The immunophenotypic and genetic characterization of pediatric T-lymphoblastic leukemia with a mature immunophenotype

T-lymphoblastic leukemia (TLBL) is a neoplasm of hematopoietic progenitors committed to the T lineage and accounts for about 15% of childhood leukemias.¹ While extensive immunophenotypic and genomic characterization exists for B-lymphoblastic leukemia, similar reproducibility between these findings in TLBL remains limited. This is largely due to the discordant expression of differentiation markers across various immunophenotypic classification schemes, recapitulating normal complex T-cell differentiation stages.² Immunophenotypically, TLBL typically expresses T-cell-associated antigens (CD2, CD5, and CD7) which together typify at least 90% of cases, along with lineage-specific CD3, either cytoplasmic (cCD3) or surface (sCD3), in nearly all cases. The vast majority of TLBL cases show immature T-cell-associated marker expression, such as TdT (in at least 90% of cases), CD1a, CD10, and CD34.³ In addition to CD34, other aberrant myeloid antigen expressions are seen in early T-cell precursor lymphoblastic leukemia/lymphoma (ETP-ALL). Occasional TLBL cases lack the expression of immature and myeloid markers,^4,5 mimicking mature T-cell neoplasms and posing significant diagnostic challenges. Accurate identification of such cases is critical, particularly in pediatric patients, where prompt diagnosis is essential for timely treatment.

Here, we characterize the features of pediatric TLBL cases lacking the expression of immature markers, with an aim to identify immunophenotypic clues, genetic drivers, and signaling pathway alterations associated with these rare and challenging cases.

This study included 15 cases diagnosed with pediatric TLBL using the World Health Organization (WHO) criteria⁶ on peripheral blood (PB) or bone marrow (BM), between January 2012 and July 2024. The study was approved by the Institutional Review Board. Immunophenotype was characterized by flow cytometry, and all cases met the following criteria: (i) positive cCD3 and sCD3; (ii) negative or <10% expression of CD1a, TdT, CD10, and CD34. Genomic analysis including chromosome analysis, fluorescence in situ hybridization (FISH), whole genome sequencing (WGS), and RNA-sequencing (transcriptome) were performed in ten cases. Of the remaining five cases, three cases (#3, 4, 6) were analyzed by targeted capture-based next-generation sequencing, with two of these (#4 and #6) also evaluated by optical genome mapping (OGM). The other two cases (#1 and #7) were assessed by conventional cytogenetics and FISH only. T-cell receptor (TCR) and immunoglobulin (IG) rearrangement analysis were performed using the clonoSEQ^® assay.

The patients’ characteristics at diagnosis and follow-up are summarized in Table 1. The cohort comprised 12 males and three females (median age, 12 years; range, 1.5-18.0). The BM was diffusely involved in all examined cases (15/15, 100%). PB blast percentages ranged from 0% to 96% (median, 45%); notably, for patients #1, 3, and 13, the original complete blood count (CBC) data were not available, and their CBC data in Table 1 were from a follow-up visit at the time of relapse. Five patients (5/13, 38%) presented with a mediastinal mass. Histologic examination of the BM biopsy showed diffuse or clustered blasts with a high nuclear/cytoplasmic ratio. On BM and PB smears, blasts in all cases showed a similar morphology: small to medium-sized, with round to irregular nuclei, dispersed chromatin, distinct nucleoli, and scant agranular cytoplasm. All 15 patients received TLBL-directed therapy. The median follow-up duration was 15 months (range, 1-123 months). Overall, nine patients had relapsed or refractory disease, while six achieved remission, including three who underwent transplantation (cases #7, #9, and #14 at 5, 7, and 19 months after diagnosis, respectively). At last follow-up, eight patients had died and seven were alive.

At immunophenotypic level, all 15 (100%) cases were positive for cCD3, sCD3, CD5, CD7, and CD45. CD45 was expressed moderately to brightly in 14 (93%) cases, and sCD3 was high (comparable to that of mature reactive T cells) in 11 (73%) cases. Cases were CD8 positive or CD4 and CD8 double-negative (DN) in 12 (80%) and three (20%), respectively. Surface TCR (sTCR) yδ was positive in seven (46%), while sTCR a|3 was positive in five (33%) cases. The remaining three cases include one with a co-expression of sTCR a|3 and yb, and two without sTCR expression. Cases with expression of sTCR a|3 (TCR a|3 +yb+: 1 case; TCR a|3 +yδ-: 5 cases) were more frequent to show expression of CD2, seen in five of six (83%) cases, as compared to only one of the seven sTCR yb+ a|3 - cases (P=0.029) (Figure 1). For immunophenotype subtype classification, we performed a hierarchical classification based on the expression status of sTCR, CD4, CD8, and CD2, and classified cases molecularly based on the most recent genomic classification of childhood T-lymphoblastic leukemia by Pölönen et al. and Milani et al.^7,8 (Figure 1).

Genomic profiling revealed recurrent rearrangements including STIL::TAL1 in three cases (#1, #2, and #5; #2 with concurrent LMO2 rearrangement and #5 with concurrent TRA::MYC), KMT2A::AFDN in two cases (#10 and #11; case #10 with concurrent SUZ12 truncation), PICALM::MLLT10 in two cases (#12 and #13), TRA/D::MYC in two cases (#5 and #6; #5 with concurrent STIL::TAL1), STAG 2 truncation in one case (#9) and concurrent ZBTB16::ABL1 and TCRB::LMO2 rearrangements with ETV6 truncation in one case (#8). In addition, one case (#7) by chromosome analysis showed t(11;14)(p11.2;q32.1), presumably involving BCL11B at 14q32. Other common genetic changes included CDKN2A/B (7/13, 53%), PHF6 and PTEN (3/13, 23% each). Copy number analysis revealed CDKN2A/B deletions in seven (53%) cases, with concurrent deletions of both CDKN2A and CDKN2B identified in six of these cases (85%). PTEN alterations were identified in three (23%) cases, RPL22 variant/deletion, ETV6, IKZF1, MYB, and TCF12 variants were identified in one (8%) case each (Figure 1). NOTCH1 alterations - frequently observed in adult LBL - were detected in only two patients (#9 and #11). TCRg rearrangement was clonal in all ten tested cases. Four tested sTCRgδ⁺ cases showed clonal TCRg rearrangement (4/4, 100%, #6, #8, #9 and # 12) with IGH rearrangement additionally seen in case #12 and β rearrangement in case #9. Three of four (75%) tested sTCRaβ⁺ cases showed concurrent TCRg and β rearrangements, while one was clonally g restricted. The remaining two cases with double-positive (#7) and double-negative (#14) sTCRaβ and gδ expression were clonally grestricted.

In summary, we characterized 15 cases of TLBL expressing a mature immunophenotype. In contrast to their immature/ blastic morphology, these neoplasms lack the expression of immature markers, including CD34, CD1a, CD10, and TdT. Bright CD45 and high sCD3 expressions, similar to mature lymphocytes by flow cytometry, make it difficult to distinguish blasts from mature T cells. The immunophenotypic profile was similar between sTCRaβ⁺ and sTCRgδ⁺ cases, except for the expression status of CD2: sTCRgδ⁺ mature TLBL more often have CD2^-CD5⁺ and express CD7 (bright). Given the negative immature markers and the expression of mature markers, these cases can be potentially misdiagnosed as mature T- or B-cell lymphoma/ leukemia. A high index of suspicion for TLBL should be warranted, especially in pediatric patients. Incorporating immunophenotype and associated molecular findings into diagnosis should be considered. Clinical and radiological presentations are also helpful, as the detection of a mediastinal mass raises the possibility of TLBL. Circulating blasts and immature morphology are other supportive features for a diagnosis of TLBL instead of mature T-cell lymphoma/leukemia.

Table 1.Clinical features of patients with T-lymphoblastic leukemia with a mature immunophenotype.

Figure 1.Molecular characterization of 15 cases of T-lymphoblastic leukemia with a mature Immunophenotype. (A) Oncoprint depicting molecular and immunophenotypic features of 15 patients with T-lymphoblastic leukemia with a mature immunophenotype. Every column represents 1 patient; case numbers are provided at the top of the plot. Orange squares indicate the presence of a gene fusion, blue squares indicate the presence of a single nucleotide variants (SNV) or insertion and/or deletion (indel) in the molecular section and positive expression of markers in the immunophenotype section, green squares indicate the presence of copy number variations (CNV), salmon squares indicate the presence of copy-neutral loss of heterozygosity (cnLOH), gray square shows negative for molecular or immunophenotypic markers, and white square indicates not tested. (B) Immunophenotypic findings in correlation with molecular classification of 15 patients with T-lymphoblastic leukemia with a mature immunophenotype. Alt: alteration; Trunc: truncation. *: detected by chromosome analysis and/ or fluorescence in situ hybridization (FISH); ~: both nonsense variants (SNV) and deletion (CNV) are present; #: different types of variants SNV (#10) or CNV are present (#3, and #4).

TLBL with an entirely mature immunophenotype has only been rarely reported in children.^4,9 Consequently, the molecular characteristics and clinical features of such cases remain largely unknown. A previous study by Cheng et al.⁴ described three cases of TLBL with a mature immunophenotype, with the most frequent molecular aberrations being CDKN2A deletion and TRA/D translocation (each observed in 2 of 3 cases, including 1 case with concurrent TRA/D::-MYC, TCRB::LMO2, and CDKN2A deletion). Two of the three patients in that series either failed to achieve remission or experienced relapse. We explored such cases in more detail in our study, and we found that at molecular level, TLBL cases with a mature immunophenotype are characterized as NOTCH1-independent pathways, manifested as frequent loss of the TLBL tumor suppressor genes, such as CDKN2A/B (53%), PTEN (23%), RUNX1/BCL11B (23%), PHF6 (23%), and ETV6 (15%). They often display genomic alterations that cause aberrant activation of the STIL::TAL1/LMO2 oncogenes, KMT2A rearrangement, PICALM::MLLT10 fusion transcript, or MYC activation (60%). Previous studies have shown that TAL1 a|3-like subtype (enriched with TCRA-expressing single-positive/mature αβ cells), LMO2 yδ-like subtype (enriched with δ/effector T cells), and TRA/D::MYC-rearranged TLBL mature leukemias cluster close together based on their gene expression signature.^7,10,11

Figure 2.Morphologic and immunophenotypic features of case #13. Aspirate smears show characteristic blasts (Wright-Giemsa), magnification x1,000, oil. Flow cytometry immunophenotyping shows blasts with moderate to bright CD45 expression merging with lymphocytes. Blasts (red color) are positive for surface CD3 (intact, without significant loss), CD5, CD7, CD8, surface TCR γδ, CD19, and surface K and negative for CD1a, CD2, CD34, surface λ,, and TCR αβ.

In all cases described in this study, tumor cells demonstrate T-cell lineage with the expression of both surface and cytoplasmic CD3. In one case (#13), tumor cells showed a mixed phenotype with PICALM::MLLT10 fusion with blasts expressed both T (sCD3, cCD3, CD5, CD7, and CD8) and B (CD19, CD79a, and K light chain) markers (Figure 2). Another case with PICALM::MLLT10 fusion transcript showed both TCRy⁺ and IGH rearrangements (#12).

Although patients in this study were treated according to different protocols and the cohort size is quite small, the results showed that TLBL with a mature immunopheno-type often presented with aggressive disease with poor prognosis, manifested as high white blood cell counts at diagnosis (>100×10⁹/L) (4/12, 33%), relapse/refractory (9/15, 60%), and fatality due to leukemia (8/15, 53%). In previous studies, LMO2 intergenic loss, TRA/D::MYC, and PTEN deletions were associated with relapse,^4,7,8 consistent with the findings of our study.

Footnotes

• Received May 15, 2025
• Accepted August 1, 2025

Correspondence

M. KHANLARI - mahsa.khanlari@stjude.org

Disclosures

No conflicts of interest to disclose.

Contributions