Introduction
Intracranial mesenchymal tumor (IMT),
FET::CREB fusion-positive is a provisional tumor type in the 2021 World Health Organization (WHO) classification of central nervous system tumors [
27]. It represents primary intracranial mesenchymal neoplasms that typically affect children and young adults. These tumors are characterized by the fusion of a
FET RNA-binding protein family gene, which includes Ewing sarcoma RNA binding protein 1 (
EWSR1) and fused in sarcoma (
FUS), to a cAMP response element-binding protein (
CREB) family gene, which includes activating transcriptase factor-1 (
ATF1), cAMP responsive element binding protein 1 (
CREB1), cAMP response element modulator (
CREM), and cAMP responsive element binding protein 3 like 3 (
CREB3L3). IMTs are rare tumors with only scattered case reports and small case series published in the literature [
35,
36,
40], so specific information regarding their characteristics, especially their clinical behavior, is not well known. In the current study, the clinicohistopathologic and molecular findings of five new cases of
FET::CREB fusion-positive IMT are detailed. Additionally, a comprehensive review of the literature to summarize relevant clinical, histopathological, genetic, treatment, and outcome information of this unique tumor type is reported.
Materials and methods
Tumor sample collection
Five IMT cases with documented FET::CREB fusion were collected from participating institutions. Fusions were confirmed by either fluorescence in situ hybridization or next-generation sequencing at referring laboratories. In addition, four of the cases were subjected to methylation profiling. The following patient characteristics and clinical data were retrieved from hospital records: age, sex, presenting symptoms, MRI characteristics, tumor location and size, extent of resection, histology, fusion partners, methylation profile, treatment, and follow-up data.
Fluorescence in situ hybridization
Fluorescence in situ hybridization (FISH) analysis was performed on representative unstained 5-μm formalin-fixed paraffin-embedded (FFPE) sections of cases 1, 2, 3, and 5 using an
EWSR1 break-apart probe set (Agilent Technologies, Santa Clara, CA) followed by reflex studies with three fusion probe sets in the following order:
EWSR1::ATF1,
EWSR1::CREB1, and/or
EWSR1::CREM. The three custom fusion probe sets utilized BAC clone cocktails that were selected based on their location in the UCSC Human Genome Browser (
http://genome.ucsc.edu) and were obtained from the BACPAC sources of BACPAC Genomics (
https://bacpacresources.org). Probes were directly labeled by nick translation and hybridized as previously described [
1]. Each clone was also hybridized to normal metaphases to confirm correct mapping, optimal signal intensity, and lack of cross-hybridization. Results were evaluated using the thresholds established by in-house validation studies.
Next-generation sequencing
Nucleic acid was extracted from representative unstained 5-μm FFPE sections of Case 4 and subjected to a solid tumor panel for sequence analyses of 238 genes and a fusion panel targeting over 700 exons of 117 cancer genes at the Children’s Hospital of Philadelphia (CHOP). Extracted DNA was fragmented and tagged using SureSelectQXT target enrichment to generate adapter-tagged libraries. Biotin-labeled probes specific to the targeted regions were used for capture hybridization. Libraries were enriched for the desired regions using streptavidin beads and then subjected to sequence analysis on Illumina MiSeq or HiSeq platform for 150 bp paired-end reads. RNA sequencing libraries were prepared using Archer Universal RNA Reagent Kit with CHOP fusion panel custom-designed primers with target specific molecular barcode. Sequencing data were analyzed using Archer™ Analysis for fusion genes.
Genome-wide DNA methylation profiling
Genome-wide methylation profiling was performed on cases 1, 2, 3, and 4. Briefly, genomic DNA was extracted from FFPE tissue sections after macro-dissection to enrich for viable tumor content using the All Prep DNA/RNA FFPE kit (Qiagen). The DNA was bisulfite converted using EZ DNA Methylation Kit (Zymo Research D5001) and subsequently processed using Infinium Methylation EPIC 850K kit according to the manufacture’s protocol (Illumina). The beadchips were scanned on iScan reader (Illumina) and output idat files were processed through the DKFZ CNS classifier versions v11b6 and v12b6 (publicly available, unpublished) of the CNS tumor methylation classifier [
8]. In addition, we requested the National Cancer Institute Laboratory of Pathology to provide a dataset of ~ 7500 brain tumors with 198 classes. Raw idat files were processed using single sample noob normalization available in minfi R package [
5]. Prior to this we removed any probes with detection
p value less than 0.05. Because this given dataset of ~ 7500 samples consists of mixed data from the 450k and EPIC arrays, we selected common probes (n = 452,453) between the two arrays. In the next step, we removed sets of probes that consisted of probes on X and Y chromosomes, single nucleotide polymorphism related probes, and probes not uniquely mapped to human reference genomes. After filtering of probes, 357,483 common probes were selected on the EPIC/450k array for further analysis. We performed unsupervised clustering on all these samples and calculated 198 principal components with most variable 20k probes and used these data to create uniform manifold approximation and projection (UMAP).
Literature search
A comprehensive review of the literature was performed to identify all reported cases of IMT with confirmed FET::CREB fusion up to September 2023 in the English literature. Specifically, PubMed was queried with the search terms “intracranial mesenchymal tumor AND fusion”, “angiomatoid fibrous histiocytoma AND intracranial”, and “(intracranial mesenchymal tumor OR angiomatoid fibrous histiocytoma) AND fusion”. All case reports, case series, and review articles that presented new cases of IMT with FET::CREB fusion were included. Additionally, references for all articles were reviewed to evaluate for any further cases that were not revealed in the initial literature search. All available clinical, pathological, genetic, treatment, and outcomes information were extracted for each individual patient. Reports that did not specify both fusion partners were excluded from further analyses.
Statistical analyses
The Kaplan–Meier method was used to estimate the probability of survival. Progression-free survival (PFS) and overall survival (OS) were analyzed for extent of resection (EOR), age, tumor location, and fusion partner. PFS was defined as the time between initial diagnosis and radiographic recurrence or last follow-up. OS was defined as the time from initial diagnosis to death. For each factor, the Log-rank test was performed to compare the survival times among the groups. Cox proportional hazard regression models were used for multivariate analysis. R packages “survival” and “survminer” and GraphPad prism version 9.5.1 (GraphPad Software, San Diego, CA, USA) were used to perform the analyses. A p-value less than 0.05 was considered statistically significant.
Discussion
The 2021 WHO classification of tumors of the central nervous system included a new provisional tumor type termed “intracranial mesenchymal tumor,
FET::CREB fusion-positive” [
27]. These tumors are rare and aspects regarding the potential breadth of clinical behavior are not well-known. In the current study, detailed clinicopathological and molecular findings of five novel cases of IMT,
FET::CREB fusion-positive are reported. In addition, an extensive literature review identified 69 additional published cases of IMT,
FET::CREB fusion-positive. The majority of these cases occurred in children and young adults with a slight female predominance. Evaluation of the clinical outcome of these patients revealed STR, younger age (< 14 years old), infratentorial location, and possible
EWSR1::ATF1 fusion as poor prognostic factors for this recently defined tumor type.
The five new cases of
FET::CREB fusion-positive IMT in the current study occurred in adolescence to early adulthood and showed a wide morphologic spectrum, from spindled/stellate cells of various cellularity (Cases 1, 2 and 5), to sheets of epithelioid cells in Case 3, and to highly-packed small blue cells with frequent mitoses in Case 4. The patients’ outcomes in the current cases were also variable, with patients 1–3 showing a stable course after GTR while patient 4 suffering an early recurrence at 3 months and later developing bilateral pulmonary metastases. These clinicopathologic features are consistent with those previously described in the literature [
35,
36,
40].
In recent years, genome-wide DNA methylation profiling has emerged as a powerful tool for CNS tumor classification. Tauziede-Espariat et al. [
40] and Sloan et al. [
36] were the first two groups to utilize DNA methylation profiling to study IMT,
FET::CREB fusion-positive. Among the 11 cases reported by Tauziede-Espariat et al. [
40], none of them was classifiable using DKFZ brain tumor version 11b4 or sarcoma classifier version 12.2 at publication, although one closely approximated the methylation class of extra-CNS “AFH”, one “clear cell sarcoma”, and two “solitary fibrous tumors”. Similarly, Sloan et al. [
36] reported that only three of their 20 IMTs aligned with the methylation class “AFH” on the DKFZ sarcoma classifier version 12.2 with a calibrated score of greater than 0.9, indicating a high confidence classification. The remaining 17 cases did not reliably classify as “AFH” or any other class on the sarcoma classifier or the CNS tumor classifier version 11b4. However, their 20 cases did resolve into two distinct epigenetic subgroups that were both divergent from all other intracranial tumors and soft tissue sarcomas. In the current cohort of five new IMT cases, four had sufficient tissue for methylation profiling and UMAP embedding analysis. Among those, two cases (Cases 1 and 2) were placed to the methylation class “intracranial mesenchymal tumor”, while the other two cases were placed to different subclasses of meningioma (Case 3 to “benign_3” and Case 4 to “intermediate_A”), which is very intriguing. To the best of our knowledge, no cases in the literature have been classified in a meningioma category before. Although IMTs show distinct genetic alterations from meningiomas, both tumor types frequently present as dural-based lesions with EMA immunoreactivity. In addition, the tumor in Case 4 eventually metastasized to bilateral lungs, which are a common extra-CNS site of metastasis for meningiomas [
37]. Taken together, the possibility of a shared cell of origin between IMT and meningioma cannot be completely excluded. In addition, the presence of a
FET::CREB fusion itself may not be enough to place a tumor to the methylation class “intracranial mesenchymal tumor”, as what happened to Cases 3 and 4 in our cohort. To further investigate this possibility, we compared the two tumors (Cases 1 and 2) that were placed to the methylation class “intracranial mesenchymal tumor” by UMAP embedding and the two tumors (Cases 3 and 4) that did not. The morphology of Cases 1 and 2, spindled/stellate tumor cells arranged in whorls, small nests, or cords in a myxoid or collagenous stroma, was similar in appearance to AFH. In contrast, the majority of tumor cells in Case 3 were epithelioid with significant nuclear pleomorphism and Case 4 featured hypercellular, tightly-packed small round cells with scattered karyorrhectic debris. It is possible that some genetic alterations other than the
FET::CREB fusion have led to the divergent morphological and epigenetic phenotypes of Cases 3 and 4. Further studies with larger cohorts are necessary to explore the full epigenetic spectrum of IMT,
FET::CREB fusion-positive and its relationship to AFH, clear cell sarcoma, solitary fibrous tumor, meningioma, and the other tumors.
To date, little is known about the prognostic factors for patients with IMT,
FET::CREB fusion-positive. Sloan et al. tried to analyze the effects of EOR and mucin-rich versus mucin-poor stroma on prognosis [
35]. However, due to limited numbers of cases (20 cases in their own cohort plus 18 cases from the literature review) included in their analysis, no statistically significant difference was observed, although STR seemed to be associated with an increased risk of death and tumor recurrence. Here, with a larger cohort of 74 patients (5 new cases and 69 cases from the literature review), we identified EOR as a significant prognostic factor for IMT,
FET::CREB fusion-positive. Kaplan–Meier analysis revealed that STR led to significantly shorter PFS and OS compared with GTR. Multivariable Cox regression analysis further confirmed STR as an independent prognostic factor associated with both inferior PFS and OS. Our results suggested that GTR should be achieved whenever possible in patients with IMT,
FET::CREB fusion-positive for the best outcome.
In addition to EOR, our study identified age as a significant prognostic factor for IMT,
FET::CREB fusion-positive. Patients younger than 14 years old had a significantly shorter PFS compared with patients of 14 years or older. Multivariable Cox regression analysis again confirmed a younger age as an independent risk factor associated with inferior PFS. Our results are consistent with what Sloan et al. reported on their epigenetic classification of IMT,
FET::CREB fusion-positive [
36]. In their study, they analyzed their cohort of 20 patients by genome-wide DNA methylation array profiling and identified two distinct epigenetic subgroups. They found that Group B tumors, which occurred most often in early childhood (median age 7 years, range 4–15 years) had an inferior PFS relative to Group A tumors, which occurred frequently in adolescence or early adulthood (median age 15 years).
Our study also suggested
EWSR1::ATF1 fusion as a possible prognostic factor for IMT,
FET::CREB fusion-positive. Among the 74 IMT cases reported to date, three patients eventually died of disease, all of whom underwent a STR for an
EWSR1::ATF1 fused tumor. Sloan et al. analyzed patient survival (OS and PFS) stratified by fusion type (
EWSR1::CREB1,
EWSR1::CREM, and
EWSR1::ATF1) but did not find a statistically significant difference [
35]. In their later analysis with epigenetic data, they found that Group A, which had a favorable outcome, contained mostly
EWSR1::ATF1 and
EWSR1-CREB1 fusions, while Group B composed of more
CREM fused tumors (either
EWSR1::CREM or
FUS::CREM) [
36]. Here, we report that although
EWSR1::ATF1 fusion did not impact PFS, tumors with
EWSR1::ATF1 fusion did show a statistically shorter OS when compared with tumors harboring other fusions. Since we still only have three patients who died of disease, this finding needs to be interpreted with caution. Further study with a larger cohort is necessary to evaluate the precise prognostic effect of different fusion types.
Besides EOR, age, and EWSR1::ATF1 fusion, we further identified another novel prognostic factor: tumor location. Although most IMT cases arose in the supratentorial areas, infratentorial tumors demonstrated shorter OS compared with their supratentorial counterparts by Kaplan–Meier analysis. However, no significant difference was observed in PFS.
In conclusion, the findings of the current study confirm that IMT, FET::CREB fusion-positive is a locally aggressive tumor with a high recurrence rate (~ 40%). The results also suggest that IMT, FET::CREB fusion-positive can be risk-stratified by several basic clinicopathologic parameters. Potential risk factors include subtotal resection, younger age, infratentorial location, and possibly EWSR1::ATF1 fusion. Larger case series are needed to better define prognostic determinants in this unique tumor type.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.