An 8-year-old female presented to the neurosurgery department with intermittent headache, nausea and vomiting for 8 days. Neurological examination revealed no abnormalities. She had no family history of cancer or visceral. Magnetic resonance imaging (MRI) of the brain revealed a heterogeneous mass in left parietal lobe that was predominantly hypointense on T1 (Fig. 1a) and inhomogeneous hyperintense on T2 (Fig. 1b), as well as the fluid attenuation inversion recovery (FLAIR) sequence was iso-intensity (Fig. 1c). After the administration of contrast, the mass displayed heterogeneous enhancement (Fig. 1d, 1e). The patient underwent a craniotomy via left parietal approach, which revealed a well-defined, soft, gray and gelatinous lesion attached to the falx cerebri (5 cm × 5 cm × 5 cm). The lesion was completely resected and postoperative course was uneventful. Postoperative CT scan showed that the intracranial tumor had been completely removed (Fig. 1f).
Light microscopy examination revealed sarcomatous neoplasms predominantly presented with spindle-shaped cells in a fascicular pattern (Fig. 2a). There was a clear boundary between tumor and peripheral brain tissue (Fig. 2b). In some areas, a myxoid stroma matrix was found with a few multinucleated tumor giant cells (Fig. 2c). The nuclei of the tumor cells were with brisk mitotic activity (Fig. 2d). Prominent cytoplasmic eosinophilic globules were easily identified (Fig. 2e, 2f).
Immuno-histochemical studies were performed using an automated Ventana Benchmark Ultra autostainer (Ventana, Tucson, Arizona, USA). Briefly, tissue sections were deparaffinized, antigens retrieved and endogenous peroxidase was blocked with 1% H2O2. The primary antibodies were used for target protein detection: Desmin (clone D33, Dako, 1:50), smooth muscle actin (SMA, clone 1A4, Dako, 1:400), MyoD1 (clone 5.8A, Dako, 1:50), myogenin (clone F5D, Dako, 1:50), CD34 (clone QBEnd10, Beckman Coulter, 1:500), S-100 (clone 4c4.9, Zytomed, 1:3000) and Ki67 (clone MIB1, Dako, 1:80). Olig2 (clone H-68, Santa Cruz, 1:200), MAP2 (Abcam Cambridge; 1:1500), Syn (Bio Genex, 1:100), glial fibrillary acidic protein (GFAP, clone 6F2, Dako Cytomation, 1:500), epithelial membrane antigen (EMA, Dako, 1:50), cytokeratin (CK, clone AE1⁄AE3, Zymed San Francisco, 1:200), NF (clone 2F11, Dako Cytomatio, 1:150), TLE1 (clone sc-9121, Santa Cruz, 1:100), STAT6 (Abcam Cambridge, 1:500), TLE1 (clone sc-9121, Santa Cruz, 1:100), SOX10 (Abcam Cambridge, 1:400), somatostatin receptor 2A (SSTR2A, cloneUMB1, Abcam, 1:200), BRG1 (clone G-7, Santa Cruz, 1:200), INI-1 (clone BAF47, Biosciences, 1:200), DICER1 (clone 4A6, Abcam, 1:100). Revelation was performed using the ultra-VIEW TM DAB systems (Ventana). All tissue sections were counterstained with hematoxylin II/Mayer’s hematoxylin (Ventana). Immunohistochemically, the MIB-1-positive cell index was 70%. The tumor cells stained positively for vimentin. The tumor cells had a focal expression of Desmin (Fig. 2g) and did not express myogenin, SMA. DICER1 protein immunostaining was extensively positive in the tumor cells, raising the possibility of DICER1 mutation (Fig. 2h). Interesting, the tumor cells stained positively for S-100 (Fig. 2i), Syn (Fig. 2j), MAP-2 (Fig. 2k) and NF (Fig. 2l), and INI-1(Fig. 2m) and BRG1 were retained. But tumor cells were negative for GFAP (Fig. 2n) and Olig-2. The Syn, MAP-2, and NF immunostaining highlighted the neural lineage differentiation of tumor cells. SOX10, CK, EMA, CD34, TLE1, STAT6 and SSTR2A were negative. The tumor cells displayed with a large amount of thin reticulin fibers (Fig. 2o). A diagnosis of primary intracranial spindle cell sarcoma with neurogenic and myogenic differentiation, possibly associated with DICER1 mutation, was favoured.
Next, WES analysis was performed in surgical formalin-fixed paraffin-embedded (FFPE) tumor tissues and normal plasm. In brief, WES analysis sequencing libraries were generated using Agilent SureSelect Human All Exon V6 kit (Agilent Technologies, CA, USA). DNA libraries were sequenced on Illumina Hiseq platform and 150 bp paired-end reads were generated and cleaned. Data was mapped to the reference human genome b37 by Burrows-Wheeler Aligner (BWA) software . SAMtools  and Picard (http://broadinstitute.github.io/picard/) were used to sort BAM files and do duplicate marking, local realignment, and base quality recalibration. Samtools mpileup and bcftools were used to do variant calling and identify SNP. Variants were selected by applying more filters on germline variants callsets. DICER1 p.E1813D (c.5439G > T) mutation located within the Ribonuclease III domain was detected in tumor tissues but not in plasm sample suggesting a somatic mutation. A final diagnosis of DCS was established. TP53 mutation (c. 560-2A > T) was found in this case. It harbored RAF1 mutation (p.R191T) which was predicted to cause activation of the MAP kinase signaling pathway. Other tumor-driving mutations including AR mutation (p.G473del), AXL mutation (p.T45P), ETV5 mutation (p.F11Y) were detected. Alterations in all genes (including not reported/undetermined significance = 158) are summarized in (Supplementary Table 1). Then we performed GO analysis to identify the potential molecular function of these driver-genes. GO analysis showed that genes involved in chromatin organization (SATB1, ATAD2 and CHD4), neuron development (TENM4, SECISBP2 and HS6ST1), histone deacetylation (SIN3B, CHD4 and MTA2) processes that were mutated in this case.
Additionally, WES indicated that the tumor presented a high level of tumor mutational burden (TMB), but did not show microsatellite instability (MSI). RNA-seq was performed for as previously described . We did not find any gene fusions by RNA-seq.
With DICER1 p.E1813D (c.5439G > T) mutation confirmed, a final diagnosis of DCS was established. The patient received postoperative radiotherapy (60 Gy/30f) with the following cyclophosphamide, doxorubicin and vincristine for the first, third, fifth and seventh chemotherapy cycles. In the second, fourth, sixth and eighth cycles ifosfamide and etoposide were used for chemotherapy. After eight cycles (median duration of 21 days among cycles) of combination chemotherapy, brain MRI revealed no evidence of tumor recurrence at the 12 months’ follow-up. The present study was conducted in accordance with the Declaration of Helsinki and under the guidelines of the institutional board on ethics of the Sanbo Brain Hospital. Written informed consent was obtained from the family member (mother) of the patient for the publication of any potentially identifiable images or data included in this article.