Frequent FGFR1 hotspot alterations in driver-unknown low-grade glioma and mixed neuronal-glial tumors

Purpose Low-grade gliomas (LGG) and mixed neuronal-glial tumors (MNGT) show frequent MAPK pathway alterations. Oncogenic fibroblast growth factor receptor 1 (FGFR1) tyrosinase kinase domain has been reported in brain tumors of various histologies. We sought to determine the frequency of FGFR1 hotspot mutations N546 and K656 in driver-unknown LGG/MNGT and examined FGFR1 immunohistochemistry as a potential tool to detect those alterations. Methods We analyzed 476 LGG/MNGT tumors for KIAA-1549-BRAF fusion, IDH1/2, TERT promotor, NF1, H3F3A and the remaining cases for FGFR1 mutation frequency and correlated FGFR1 immunohistochemistry in 106 cases. Results 368 of 476 LGG/MNGT tumors contained non-FGFR1 alterations. We identified 9 FGFR1 p.N546K and 4 FGFR1 p.K656E mutations among the 108 remaining driver-unknown samples. Five tumors were classified as dysembryoplastic neuroepithelial tumor (DNT), 4 as pilocytic astrocytoma (PA) and 3 as rosette-forming glioneuronal tumor (RGNT). FGFR1 mutations were associated with oligodendroglia-like cells, but not with age or tumor location. FGFR1 immunohistochemical expression was observed in 92 cases. FGFR1 immunoreactivity score was higher in PA and DNT compared to diffuse astrocytoma, but no correlation between FGFR1 mutation in tumors and FGFR1 expression level was observed. Conclusion FGFR1 hotspot mutations are the fifth most prevailing alteration in LGG/MNGT. Performing FGFR1 sequencing analysis in driver-unknown low-grade brain tumors could yield up to 12% FGFR1 N546/K656 mutant cases.

Challenges in the Intraoperative Consultation of Low-Grade Epilepsy-Associated Neuroepithelial Tumors by Cytomorphology in Squash Preparations

<b><i>Introduction:</i></b> Low-grade epilepsy-associated neuroepithelial tumors (LEATs) create a diagnostic challenge in daily practice and intraoperative pathological consultation (IC) in particular. Squash smears are extremely useful in IC for accurate diagnosis; however, the knowledge on cytopathologic features of LEATs is based on individual case reports. Here, we discuss the 3 most common and well-established entities of LEATs: ganglioglioma (GG), dysembryoplastic neuroepithelial tumor (DNT), and papillary glioneuronal tumor (PGNT). <b><i>Methods:</i></b> Thirty patients who underwent surgery for GG, DNT, and PGNT between 2001 and 2021 were collected. Squash smears prepared during intraoperative consultation were reviewed by 1 cytopathologist and an experienced neuropathologist. <b><i>Results:</i></b> Among the 30 tumors, 16 (53.3%) were GG, 11 (36.6%) DNT, and 3 (10%) PGNT. Cytomorphologically, all of the 3 tumor types share 2 common features such as dual cell population and vasculocentric pattern. GG smears were characteristically composed of dysplastic ganglion cells and piloid-like astrocytes on a complex architectural background of thin- to thick-walled vessels. DNT, on the other hand, showed oligodendroglial-like cells in a myxoid thin fibrillary background associated with a delicate capillary network. Common cytological features of PGNT were hyperchromatic cells with narrow cytoplasm surrounding hyalinized vessels forming a pseudopapillary pattern and bland cells with neuroendocrine nuclei dispersed in a neuropil background. <b><i>Conclusion:</i></b> A higher diagnostic accuracy can be obtained when squash smears are applied with frozen sections. However, it is important to integrate clinical and radiologic features of the patient as well as to know the cytopathologic features of the LEAT spectrum in the context of differential diagnosis to prevent misinterpretation in the IC.

A Whole-Slide Image Managing Library Based on Fastai for Deep Learning in the Context of Histopathology: Two Use-Cases Explained

Background: Processing whole-slide images (WSI) to train neural networks can be intricate and labor intensive. We developed an open-source library dealing with recurrent tasks in the processing of WSI and helping with the training and evaluation of neuronal networks for classification tasks. Methods: Two histopathology use-cases were selected and only hematoxylin and eosin (H&E) stained slides were used. The first use case was a two-class classification problem. We trained a convolutional neuronal network (CNN) to distinguish between dysembryoplastic neuroepithelial tumor (DNET) and ganglioglioma (GG), two neuropathological low-grade epilepsy-associated tumor entities. Within the second use case, we included four clinicopathological disease conditions in a multilabel approach. Here we trained a CNN to predict the hormone expression profile of pituitary adenomas. In the same approach, we also predicted clinically silent corticotroph adenoma. Results: Our DNET-GG classifier achieved an AUC of 1.00 for the ROC curve. For the second use case, the best performing CNN achieved an area under the curve (AUC) of 0.97 for the receiver operating characteristic (ROC) for corticotroph adenoma, 0.86 for silent corticotroph adenoma, and 0.98 for gonadotroph adenoma. All scores were calculated with the help of our library on predictions on a case basis. Conclusions: Our comprehensive and fastai-compatible library is helpful to standardize the workflow and minimize the burden of training a CNN. Indeed, our trained CNNs extracted neuropathologically relevant information from the WSI. This approach will supplement the clinicopathological diagnosis of brain tumors, which is currently based on cost-intensive microscopic examination and variable panels of immunohistochemical stainings.

A Whole Slide Image Managing Library Based on Fastai for Deep Learning in the Context of Histopathology: Two Use-Cases Explained

Background: Processing whole-slide images (WSI) to train neural networks can be intricate and laborious. We developed an open-source library covering recurrent tasks in processing of WSI and in evaluating the performance of the trained networks for classification tasks. Methods: Two histopathology use-cases were selected. First we aimed to train a CNN to distinguish H&amp;amp;E-stained slides obtained from neuropathologically classified low-grade epilepsy-associated dysembryoplastic neuroepithelial tumor (DNET) and ganglioglioma (GG). The second project we trained a convolutional neural network (CNN) to predict the hormone expression of pituitary adenoms only from hematoxylin and eosin (H&amp;amp;E) stained slides. In the same approach, we addressed the issue to also predict clinically silent corticotroph adenoma. We included four clinico-pathological disease conditions in a multilabel approach. Results: Our best performing CNN achieved an area under the curve (AUC) of 0.97 for the receiver operating characteristic (ROC) for corticotroph adenoma, 0.86 for silent corticotroph adenoma and 0.98 for gonadotroph adenoma. Our DNET-GG classifier achieved an AUC of 1.00 for the ROC curve. All scores were calculated with the help of our library on predictions on a case basis. Conclusions: Our comprehensive library is most helpful to standardize the work-flow and minimize the work-burden in training CNN. It is also compatible with fastai. Indeed, our new CNNs reliably extracted neuropathologically relevant information from the H&amp;amp;E staining only. This approach will supplement the clinico-pathological diagnosis of brain tumors, which is currently based on cost-intense microscopic examination and variable panels of immunohistochemical stainings.

Magnetic Resonance-Guided Laser Interstitial Thermal Therapy (MR-gLiTT) in Pediatric Epilepsy Surgery: State of the Art and Presentation of Giannina Gaslini Children's Hospital (Genoa, Italy) Series

Magnetic resonance-guided laser interstitial thermal therapy (MR-gLiTT) is a novel minimally invasive treatment approach for drug-resistant focal epilepsy and brain tumors. Using thermal ablation induced by a laser diode implanted intracranially in a stereotactic manner, the technique is highly effective and safe, reducing the risk associated with more traditional open surgical approaches that could lead to increased neurological morbidity. Indications for MR-gLiTT in pediatric epilepsy surgery include hypothalamic hamartoma, tuberous sclerosis complex, cavernoma-related epilepsy, SEEG-guided seizure onset zone ablation, corpus callosotomy, periventricular nodular heterotopia, mesial temporal lobe epilepsy, and insular epilepsy. We review the available literature on the topic and present our series of patients with drug-resistant epilepsy treated by MR-gLiTT. Our experience, represented by six cases of hypothalamic hamartomas, one case of tuberous sclerosis, and one case of dysembryoplastic neuroepithelial tumor, helps to confirm that MR-gLiTT is a highly safe and effective procedure for several epilepsy conditions in children.

A Whole Slide Image Managing Library Based on Fastai for Deep Learning in the Context of Histopathology: Two Use-Cases Explained

Background: Processing whole-slide images (WSI) to train neural networks can be intricate and laborious. We developed an open-source library covering recurrent tasks in processing of WSI and in evaluating the performance of the trained networks for classification tasks. Methods: Two histopathology use-cases were selected. First we aimed to train a CNN to distinguish H&amp;amp;E-stained slides obtained from neuropathologically classified low-grade epilepsy-associated dysembryoplastic neuroepithelial tumor (DNET) and ganglioglioma (GG). The second project we trained a convolutional neural network (CNN) to predict the hormone expression of pituitary adenoms only from hematoxylin and eosin (H&amp;amp;E) stained slides. In the same approach, we addressed the issue to also predict clinically silent corticotroph adenoma. We included four clinico-pathological disease conditions in a multilabel approach. Results: Our best performing CNN achieved an area under the curve (AUC) of 0.97 for the receiver operating characteristic (ROC) for corticotroph adenoma, 0.86 for silent corticotroph adenoma and 0.98 for gonadotroph adenoma. Our DNET-GG classifier achieved an AUC of 1.00 for the ROC curve. All scores were calculated with the help of our library on predictions on a case basis. Conclusions: Our comprehensive library is most helpful to standardize the work-flow and minimize the work-burden in training CNN. It is also compatible with fastai. Indeed, our new CNNs reliably extracted neuropathologically relevant information from the H&amp;amp;E staining only. This approach will supplement the clinico-pathological diagnosis of brain tumors, which is currently based on cost-intense microscopic examination and variable panels of immunohistochemical stainings.

5-Aminolevulinic acid–assisted resection of pediatric dysembryoplastic neuroepithelial tumor: illustrative case

BACKGROUND 5-Aminolevulinic acid (5-ALA) is approved as an adjunct for the resection of high-grade gliomas and is associated with improved outcomes. Dysembryoplastic neuroepithelial tumors (DNETs) are benign glioneural tumors occurring primarily in pediatric patients and often manifesting with seizure disorder. The goal of the surgical intervention is to obtain gross-total resection, which is associated, in the majority of cases, with seizure freedom. OBSERVATIONS The authors present the first case report of a pediatric patient who underwent gross-total resection of a 5-ALA–positive DNET with no evidence of recurrent seizures (Engel class I). LESSONS Fluorescence-guided surgery using 5-ALA facilitated gross-total resection of the mass.

The spectrum of mitochondrial DNA (mtDNA) mutations in pediatric central nervous system (CNS) tumors

Background We previously established the landscape of mitochondrial DNA (mtDNA) mutations in 23 subtypes of pediatric malignancies, characterized mtDNA mutation profiles among these subtypes, and provided statistically significant evidence for a contributory role of mtDNA mutations to pediatric malignancies. Methods To further delineate the spectrum of mtDNA mutations in pediatric CNS tumors, we analyzed 545 tumor-normal paired whole genome sequencing data sets from the Children’s Brain Tumor Tissue Consortium. Results Germline mtDNA variants were used to determine the haplogroup, and maternal ancestry, which was not significantly different among tumor types. Among 166 (30.5%) tumors we detected 220 somatic mtDNA mutations, primarily missense mutations (36.8%), as well as 22 loss-of-function mutations. Different pediatric CNS tumor subtypes had distinct mtDNA mutation profiles. The number of mtDNA mutations per tumor ranged from 0.20 (dysembryoplastic neuroepithelial tumor) to 0.75 (meningiomas). The average heteroplasmy was 10.7%, ranging from 4.6% in atypical teratoid/rhabdoid tumor (AT/RT) to 26% in diffuse intrinsic pontine glioma. High-grade gliomas had a significant higher number of mtDNA mutations per sample than low-grade gliomas (0.6 vs. 0.27) (p = 0.004), with almost twice as many missense mtDNA mutations per sample (0.24 vs. 0.11), and higher average heteroplasmy levels (16% vs. 10%). Recurrent mtDNA mutations may represent hotspots which may serve as biologic markers of disease. Conclusions Our findings demonstrate varying contributions of mtDNA mutations in different subtypes of CNS tumors. Sequencing the mtDNA genome may ultimately be used to characterize CNS tumors at diagnosis and monitor disease progression.