Deep learning-based algorithm for postoperative glioblastoma MRI segmentation: a promising new tool for tumor burden assessment.

OBJECTIVE: Clinical and surgical decisions for glioblastoma patients depend on a tumor imaging-based evaluation. Artificial Intelligence (AI) can be applied to magnetic resonance imaging (MRI) assessment to support clinical practice, surgery planning and prognostic predictions. In a real-world context, the current obstacles for AI are low-quality imaging and postoperative reliability. The aim of this study is to train an automatic algorithm for glioblastoma segmentation on a clinical MRI dataset and to obtain reliable results both pre- and post-operatively. METHODS: The dataset used for this study comprises 237 (71 preoperative and 166 postoperative) MRIs from 71 patients affected by a histologically confirmed Grade IV Glioma. The implemented U-Net architecture was trained by transfer learning to perform the segmentation task on postoperative MRIs. The training was carried out first on BraTS2021 dataset for preoperative segmentation. Performance is evaluated using DICE score (DS) and Hausdorff 95% (H95). RESULTS: In preoperative scenario, overall DS is 91.09 (± 0.60) and H95 is 8.35 (± 1.12), considering tumor core, enhancing tumor and whole tumor (ET and edema). In postoperative context, overall DS is 72.31 (± 2.88) and H95 is 23.43 (± 7.24), considering resection cavity (RC), gross tumor volume (GTV) and whole tumor (WT). Remarkably, the RC segmentation obtained a mean DS of 63.52 (± 8.90) in postoperative MRIs. CONCLUSIONS: The performances achieved by the algorithm are consistent with previous literature for both pre-operative and post-operative glioblastoma's MRI evaluation. Through the proposed algorithm, it is possible to reduce the impact of low-quality images and missing sequences.

A novel 3D anatomical visualization system to avoid injuries of nerves in retrosigmoid approach.

The retro-sigmoid approach (RA), widely used during different neurosurgical procedures, is burdened by the risk of injuries of the nerves that cross that region contributing to possible postoperative complications. By using, anatomage table (AT), a novel 3D anatomical visualization system, we described the nerves passing through the retromastoid area including the great occipital nerve (GON), the lesser occipital nerve (LON) and the great auricular nerve (GAN), and their courses from the origins, till terminal branches. Moreover, using dedicated software, we measured distances between the nerves and well-recognizable bony landmarks. After identifying the nerves and their distances from bony landmarks, we observed that the safest and risk-free skin incision should be made in an area delimited, superiorly from the superior nuchal line (or slightly higher), and inferiorly from a plane passing at 1-1.5 cm above the mastoid tip. The lateral aspect of such an area should not exceed 9.5-10 cm from the inion, while the medial one should be more than 7 cm far from the inion. This anatomical information has been useful in defining anatomical landmarks and reducing the risk of complications, mainly related to nerve injury, in RA. In-depth neuroanatomic knowledge of the cutaneous nerves of the retromastoid area is essential to minimize the complications related to their injury during different neurosurgical approaches. Our findings suggest that the AT is a reliable tool to enhance understanding of the anatomy, and thus contributing to the refinement of surgical techniques.

Primary Intracranial Gliosarcoma: Is It Really a Variant of Glioblastoma? An Update of the Clinical, Radiological, and Biomolecular Characteristics.

Gliosarcomas (GS) are sporadic malignant tumors classified as a Glioblastoma (GBM) variant with IDH-wild type phenotype. It appears as a well-circumscribed lesion with a biphasic, glial, and metaplastic mesenchymal component. The current knowledge about GS comes from the limited literature. Furthermore, recent studies describe peculiar characteristics of GS, such as hypothesizing that it could be a clinical-pathological entity different from GBM. Here, we review radiological, biomolecular, and clinical data to describe the peculiar characteristics of PGS, treatment options, and outcomes in light of the most recent literature. A comprehensive literature review of PubMed and Web of Science databases was conducted for articles written in English focused on gliosarcoma until 2023. We include relevant data from a few case series and only a single meta-analysis. Recent evidence describes peculiar characteristics of PGS, suggesting that it might be a specific clinical-pathological entity different from GBM. This review facilitates our understanding of this rare malignant brain tumor. However, in the future we recommend multi-center studies and large-scale metanalyses to clarify the biomolecular pathways of PGS to develop new specific therapeutic protocols, different from conventional GBM therapy in light of the new therapeutic opportunities.