Introduction

T cells develop and mature in the thymus. Immature T cells are usually not detected at extrathymic sites. Recently, a pathologic condition of indolent T-lymphoblastic proliferation (iT-LBP) has been described [1, 2]. iT-LBP is a nonclonal extrathymic proliferation of immature thymocytes (T-lymphoblasts). It is commonly seen in association with Castleman disease with or without follicular dendritic neoplasm [2,3,4,5,6,7,8]. Association with other conditions has also been reported [2, 7]. The clinical course of iT-LBP is indolent, and therapy is usually not required.

iT-LBP is a rare condition and a diagnosis of exclusion. As a rule, it requires sophisticated techniques to exclude T-lymphoblastic leukemia/lymphoma (T-LBL/L), which is an aggressive malignant disease and needs prompt treatment. We present a T-LBL/L case with interfollicular proliferation pattern, preserved nodal architecture, and no clonal T-cell receptor (TCR) gene rearrangement, which mimicked iT-LBP.

Case

A 13-year-old previously healthy male presented with a 3-month history of neck swelling, which had slowly increased in size. Patient had no fevers, no pain, no night sweats, no weight loss, no difficulty breathing or swallowing, and no skin discoloration or drainage. There were no masses or swelling noted in other parts of the body. Patient had no history of foreign travel or cat exposure. Ultrasound showed two adjacent enlarged submental lymph nodes (LNs) measuring 1.4 × 3 × 3 cm and 1.5 × 1.4 × 2.4 cm, respectively. Laboratory tests were unremarkable. Blood cell counts were within normal limits. An excisional biopsy was performed.

Microscopically, the LN showed preserved nodal architecture with focal interfollicular expansion (Fig. 1a). There were frequent follicles with atrophic germinal centers (highlighted by CD21 immunostaining) and mantle cell hyperplasia (Fig. 1b). Hyalinized vessels and increased high endothelial venules were also seen. These features were compatible with hyaline vascular Castleman disease. There were frequent atypical lymphoid cells in the interfollicular area (Fig. 1c), which were intermediate to large with round, oval, or irregular nuclei, vesicular or finely dispersed nuclear chromatin, indistinct to prominent nucleoli, and scant cytoplasm. Mitoses were easily seen. By immunohistochemistry (IHC) stains, the atypical cells were positive for CD3, TdT, and CD99, partially positive for CD4, negative for CD1a, and mostly negative for CD8 (Fig. 1). Flow cytometry (FCM) was performed and identified a small T lymphoblast population, which represented less than 2% of total cells, and expressed cytoplasmic CD3 (cCD3), TdT, CD7, CD5, CD2 (dim), with no CD1a, surface CD3 (sCD3), CD4, and CD8. No clonal TCR gamma (TCRγ) gene rearrangement was detected by the assessment using Biomed-2 primers, polymerase chain reaction (PCR) and fragment analysis (Fig. 1d). The case was signed out as atypical T-lymphoblastic proliferation with recommendation for close follow-up. A second opinion was requested from an outside hematopathology consultation service, which interpreted the histomorphologic and reported laboratory findings as consistent with hyaline vascular Castleman disease with T-lymphoblastic proliferation with comments on the atypical cytomorphology and high extent of the TdT+ T-lymphoblastic proliferation, and recommended close clinical monitoring.

Fig. 1
figure 1

Histopathology, immunohistochemical staining, and TCR gamma gene rearrangement of the first lymph node specimen. Hematoxylin and eosin-stained section at low power (a) showed preserved nodal architecture with interfollicular expansion. High power of the lymphoid follicle (b) showed atrophic germinal center, mantle cell hyperplasia, hyalinized vessels, and high endothelial venule hyperplasia. High power of interfollicular area (c) showed frequent medium to large--sized atypical lymphoid cells with round, oval, or irregular nuclei, vesicular nuclear chromatin, indistinct to prominent nucleoli and scant cytoplasm. Mitoses were easily seen. CD21 immunostain showed CD21+ B cells (dim to intermediate) and follicular dendritic meshwork (strong), highlighting the regressed germinal centers. CD20 immunostain highlighted B cells in the follicles. The interfollicular atypical cells were positive for CD3, TdT, and CD99, partially positive for CD4, negative for CD1a, and mostly negative for CD8. No clonal TCR gamma gene rearrangement was detected by polymerase chain reaction using Biomed-2 primers (d)

Full size image

The right neck swelling recurred 1 month after the first biopsy, and slowly increased in size. A PET scan done 6 months later demonstrated increased activity in the LNs of the right neck; thus, a second excisional biopsy was performed. Microscopically, the LN showed architectural distortion by prominent interfollicular proliferation of medium- to large-sized atypical lymphoid cells with morphology and phenotype similar to those seen in the first specimen (Fig. 2a). By FCM, a T-lymphoblast population (Fig. 2b, purple) was identified, which had low to intermediate side scatter and represented approximately 25% of total cells analyzed. The abnormal cells were positive for cCD3, CD7 (bright), CD5, TdT, CD2 (dim to negative), CD38, CD71 (bright), CD123 (variable), CD13, CD33, CD19 (partial), and CD45 (intermediate); they were negative or mostly negative for sCD3, CD4, CD8, CD16, CD34, and CD56. A bone marrow aspirate demonstrated 21% blasts and a diagnosis of stage 4 T lymphoblastic lymphoma was made.

Fig. 2
figure 2

Histopathology and immunophenotype of the second lymph node specimen. (a) Hematoxylin and eosin-stained section at low power (HE1) showed distorted nodal architecture by significant interfollicular expansion. High power of interfollicular area (HE2) showed sheets of medium to large-sized atypical lymphoid cells with round, oval, or irregular nuclei, fine nuclear chromatin, distinct one or multiple nucleoli, and scant cytoplasm. Mitoses were easily seen. CD20 stained significantly diminished residual B cells. The interfollicular atypical cells were positive for CD3 and TdT, negative for CD1a, and mostly negative for CD4 and CD8. (b) By flow cytometry, an abnormal cell population (purple) was identified with low to intermediate side scatter and represented approximately 25% of total cells analyzed. The abnormal cells were positive for cytoplasmic CD3 (cCD3), CD7 (bright), CD5, TdT, CD2 (dim to negative), CD38, CD71 (bright), CD123 (variable), CD13, CD33, CD19 (partial), and CD45 (intermediate); they were negative or mostly negative for surface CD3 (sCD3), CD4, CD8, CD16, CD34, and CD56

Full size image

Genetic studies at the time of the second biopsy included chromosome, FISH, microarray, and next-generation sequencing analyses of BM, and TCR clonality, microarray, and next-generation sequencing analyses of the LN. BM studies showed the following karyotype (Fig. 3a): 47,XY,add(12)(p12),+21[5]/49,sl,+8,+9[1]/46,XY[14]/92,XXYY[5]. Fluorescence in situ hybridization (FISH) analysis showed loss of ETV6 (12p13) and gain of 21q22 (RUNX1) in 11.5%, and gain of 9q34 (ASS1, ABL1) in 4.5% of nuclei (Fig. 3b and c). Microarray analysis showed segmental losses within 1p34.3 (~3.7 Mb), 12p13.31p13.1 (~7 Mb) including ETV6, and 12q23.3q24.13 (~3.6 Mb); gain of chromosome 21; and low-level gain of 9q (Fig. 3d). The LN specimen showed no clonal TCRγ gene rearrangement. Microarray showed the same aberrations with two additional findings of low-level 1q21.2qter gain and 2q24.1qter loss. Next-generation sequencing identified a pathogenic mutation in SETD2 (c.4715 C>A; p.Ser1572*; 31% VAF) (Fig. 3e, upper panel), a likely pathogenic mutation in SH2B3 (c.516_533delinsACCGCCA, p.Ala174Profs*90), and other variants of uncertain significance. Of note, the BM sample showed 2 SETD2 mutations (Fig. 3e). The 2nd SETD2 mutation (Fig. 3e, lower panel), c.4715+3_4715+5delinsGGTCC (p.?), was seen at a VAF of ~5%. This allele was on a different allele relative to the single nucleotide nonsense mutation. It was predicted by in silico programs to abolish the adjacent splice site leading to aberrant splicing and allelic loss-of-function; however, RNA studies are needed to fully assess the functional impact of this variant.

Fig. 3
figure 3

Cytogenetic and next-generation sequencing findings of the lymphoma. (a) Karyotype showed extra material on 12p (ETV6) and an extra copy of chromosome 21. (b) FISH analysis with ETV6 (red) and RUNX1 (green) probes showed one copy of ETV6 (12p) and three copies of RUNX1 (21q). (c) FISH analysis using ASS1/ABL1/BCR probes showed an extra copy of 9q34 (ASS1, ABL1). (d) Microarray analysis showed segmental losses within 1p34.3 (~3.7 Mb), 12p13.31p13.1 (~7 Mb) including ETV6, and 12q23.3q24.13 (~3.6 Mb); gain of chromosome of 21; and low-level gain of 9q. (e) Next-generation sequencing (whole exome sequencing with an average depth of coverage of 489x) of the lymph node sample (upper panel) and bone marrow specimen (lower panel). The lymph node sample showed a pathogenic loss-of-function mutation: c.4715 C>A (p.Ser1572*) at a VAF of 19% in SETD2. This same mutation was seen in the bone marrow sample at a VAF of 25%. A 2nd SETD2 mutation, seen in only the bone marrow sample, was also identified. This variant, c.4715+3_4715+5delinsGGTCC (p.?), represented by one read in the lower panel, was seen at a VAF of ~5% (6 out of 122 reads). This allele is on a different allele relative to the single nucleotide nonsense mutation. It is predicted by in silico programs to abolish the adjacent splice site leading to aberrant splicing and allelic loss-of-function; however, RNA studies are needed to fully assess the functional impact of this variant

Full size image

The patient was treated using the Children’s Oncology Group AALL0434 protocol [9]. He developed methicillin-susceptible Staphylococcus aureus (MSSA) bacteremia followed by multiple MSSA abscesses at the end of induction therapy and had a prolonged hospitalization. He had resolution of PET-avid lymphadenopathy after induction treatment with minimal residual disease (MRD) on day 29 of 1.7%. The severity of his infection necessitated the use of bridging chemotherapy with mercaptopurine and methotrexate for 1 month during recovery. MRD following this treatment was negative and therapy was resumed as planned.

Discussion

T-LBL/L usually shows sheets of medium-sized blast cells with complete or partial effacement of normal lymphoid architecture. Predominantly interfollicular growth pattern with respect to normal nodal architecture as seen in this case is uncommon and can be a diagnostic pitfall. In the absence of a clonal TCR gene rearrangement, T-LBL/L with interfollicular growth pattern could be confused with iT-LBP, especially in cases with hyaline vascular Castleman disease features. iT-LBP is clinically indolent, and unlike the malignant counterpart, the proliferation of nonclonal TdT+ T cells does not require treatment. Distinguishing between iT-LBP and T-LBL/L is clinically critical. Ohgami et al. [2] have proposed specific criteria to accurately diagnose iT-LBP, notably (1) confluent groups of TdT+ T cells in a biopsy specimen, (2) relative preservation of surrounding normal lymphoid architecture, (3) TdT+ T cells without morphologic atypia, (4) absence of thymic epithelium, (5) nonclonal TdT+ T cells, (6) immunophenotype of developmentally normal immature thymic T cells, and (7) clinical evidence of indolence (follow-up >6 months without progression). The extent of T-lymphoblast proliferation in iT-LBP is variable and could be as high as 50% of total volume [4], and thus, there is no established cutoff of the amount of T-lymphoblasts to reliably rule out or rule in iT-LBP. Our first biopsy specimen met the criteria of 1, 2, 4, and 5, but did show some morphologic and phenotypic atypia. Instead of small to medium immature lymphoid cells with blastic chromatin as reported in most cases of iT-LBP [2], the T-lymphoblasts of our specimen were mostly intermediate to large and showed significant nuclear irregularity. Instead of TdT positive, CD4/CD8 double-positive immature thymocyte immunophenotype as seen in most reported iT-LBP cases [2], our case was mostly negative for CD4 and CD8. Follow-up showed disease progression. This case highlights the difficulty to reliably distinguish between iT-LBP and T-LBL/L sometimes, and the importance of applying stringent diagnostic criteria for iT-LBL. As seen in this case, some early immature T-LBL/L cases may not demonstrate a clonal TCR rearrangement due to their germline status. Also, TCR beta gene rearrangement analysis was not performed on our case. The addition of TCR beta gene rearrangement will increase the sensitivity of T-cell clonality detection by approximately 5% [10]. Therefore, a negative clonal TCR rearrangement result cannot be used to completely rule out a diagnostic possibility of T-LBL/L.

Normal thymocytes show consistent and characteristic antigen maturation pattern, which can be demonstrated by FCM study [11]. Early thymocytes express CD34, CD7, TdT, and CD2, then lose CD34 and gradually gain CD5, CD4, cCD3, both CD4 and CD8, and CD1a to become intermediate thymocytes. Late-stage thymocytes lose TdT and CD1a, gradually gain sCD3 and TCR, and eventually become CD4 or CD8 single-positive mature T cells. Intermediate thymocytes are predominant in thymus and show characteristic winged CD4xCD8 differentiation pattern. However, the differentiation of thymocytes outside of thymus can be significantly skewed [12, 13], CD4/CD8 double-positive intermediate thymocyte population and /or winged pattern could be absent. A recent FCM study of 10 iT-LBP cases revealed that most of the cases showed a major CD4-CD8- subset, and one case even showed a major CD8+CD4- subset. These findings are different from most IHC stain results reported, which were usually CD4+CD8+. TdT and cCD3 were almost always positive by either FCM anylysis or IHC stains, but the expression of sCD3, CD2, CD5, CD7, CD10, CD1a, and CD34 was variable [2, 14]. CD33, a myeloid marker, is reportedly expressed in some iT-LBP cases [5, 8]. Therefore, it is not easy to reliably rule out iT-LBP or T-LBL/L based on immunophenotyping. FCM analysis performed on our 1st LN specimen identified a discrepantly much smaller T-lymphoblast population, which expressed cCD3, TdT, CD7, CD5, CD2 (dim) without CD1a, sCD3, CD4, and CD8. This phenotype was not abnormal enough to exclude iT-LBP. The reason for discrepantly much lower T-lymphoblastic proportion by FCM was not clear and could be related to disproportionally greater loss of the T-lymphoblasts during sample process.

More comprehensive FCM immunophenotyping was performed on the 2nd LN specimen and identified a large abnormal T-lymphoblastic population. Overall, the case showed early immature T-LBL/L phenotype (T-II acute lymphoblastic leukemia based on EGIL classification [15]) with the absence of CD1a, CD4, CD8, and sCD3, expression of TdT, and dim expression of CD2. This case also showed expression of myeloid markers CD13 and CD33, but did not fulfill the diagnostic criteria for early T-cell precursor lymphoblastic leukemia (ETP-LL) because of the strong expression of CD5, instead, it can be called near-ETP-LL based on one study [16]. This large series study in children and young adult patients found no statistical difference in overall survival and event-free survival among ETP, near-ETP, and not-ETP groups despite significantly higher rates of induction failure in ETP and near-ETP groups [16]. This case also showed partial expression of CD19, a B-cell marker, which is very unusual and has not been reported in literature. Cytogenetic and molecular genetic findings of this case are compatible with T-LBL/L, especially ETP type [17,18,19,20,21]. SETD2 (SET domain containing 2) is a histone methyltransferase; its downregulation due to a genetic lesion can contribute to both initiation and progression during leukemia development. SETD2 alterations can impair DNA damage recognition and lead to resistance to chemotherapy in leukemia as well [20, 22].

The relatively short history of enlarged LNs argues against the preexistent hyaline vascular Castleman disease, a nonclonal lymphoproliferative disorder. The pathogenesis of Castleman disease remains unclear, but many studies have demonstrated the important role of inflammatory cytokine, especially IL6, in the development of the disease. Some disease conditions may result in similar histopathologic growth pattern with similar inflammatory condition. These conditions include systemic disease such as IgG4-related disease [23], and lymphomas such as angioimmunoblastic T-cell lymphoma [24], lymphocyte-rich classic Hodgkin lymphoma [24], and other lymphomas [25, 26]. Recently, a next-generation sequencing study of Castleman disease revealed multiple mutations in some signaling pathways [27]. SETD2 mutation was seen in one case of hyaline vascular Castleman disease with follicular dendritic cell sarcoma. Next-generation sequencing was not performed on our 1st LN biopsy but on the 2nd LN biopsy and BM, and identified a pathogenic SETD2 mutation in both specimens and a likely pathogenic 2nd SETD2 mutation in BM only. The association of SEDT2 mutation and hyaline vascular Castleman disease-like features present in this case is not clear.

Conclusion

T-LBL/L may occasionally have predominant interfollicular growth pattern, hyaline vascular Castleman-like morphology, and no clonal TCR gene rearrangement, which are the common features of iT-LBP. However, T-LBL should always show some degree of morphologic and/or phenotypic atypia, which is usually not seen in iT-LBP. This case highlights the importance of applying stringent criteria for the diagnosis of iT-LBP. In the presence of any atypical features, more comprehensive workup including cytogenetic and molecular genetic studies should be applied, and close follow-up should be recommended.