Lack of classical astroblastoma features in pediatric MN1::BEND2‐fused brain tumors

Abstract

Three distinct MN1::BEND2 fusion‐positive tumors in pediatric patients. (A) Clinical course for each patient was variable in part due to differences in initial diagnosis. Each patient responded favorably to gross total resection and is stable at last follow‐up. (B) Histologic diversity, lack of prominent classical astroblastoma features, and variable immunoexpression of key markers makes microscopic diagnosis challenging.

Astroblastoma histology has a strong association with MN1 alteration, particularly in the pediatric population, yet clinical reports of MN1‐altered central nervous system (CNS) tumors are limited, and pediatric‐focused series are lacking [1, 2, 3]. The methylation class first described for MN1‐altered brain tumors was “high‐grade neuroepithelial tumors with MN1 alteration” (HGNET, MN1), and the reported frequency of occurrence among patients under 15 years old is 10 times greater than among those aged 15 and above. [4, 5, 6, 7, 8]. Clinical recurrence is commonly reported, with 5‐ and 10‐year progression‐free survival (PFS) rates of 38% and 0%, and overall survival (OS) rates for 5 and 10 years reported at 89% and 55%, respectively [6]. Gross total resection (GTR) is associated with improved prognosis compared to subtotal resection (STR), but optimal therapy is still under investigation [6, 9].

We report a case series of three pediatric patients with tumors harboring the most common MN1 fusion, MN1::BEND2, along with a focused literature review (PubMed and Scopus) for pediatric (<18 years at time of diagnosis) MN1::BEND2 fusion CNS tumors with documented treatment and/or clinical follow‐up. Available data including OS and PFS from this review is shown in Table 1. The three cases in our series showed DNA methylation profiles that aligned with Astroblastoma, MN1‐altered with high‐calibrated scores observed in Table 1.

TABLE 1.

A review of the literature for the clinicopathologic features of pediatric brain tumors with MN1::BEND2 fusion.

Patient ID	References	DKFZ methylation classifier version (calibrated score)	Age at diagnosis (years)	Sex	Histology at diagnosis	Tumor location	Treatment	PFS	OS/last follow‐up
1	Case 1 (current paper)	v12.5 (0.99988)	6	F	Astroblastoma	Right occipital‐parietal	STR + RT (5940 Gy)/chemo (Temodar) ➔ GTR+ RT (5400 Gy)/chemo (Avastin + Irinotecan)	7 years	17 years (alive)
2	Case 2 (current paper)	v12.5 (0.99999)	16	F	Papillary Ependymoma	Left parietal lobe	NTR ➔ STR ➔ GTR + no AT ➔ GTR + RT (5400 Gy)	1 year	10 years (alive)
3	Case 3 (current paper)	v11b6 (0.99999)	10	F	Malignant‐appearing Papillary Neoplasm	Lateral ventricles	GTR + RT (5340 cGy) /Chemo (cisplatin, cyclophosphamide, and vincristine)	5 years	5 years (alive)
13	Boisseau et al., [1]	N/A	15	F	Recurrent Astroblastoma	Parietal and occipital	Surgery ➔ Surgery + no AT	9 years	19 years (alive)
01	Burford et al., [4]	N/A	6	F	Recurrent Astroblastoma	Left parietal	Surgery ➔ 9 Surgeries + RT/Chemo	1 year	10 years (expired)
01	Chen et al. [6]	N/A	9	F	Yolk sac carcinoma	Left frontal‐parietal	NTR + Chemo ➔ gamma knife radiosurgery + Chemo ➔ GTR	8 months	2 years
18	Lake et al. [5]	N/A	5	F	Astroblastoma	Frontal	GTR	7 years	7 years
28	Frederico et al. [9]	N/A	5	F	Ependymoma	Left fronto‐parietal	Surgery ➔ RT ➔ GTR	2 years	21 years

Patient #1 was a 6‐year‐old female who presented with seizures and a left occipital‐parietal region mass. Her clinical course included multiple resections and is summarized in Figure 1A. At last follow‐up, 16 years post‐diagnosis, she had visual field defects and new progression of disease in the anterior third ventricle (Figure 1A and Table 1).

FIGURE 1.

Clinicopathologic details for three pediatric patients with MN1::BEND2 fused CNS tumors. (A) Clinical course for each patient in the series. Swimmer's plot shows length of follow‐up (blue bar) with significant timepoints including surgery, chemotherapy, and radiation. Patient 1 underwent one surgery resulting in a non‐tumor diagnosis (i.e., pseudoprogression), marked by hexagonal data point. (B) Varied histologic findings among the three cases. H&E stains show variable extents of papillary architecture and/or pseudorosetting; no tumors showed overt stromal hyalinization. Extent of immunoreactivity for GFAP, OLIG2, and EMA was highly variable, from minimal to extensive, among the three cases.

Histologically, the tumor showed sheet‐like and pseudopapillary growth patterns. Neither vascular hyalinization nor astroblastomatous pseudorosettes were identified in any of the resection samples. Focal necrosis and frequent mitotic activity were present. OLIG2 staining in the tumor was minimal and focal. GFAP expression was highly variable, with some areas showing strong diffuse staining and others essentially negative. Ki‐67 labeling index was quantified at 3.25%. EMA expression was diffuse with prominent membranous staining in scattered tumor cells (Table S1). Whole transcriptome RNA sequencing (RNA‐seq) revealed an in‐frame MN1::BEND2 fusion, with the 5′ portion of MN1 (exons 1 of 2, ENST00000302326) fused to the 3′ portion of BEND2 (exons 9–14 of 14, ENST00000380033), retaining the transactivation domain of MN1 and the BEN domain of BEND2. Chromosomal microarray analysis showed copy number losses including loss of chromosomes X and 9; terminal deletions affecting the long arms of chromosomes 16 and 21; as well as segmental deletions in 7p, 10p, 11p, 14q, 15q, 20q, and 22q. Notably, the BEND2 locus on the X chromosome retained two copies in the context of loss of chromosome X. The interstitial deletion on chromosome 22 had a breakpoint within the MN1 gene.

Patient #2 was a 16‐year‐old female with 1 ½ years of progressive neurological symptoms who was found to have an extra‐axial mass of the left parietal region. She underwent near‐total resection (NTR), and no adjuvant therapy was given. After local recurrence a year later, she underwent a two‐stage surgery resulting in a GTR, and subsequently radiation therapy. At the last follow‐up, 9‐years post‐diagnosis, she presented at clinical baseline with interval enlargement of a nodule, which remains under observation (Figure 1A and Table 1).

Histologically, the tumor showed papillary architecture and minimal vascular thickening. GFAP staining was patchy overall but accentuated around papillary cores with some astroblastomatous pseudorosettes with stout, outwardly tapered, GFAP‐positive cytoplasmic processes abutting vessels. OLIG2 expression was present in a large subset of tumor cells. EMA expression was limited to focal dot‐like and luminal staining and rare positive cells. Ki‐67 labeling index was quantified at 13.22%. Molecular testing identified an X;22 translocation resulting in MN1::BEND2 fusion. At tumor recurrence, the tumor's papillary architecture was significantly less prominent, with high cellularity and confluent sheet‐like growth in most areas.

Patient #3 was a 10‐year‐old female who presented with headaches and emesis. MRI demonstrated a midline tumor involving the lateral ventricles. She underwent GTR and adjuvant therapy at last follow‐up, 4 years post‐diagnosis, she demonstrated stable disease (Figure 1A and Table 1).

Histologic review of the tumor showed a highly cellular tumor with prominent true rosettes or microlumens, and papillary architecture. Vascular hyalinization and astroblastomatous pseudorosettes were not present. Focal extracellular material resembling mucin prompted consideration of an epithelial neoplasm. Immunoreactivity for GFAP was minimal overall, and OLIG2 was mostly negative. EMA was diffusely positive including strong labeling of the luminal surfaces. Ki‐67 labeling index was quantified at 14.98%. RNA‐seq analysis revealed an in‐frame MN1::BEND2 fusion with exon 1 of the MN1 gene fused to exon 7 of the BEND2 gene. The transactivation domain of MN1 and BEN domain of BEND2 were retained in the fusion transcript. No clinically significant copy number variants were detected by chromosomal microarray. The fusion is likely the result of a balanced rearrangement between chromosomes X and 22.

This case series illustrates several typical elements of MN1::BEND2 fusion tumors: predilection for female patients, favorable outcome for GTR, and recurrence in the setting of STR. All three cases displayed perivascular arrangement of tumor cells, with cases 2 and 3 showing overt papillary growth. All cases had focal necrosis, and showed at least focal positivity for GFAP, although staining was heterogeneous, and with sampling error, total negativity could be observed. EMA immunoexpression ranged from diffuse and strong in two cases to focal luminal expression in one case.

The lack of prominent classical astroblastoma features in these cases resulted in differences in diagnosis at different timepoints of resection, including papillary ependymoma (patient 2) and malignant papillary neoplasm NOS (patient 3). The latter displayed a highly unusual pattern with numerous intratumoral lumens and EMA‐positive apical surfaces, reminiscent of true rosettes. True rosettes have not been previously noted in association with MN1‐altered brain tumors, including in a large meta‐analysis [9]. The lack of stromal hyalinization is also notable, as this is considered a key feature of astroblastoma.

The variable appearances of these tumors in combination with the wide range of expression of GFAP and EMA underscores the challenge in accurate histologic diagnosis, and the importance of molecular testing. The Ki‐67 labeling indices for case 2 (13%) and 3 (15%) corresponded to the typical astroblastoma phenotype (>10%), whereas case 1 was low overall at 3%. This may in part account for the favorable prognosis among these pediatric cases. In all three cases, both methylation classification and RNA analysis enabled more specific diagnosis. The diagnostic or prognostic significance of the observed differences in histologic or immunohistochemical features, if any, is unclear. Although copy number profiling of pediatric cases is limited, our cases represent a variable pattern of copy number alterations. In case 1, the copy number losses involved chromosomes 14, 16, 22q, and X. Additionally, a relatively stable copy number genome, as seen in case 3, has also been previously documented [4]. The copy number loss on chromosome 22q affecting the MN1 locus may suggest a potential rearrangement, but the absence of copy number alteration in this locus does not rule out the presence of the MN1 fusion, as observed in case 3.

Limited data on treatment efficacy exists because of the recent introduction of this diagnostic category. Patients in this series received diverse treatments based on the variety of histologic diagnoses, complicating direct comparison. Today, RNA sequencing would enable precise up‐front diagnosis and standardized therapy. At our institution, pediatric MN1::BEND2 tumors responded favorably to GTR and radiation therapy. Further study of clinical responses and long‐term outcomes for pediatric patients with MN1‐altered tumors is needed.

AUTHOR CONTRIBUTIONS

Jennifer A. Cotter conceived and executed the study; analyzed and interpreted the histological data; and wrote and revised the manuscript. Nicholas Chapman performed literature search, provided the clinical and imaging data, and wrote and revised the manuscript. Mohammed Iqbal analyzed and provided the clinical data and was a major contributor to the revised manuscript. Adam D. Walker assisted with digital histology analysis of the cases and revised the manuscript. Debra Hawes performed the histological examination of the cases, contributed to writing the pathological findings, and revised the manuscript. Tom Belle Davidson provided and revised the clinical oncological portion of the manuscript. Nathan Robison provided and revised the clinical oncological portion of the manuscript. Benita Tamrazi obtained imaging data for cases and revised the manuscript. Jianling Ji analyzed the molecular data and revised the manuscript. Mark D. Krieger analyzed and provided the clinical and imaging data and revised the manuscript.

All authors read and approved the final manuscript.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest.

ETHICS STATEMENT

This study was conducted in compliance with the policies and procedures of the CHLA IRB and Human Subjects Protection Program. In addition, patients and their families provided informed consent for research.

Supporting information

Table S1. Antibodies used in the study.

Chapman N, Iqbal M, Walker AD, Hawes D, Davidson TB, Robison N, et al. Lack of classical astroblastoma features in pediatric MN1::BEND2‐fused brain tumors. Brain Pathology. 2025;35(2):e13297. 10.1111/bpa.13297

DATA AVAILABILITY STATEMENT

Whole slide images of H&E sections from the three original cases (1, 2, and 3) and the recurrence of case 2 (2R) are available at https://doi.org/10.6084/m9.figshare.25629561.v2. Other data supporting the findings of this study are available from the corresponding author on reasonable request.