Relapsed B‐acute lymphoblastic leukemia with aberrant myeloperoxidase expression following CAR T‐cell therapy: A diagnostic challenge

An 18-year-old man presented with fever to another hospital in 2009. A complete blood cell count (CBC) showed: WBC 3.6 K/μL, hemoglobin 9.6 g/dL, and platelets 80 K/μL. Bone marrow biopsy showed 93% blasts. Flow cytometry analysis showed the blasts were positive for CD10, CD19, CD20, CD22, CD45, HLA-DR, TdT, and negative for CD2, CD3, CD5, CD7, CD13, CD14, CD15, CD33, CD34, CD56, CD64, CD71, CD117, myeloperoxidase (MPO), consistent with B-acute lymphoblastic leukemia (B-ALL). Cytogenetic analysis showed a complex karyotype: 56~60,XY,+X,+4,+4,+6,i(7)(q10),+8,+10,der(12)t(1,12)(q23;p13),+14,+18,+21,+21,+mar[cp9]/46,XY[11]. Fluorescence in situ hybridization (FISH) analysis was negative for KMT2A/MLL rearrangement and BCR-ABL1 or TEL-AML1. His past medical history was significant for hepatoblastoma diagnosed at age 2, treated by liver resection and chemotherapy with carboplatin, 5-FU and vincristine. He subsequently developed lung metastasis that was managed with lobectomy, platinum-based chemotherapy, radiation, and autologous stem cell transplant. Following the initial diagnosis of B-ALL he received induction chemotherapy with methotrexate, dexamethasone, prednisone, and daunorubicin, followed by induction 2 with vincristine, and intensification with high-dose cytarabine and methotrexate for four courses. He refused further intensification and started maintenance with 6-mercaptopurine, vincristine, methotrexate, and prednisone. He relapsed in 2014 and was treated with vincristine, dexamethasone, and asparaginase with intrathecal therapy, and consolidation with ifosfamide and etoposide, followed by fludarabine and high-dose cytarabine which led to negative minimal residual disease (MRD) by flow cytometry. He relapsed again in June, 2015 and proceeded with CD19 chimeric antigen receptor (CAR)-T cell therapy. He experienced mild cytokine release syndrome and reversible neurologic adverse events. He remained in MRD-negative complete remission (CR) for 1 year until June, 2016 when he relapsed with CD19-negative leukemia. He received CD22 CAR-T cell therapy in September, 2016 and a second infusion in November, 2016, achieving MRD-negative CR that persisted until March, 2017. At that time, he had another relapse and received humanized CD22 CAR-T cells, with an engraftment failure. He received inotuzumab in June, 2017 and achieved MRD-negative CR in July, 2017. In May, 2018, he developed pancytopenia and came to our institution for further evaluation. At our hospital, a CBC showed: WBC 3.6 K/μL, hemoglobin 14.3 g/dL, platelets 32 K/μL, with 20% blasts. Bone marrow biopsy showed 80% blasts with 50% positive for MPO by cytochemistry (Figure 1). Flow cytometry immunophenotyping revealed the blasts were positive for CD22 (dim partial), CD33 (partial), CD34, CD38, CD45, cytoplasmic CD79a, CD123, HLA-DR, MPO (22%), TdT, and negative for CD2, CD3, CD4, CD5, CD7, CD10, CD13, CD14, CD15, CD19, CD25, CD36, CD41, CD56, CD64, CD117, cytoplasmic IgM. Regarding B-cell markers, CD19 was completely absent, CD22 was dim and partial, and cytoplasmic IgM was negative, but CD79a was mostly positive (Figure 2). Immunohistochemical analysis showed the blasts were positive for MPO (20%), CD79a and PAX-5 (strong), CD22 (weak partial), and negative for CD117 (Figure 1). Cytogenetic analysis showed 57~58,XY,add(1)(q32),+4,+4,+6,del(7)(p13),i(7)(q11.2),-9,+11,+14,+15,+17,add(17)(p11.2)x2,+18,+21,+21,+der(?)t(1,?)(q21;?),+1-6mar[cp14]/46,XY,del(20)(q11.2q11.3)[2]/46,XY,t(4,15)(p16;q22)[1]/46,XY[3]. FISH analysis revealed TP53 deletion and monosomy 17, and no evidence of BCR-ABL1 or KMT2A/MLL rearrangement. Nanofluidics-based multiparametric RT-PCR did not detect any translocations that are commonly seen in acute leukemia. Next-generation sequencing assessing 81 genes that are commonly mutated in hematopoietic neoplasms detected two mutations in TP53 (c.456del p.G154fs, VAF 22.8% and c.574_576delinsTAGAGGGG p.H193fs, VAF 4.8%). The leukemia was considered to be relapsed B-ALL with aberrant MPO expression and CD19 loss. The patient received venetoclax, navitoclax, asparaginase, vincristine, and dexamethasone, but treatment was complicated by fulminant sepsis and death due to multiorgan failure in July, 2018. We present this case because it illustrates two difficult issues in the diagnosis and management of B-ALL. The first is the impact that novel therapies have on its immunophenotype. The second is the challenge of MPO expression in B-ALL and its boundary with mixed phenotype acute leukemia (MPAL). CD19-negative relapsed B-ALL is observed in up to 20% of patients after CD19-directed CAR T-cell therapy1 and 22% of patients treated with blinatumomab, an anti-CD3/CD19 bispecific T-cell engager antibody.2 It may be attributed to several biological mechanisms. One is loss of CD19 epitope as a result of hemizygous deletion spanning the CD19 locus, de novo frameshift and missense mutations of CD19, or alternative splicing of CD19 mRNA allowing expression of the N-terminally truncated CD19.1 Another is selection of a minor CD19-negative clone under the strong selective pressure of CD19-directed therapy.3 The third hypothesis is lineage switch as a result of cell reprogramming or dedifferentiation of previously committed B-lymphoblasts, or myeloid differentiation of a noncommitted precursor or selection of a preexisting myeloid clone after CD19 CAR-T-cell therapy.4 Acute leukemia with >3% blasts positive for MPO as detected by cytochemistry is generally considered as myeloid lineage. Although such a threshold has not been established for flow cytometry, the European Group for the Immunological Classification of Leukemia has recommended a threshold of 10%.5 In fact, ALL blasts might express MPO depending on the maturation of the progenitor cells, and trace MPO has been found in lymphoblasts at the mRNA and protein level.6 MPO activity has been reported in B-lymphoblasts in several studies, however, in most cases MPO was at a low level, that is, <10% by cytochemistry. Cases of B-ALL with high MPO activity are rare. Rytting et al. reported five cases of MPO-positive, otherwise typical ALL without expression of other myeloid markers; all responded to ALL therapy.7 Steiner et al. reported two cases of MPO-positive acute leukemia classified as B lymphoblastic by immunophenotyping; both did not require acute myeloid leukemia (AML) treatment.8 More recently, Oberley et al. reported a large group of pediatric patients with B-ALL that expressed MPO without other myeloid markers and found this group had an increased relapse rate and poorer event-free survival than patients with B-ALL who did not express MPO.9 With complete absence of CD19 and CD10 and substantial MPO expression, the differential diagnosis of CD19-negative B-ALL, AML, or MPAL is in debate. Very weak expression of CD22 due to CD22-directed therapy added additional difficulty in defining B-lineage. The studies of additional B-cell markers CD79a and PAX-5 aided in confirming B-cell lineage. Other myeloid markers including CD13, CD14, CD15, CD64, and CD117 were all negative. The karyotype was complex, but shared a number of common abnormalities as the original B-ALL, indicating clonal evolution rather than a completely different clone. Furthermore, the patient had received chemotherapy and radiation for hepatoblastoma. Therefore, his B-ALL could be considered as therapy-related. It has been reported that at least a subset of secondary B-ALL is therapy-related, and linked to previous treatment with alkylating agents or topoisomerase II inhibitors. These cases usually exhibits characteristic cytogenetic alterations related to different exposures, including t(4;11)(q21;q23) and hypodiploidy with loss of chromosomes 5, 7, and 17, infrequent class I mutations (FLT3, KRAS, NRAS, KIT), and poor prognosis.10 In summary, we report a diagnostically challenging case of relapsed B-ALL that lost CD19, down-regulated CD22, and acquired strong MPO expression after treatment with CD19 and CD22 CAR T-cell therapies. Our case show that MPO expression as the sole myeloid marker in an otherwise typical B-ALL is not sufficient for a diagnosis of MPAL or AML, and this concept is especially important in patients who received B-cell marker-directed immunotherapy (CD19, CD22). Immunophenotypic studies for additional B-cell and myeloid markers, as well as cytogenetic and molecular comparisons to the original leukemia are essential to reach a correct diagnosis.