Glioblastoma, IDH-Wildtype Harboring a KIAA1549::BRAF Fusion: Report of a Case With Comprehensive Molecular Profiling

To the Editor:

KIAA1549::BRAF fusions have been reported in adult-type diffuse gliomas (1, 2) but such cases have not been comprehensively characterized. We describe a case of glioblastoma, IDH-wildtype harboring a KIAA1549::BRAF fusion that has been characterized by sequencing, copy number, and methylation profiling.

A 52-year-old man presented with a 1-month history of headaches and confusion. Family history was significant for an uncle who reportedly had a glioblastoma. Imaging studies revealed a right parietal enhancing solid-cystic tumor (Fig. 1A). Gross total resection was performed (Fig. 1B) and revealed a mitotically active infiltrating astrocytoma with microvascular proliferation and necrosis (Fig. 1D, E). Rosenthal fibers and eosinophilic granular bodies were absent. No areas with low-grade features or solid non-infiltrating growth pattern that would suggest a precursor lesion were observed. IDH1-R132H immunostain was negative and ATRX immunostain showed preserved protein expression. Targeted 187-gene neuro-oncology mutation (using DNA) and fusion (using RNA with RT-PCR) next-generation sequencing panel detected TERT promoter, PTEN, and RPL5 mutations at 41%, 38%, and 20% variant allele frequency, respectively (tumor purity visually estimated as 70%), and no IDH or histone H3 mutations were identified. A KIAA1549::BRAF fusion (exon 16::exon 9) was also detected and confirmed on re-extracted RNA, excluding the possibility of sample contamination/mix-up. OncoScan chromosomal microarray revealed multiple copy number changes including a gain of whole chromosome 7 along with the additional gain of 7q34, loss of entire chromosome 10, and CDKN2A/CDKN2B homozygous deletion (Fig. 2A, B). The 7q34 gain appeared to disrupt KIAA1549 and BRAF as the proximal and distal breakpoints seemed to be within each gene, within the confidence of the detection limits of OncoScan chromosomal microarray testing platform (Fig. 2C). By genome-wide methylation array, this tumor matched to the “Glioblastoma, IDH-wildtype, RTK1” methylation subclass with a calibrated score of 0.999 (classifier version 12.5). Collectively, the mutational, copy number, and methylation pattern in this case were in keeping with a glioblastoma, IDH-wildtype with the unusual finding of a KIAA1549::BRAF fusion. MGMT promoter methylation derived from methylation array and by real-time PCR interrogating 8 downstream CpG sites was negative. Postoperatively, the patient opted for radiotherapy only, which was followed by Optune tumor treating fields with minimal compliance due to severe fatigue. Eight months after diagnosis, he presented with progressive left upper extremity paresis due to tumor recurrence (Fig. 1C). The patient underwent reirradiation and was started on bevacizumab and temozolomide. He has clinically progressive disease at 13 months.

FIGURE 1.

Radiological findings (post-contrast T1-weighted MRI). (A) Preoperative: 5.2 cm enhancing complex solid-cystic tumor in the right parietal lobe associated with right-to-left midline shift. (B) Immediate postoperative: Surgical cavity after gross total tumor resection. (C) Eight-month postoperative: Tumor recurrence with increased areas of enhancement along the resection cavity and surrounding edema in the brain parenchyma. Histopathological findings (H&E, 200×). (D, E) High-grade diffuse astrocytic glioma showing (D) microvascular proliferation and (E) tumor necrosis.

FIGURE 2.

Copy number profile by chromosomal microarray. (A) Combined gain of chromosome 7 and loss of chromosome 10 (+7/−10), CDKN2A/CDKN2B homozygous deletion (asterisk) and 7q34 gain (black arrow) among other partial/whole chromosomal copy number changes. (B) Whole chromosome 7 gain and a focal superimposed gain of 7q34 (grey curtain) that (C) appeared to disrupt KIAA1549 and BRAF with the proximal and distal breakpoints indicated by the black solid lines; the probes (burgundy dots) between the 2 solid black lines seemed to consistently be at a higher level than the adjacent probes.

KIAA1549::BRAF fusions are characteristic of pilocytic/pilomyxoid astrocytoma and diffuse leptomeningeal glioneuronal tumor (3) and among the diagnostically supportive MAPK pathway alterations of high-grade astrocytoma with piloid features (4). These fusions have also been reported in 17 (of 180; 9.4%) adult-type diffuse gliomas which were histologically classified without methylation profiling and tested by RT-PCR/Sanger sequencing: 11 IDH-mutant tumors—6 also 1p/19q-codeleted; 5 IDH-wildtype tumors—1 histological glioblastoma who survived 8 months; and 1 tumor without IDH status (1). Two additional cases classified by methylation profiling (1 glioblastoma, IDH wildtype, subclass mesenchymal, and 1 IDH-mutant glioma, subclass high-grade astrocytoma) were reported to show the 7q34 gain indicative of a KIAA1549::BRAF fusion by automated screening of the methylation array-based copy number data; however, the detected 7q34 gains were reportedly not entirely prototypical by visual evaluation and fusion confirmation was not performed (2). The relatively high frequency of KIAA1549::BRAF fusions reported in histologically defined cases was not replicated by larger comprehensive multiplatform histomolecular studies of adult-type diffuse gliomas (5, 6). The reason for this discordance is unclear for IDH-mutant tumors as tumor misclassification is unlikely; for IDH-wildtype diffuse gliomas, KIAA1549::BRAF fusion frequency in histologically assessed tumors could have been overestimated by tumor misclassification because methylation profiling data were unavailable and alternative diagnoses, such as high-grade astrocytoma with piloid features had not yet been described.

Although overlapping molecular features with high-grade astrocytoma with piloid features, including the KIAA1549::BRAF fusion and CDKN2A/CDKN2B homozygous deletion, were observed, this reported case has the characteristic mutational, copy number, and methylation profile of a glioblastoma, IDH-wildtype. The detected KIAA1549::BRAF fusion junction encompassed KIAA1549 exon 16 and BRAF exon 9 like the canonical KIAA1549::BRAF fusions. Also like the typical KIAA1549::BRAF fusions seen in pilocytic/pilomyxoid astrocytoma and diffuse leptomeningeal glioneuronal tumor, the 7q34 gain by chromosomal microarray had proximal 5′ and distal 3′ breakpoints that appeared to involve and disrupt KIAA1549 and BRAF, respectively.

In pediatric low-grade gliomas, KIAA1549::BRAF fusions are associated with favorable prognosis (7). Conversely, the clinical significance of a KIAA1549::BRAF fusion in the context of adult-type diffuse gliomas, including glioblastoma, IDH-wildtype, is not fully understood. The single reported case with available follow-up data was histologically defined and had an aggressive clinical course (1). The progressive clinical evolution of our patient also argues against a favorable clinical impact in this clinicopathological context. Preliminary efficacy data of MEK inhibitors for pediatric patients with refractory/progressive KIAA1549::BRAF fusion-positive low-grade gliomas are encouraging (8, 9). Although not directly comparable, it may be extrapolated that such targeted therapies could represent an alternative therapeutic option for other KIAA1549::BRAF fusion-positive tumors like this glioblastoma, IDH-wildtype, especially in the context of a negative MGMT promoter methylation test result.

In conclusion, this case demonstrates that KIAA1549::BRAF fusions rarely occur in glioblastoma, IDH-wildtype, and underscores the power of routine clinical molecular profiling in unraveling unusual patterns as well as the complementarity of multi-platform molecular profiling in characterizing such unusual patterns encountered with increasing brain tumor molecular testing.

COMPETING INTERESTS

The authors have no duality or conflicts of interest to declare.