms with eosinophilia and rearrangement of PDGFRA, PDGFRB, or FGFR1, or with PCM1-JAK2.

Image 5.

Case 123. A, Hypercellular marrow with sheets of sheets of blasts (H&E). B, Peripheral blood smear showed small to medium-sized blasts with round nuclei, fine chromatin, and scant cytoplasm (Wright-Giemsa). C, Diagram showing domain structure of GOLGA5, JAK2, and GOLGA5-JAK2 fusion. D, Results from reverse transcription polymerase chain reaction (RT-PCR) assays confirming the GOLGA5-JAK2 fusion. E, Direct sequencing of RT-PCR product shows GOLGA5-JAK2 fusion with preservation of JAK2 kinase domain. Courtesy of Yu Hui, Marilyn M Li, Sarah K Tasian, Michele Paessler, Vinodh Pillai, and Gerald Wertheim.

Because of the diversity of genetic changes that underlies BCR-ABL1-like B-ALL, screening for this subtype is challenging. To screen every case for every possible fusion by FISH or RT-PCR would be expensive and unwieldy, while broader RNA sequencing approaches have not been widely accessible. Algorithms to focus testing on patients most likely to benefit may be helpful. One strategy is to use expression profiling to help identify the subset of cases likely to harbor BCR-ABL1-like genetics.[53] Targeted RNA sequencing is also becoming more widely available, and has the potential to identify a wide variety of fusions with a single assay.[54,55] In the reported literature, CRFL2 rearrangements comprise approximately 50% of BCR-ABL1-like cases, while 80% of the BCR-ABL1-like cases submitted to the workshop had CRLF2 rearrangements. This high proportion of CRLF2-rearranged cases may reflect the fact that flow cytometry for CRLF2 overexpression is a relatively simple addition to the testing algorithm while other types of BCR-ABL1-like ALL are often missed with current testing. The increasing availability of targeted RNA sequencing and expression analysis will allow more routine identification of BCR-ABL1-like B-ALL.

Challenges in screening for the BCR-ABL1-like phenotype were further illustrated in case 119 Image 6. Karyotype was normal; however, a SNP array demonstrated the P2RY8-CRLF2 rearrangement. This fusion was challenging to prove on FISH due to the loss of the chromosomal region covered by the red-orange probe. JAK1 and CDKN2A mutations were also detected, consistent with a BCR-ABL1-like signature. The interpretation of this case was further complicated by prominent expression of B-cell markers and partial expression of lineage-specific MPO seen on flow cytometry, cytochemistry, and immunohistochemistry. Two diagnoses were considered by the panel: MPAL, B/myeloid, with concomitant PR2Y8-CRLF2 fusion and JAK1 mutation, and B-ALL, BCR-ABL1-like, with aberrant MPO expression. The expression of MPO as a sole marker of myeloid lineage in otherwise typical cases of B-ALL represents a significant proportion of B-ALL cases and can be challenging to interpret.[56] The WHO classification monograph discusses additional features that can be helpful in differentiating between B-ALL with isolated MPO expression and MPAL. These include a confirmation of MPO expression by cytochemistry and demonstration of heterogeneity of antigen expression and scatter characteristics by flow cytometry, which are seen in true biphenotypic/bilineal immunophenotype (illustrated in another workshop case, Image 4). In case 119, even though MPO was positive by flow cytometry, cytochemistry, and immunohistochemistry, flow cytometry demonstrated no significant variability of scatter characteristics or difference in antigen expression between MPO-positive and MPO-negative blasts. Interestingly, in a previously published cohort of B-ALL, the MPO-positive group clustered separately from BCR-ABL1-like cases, which further underscores the uniqueness of this case.[57] In case 119, in addition to therapeutic implications of a BCR-ABL1-like signature, there is a prognostic significance of isolated MPO expression when associated with a precursor B-cell phenotype. Within the B-ALL group, cases with isolated MPO expression showed a significantly increased risk of relapse, similar to that of MPAL, and were proposed to be excluded from the standard risk category.[56,58]

Image 6.

Case 119. A, Bone marrow aspirate smear showing some variability in blast morphology. B, Flow cytometry showing a partial expression of myeloperoxidase (MPO). C, Diagram of CRLF2 and P2RY8 losses leading to P2RY8-CRLF2 rearrangement as demonstrated by single-nucleotide polymorphism array. Break apart FISH showing loss of the red-orange probe. FITC, fluorescein isothiocyanate; PE, phycoerythrin. Courtesy of Sarah M. Choi, John K. Frederiksen, Charles W. Ross, Dale L. Bixby, and Lina Shao.

Fusions between MYC and immunoglobulin genes have been previously described in B-ALL and, in some instances, overlap with MYC-rearranged Burkitt lymphoma was considered. A case with blast-like morphology, precursor B-cell immunophenotype (partial expression of terminal deoxynucleotidyl transferase [TdT]) and t(8;14)(q24.1;q32) was submitted to the workshop (case 265). A recent series of MYC-rearranged neoplasms with precursor B-cell immunophenotype found evidence of aberrant VDJ (variable-diversity-joining region) recombination within the MYC-IGH translocation, analogous to that found in B-cell precursors.[59] Moreover, the mutational and methylation profiles of these cases were similar to other B-ALLs and not Burkitt lymphoma. Based on this evidence, the authors suggested adding MYC-rearranged cases with a precursor B-cell immunophenotype to the category of B-ALL with recurrent genetic abnormalities.

Two additional B-ALL cases had unusual mutations, more commonly associated with myeloid malignancies. One patient was a young male with B-ALL, NOS with normal karyotype and U2AF1 mutation (case 66). U2AF1 is mutated in a variety of myeloid neoplasms such as MDS, AML, MPN, MDS/MPN, and also in B-cell lymphomas, T-cell leukemias and lymphomas, histiocytic neoplasms, and solid tumors. Spliceosome mutations, including identical U2AF1 mutations, have been rarely reported in B- and T-ALL.[60,61] The importance of functional studies was emphasized in case 367 describing B-ALL with atypical BCR-ABL1 fusion and ATRX p. N1860S mutation at a VAF of 99.8%. Initially, this likely germline mutation was considered potentially pathogenic; however, in a recent report it was shown to lack the epigenetic signature of other germline ATRX mutations and therefore is best considered a benign polymorphism.[62] In addition to germline ATRX mutations associated with α thalassemia/mental retardation X-linked syndrome, somatic ATRX mutations have been reported in patients with B- and T-ALL, myeloid neoplasms, lymphomas, and a variety of solid tumors.