Two novel tumours with NTRK2 fusion in the methylation class of extraventricular neurocytomas, including one intraventricular

Abstract

We report here about two novel tumours classified as extraventricular neurocytomas (EVN) using DNA‐methylation profiling, associated with NTRK2 fusions instead of the usual FGFR1 alterations so far attributed to this tumoural entity. We present the second detailed case of an intraventricular presentation in the MC EVN. Our findings broaden the spectrum of MC EVN and have implications in terms of diagnosis, therapy and terminology.

We report here about two novel tumours classified as extraventricular neurocytomas (EVN) using DNA‐methylation profiling, associated with NTRK2 fusions instead of the usual FGFR1 alterations so far attributed to this tumoural entity [1, 2].

Case 1 concerned a 23‐year‐old woman who had a personal history of migraines, with revealing generalised epileptic seizures. MRI showed a centimetric lesion in T2 and FLAIR hypersignal in the posterior part of the left superior frontal gyrus, without spontaneous T1 hypersignal and a light contrast enhancement on injected T1 sequences. The imaging was stable for 4 years and the patient was asymptomatic under antiepileptic monotherapy. Over the next 3 years, a progressive worsening of the epilepsy was noted (under triple therapy) as well as increased contrast enhancement of the lesion (Figure 1A). A total removal of the lesion was then performed. At 14 months post‐operatively, no neurological deficits or epileptic seizures were observed.

FIGURE 1.

MRI on the T1 sequence shows discrete contrast enhancement of cortical (case 1) (A) or bilateral intraventricular (case 2) (B) lesions. CNV prediction from genome‐wide methylation analysis with http://www.bioconductor.org/packages/devel/bioc/html/conumee.html, showing a flat profile except for chromosome 4 loss in case 1 (C) and a more complex profile in case 2 (D). Schematic representation of NTRK2 fusions (E) in cases 1 and 2 alongside the case reported by Gubiotti et al. [6]. t‐SNE analysis (F). H&E showed oligo‐like component lying in a neurofibrillary background in case 1 (G) with ganglionic sometimes binucleated cells in case 2 (H). Neurofilaments showed a less circumscribed tumour in case 1 (I) than in case 2 (J). Synaptophysin positivity was observed in both cases: case 1 (K) and case 2 (L) as a strong nuclear staining with olig 2: case 1 (M) and case 2 (N).

Histopathologically, the tumour had an oligo‐like component lying in a neurofibrillary background with strong olig2 and synaptophysin positivities (Figure 1G,K,M). The neurofilament network was poor in the centre of the tumour and denser at the periphery in favour of an at less partially circumscribed glioma. ATRX expression was preserved and GFAP was expressed. Mib1 proliferative index was below 5%. Diagnoses of pilocytic astrocytoma, ganglioglioma, gangliocytoma, diffuse leptomeningeal glioneuronal tumour or glioneuronal tumour kinase‐fused (GNT_KinF_A) could be discussed, highlighting the previously reported wide overlap of EVN with various glioneuronal entities [1, 2, 3, 4].

In case 2, intracerebral calcifications were detected on profile skull x‐rays during orthodontic treatment in an asymptomatic 15‐year‐old girl. MRI showed a pure intraventricular lesion in the frontal horn of the left lateral ventricle, extended to both lateral ventricles through an invasion of the septum pellucidum. This did not affect the corpus callosum. Contrast enhancement was seen in the septum pellucidum and along the outer wall of the frontal horn of the left lateral ventricle (Figure 1B). A total removal of the lesion was performed. Microscopically, the tumour had two components: one was oligo‐like, while the other corresponded to more differentiated cells with ganglionic appearance suggestive of ganglioglioma (Figure 1H). The neurofilament network showed a well circumscribed tumour (Figure 1J). Both components were strongly stained by synaptophysin and olig2 (Figure 1L,N). ATRX expression was preserved and GFAP was expressed. Mib1 proliferative index was below 5%. Both ganglionic cell differentiation and microcalcifications have been described in EVN [1, 2]. Five years after the initial surgery, there has been no recurrence and the patient is in good health.

For the two cases, the targeted NGS sequencing was negative, and RNA‐sequencing retrieved a canonic NTRK2 fusion with SOX6 in case 1 and PXDC1 in case 2 (Figure 1E). Genome‐wide DNA methylation profiling indicated a diagnostic score of >0.9 for the extraventricular neurocytoma methylation class (MC) in v12.5 for both tumours (0.93119 and 0.93630 for case 1 and 2, respectively) [5]. We collected the IDAT files of 7 FGFR1::TACC1 fusion EVN (4 from the RENOCLIP Network and 3 from the GEO Database) to enrich the DKFZ Database and perform a t‐SNE [5]. Both cases were in the vicinity of other EVN (Figure 1F). A previous case of EVN with a KIFF5B::NTRK2 fusion was recently reported and u‐MAP analysis also placed this case in the EVN methylation class [6].

Despite definition of the MC EVN from a series of initially histopathological diagnoses of EVN, this nosology seems to present key issues [2]. First, the study describing the MC EVN lacked a central neuroradiological review confirming the absence of ventricular contact of these tumours. Additionally, 7/26 (27%) DNA‐methylation proven cases have no information concerning their precise location in the CNS [2]. Recently, Sato et al. reported a left intraventricular tumour in a 30‐year‐old woman with histology suggestive of central neurocytoma (but the authors did not perform olig2 immunohitochemistry), FGFR1‐TACC1 fusion and a methylation score of 0.99 for the MC class of EVN [7]. This case and the current case 2 seem to suggest that a subset of tumours classified as MC EVN are not extraventricular. The cellular origin of EVN is unknown, but the fact that they form a distinct molecular group suggests that these tumours may arise from a specific precursor cell population, differing from the cells that give rise to central neurocytomas. Unlike the ventricular precursors of central neurocytomas, this specific cell population could possibly be extra or intraventricular.

Furthermore, the terminology ‘neurocytoma’ suggests that tumour cells express neuronal markers such as synaptophysin, but no glial markers such as Olig2. The current WHO classification states that olig2 expression has only been observed in a subset of extraventricular neurocytomas, but the cited publications report very few cases and date from a pre‐methyloma era. Although the only article reporting a series of 26 MC EVN cases does not mention olig2 status, one published case of MC EVN and our two cases showed immunopositivity for this glial marker [2, 6].

The MC of EVN was reported as enriched in FGFR1::TACC1 fusion [2]. Since they were first described, one case of MC EVN in the literature and the two current cases have been reported to harbour NTRK2 fusions with three different genes: KIFF5B, SOX6 in the current case 1 and PXDC1 in the current case 2 (Figure 1E) [6]. The KIFF5B::NTRK2 fusion was also reported in a GNT_KinF_A [3]. The SOX6::NTRK2 was previously reported in one glioneuronal tumour with ATRX alteration, kinase fusion and anaplastic features (GTAKA) whereas no PXDC1::NTRK2 fusion has previously been described to our knowledge [8]. Several low‐grade glioneuronal tumours share many overlapping histopathological (including neurocytic and oligodendroglioma‐like morphology) and molecular features (including FGFR1 and NTRK2 fusions) [2, 3, 6, 8]. Further comprehensive studies including radiology, histopathology and genetics are needed to better characterize these tumours.

With these 2 novel cases, we have validated NTRK2 fusions as alternative driver mutations for the MC EVN, which can be used as potential therapeutic targets. We present the second detailed case of an intraventricular presentation in the MC EVN. Our findings broaden the spectrum of MC EVN and have implications in terms of diagnosis, therapy and terminology. Further studies including radiology, histopathology, genetic and epigenetic analyses are necessary to suggest a consensual nosology proposition for these tumours.

DATA AVAILABILITY STATEMENT

Data sharing is not applicable to this article as no new data were created or analyzed in this study.