Key Points Question How does augmentation with a deep learning segmentation model influence the performance of clinicians in identifying intracranial aneurysms from computed tomographic angiography examinations? Findings In this diagnostic study of intracranial aneurysms, a test set of 115 examinations was reviewed once with model augmentation and once without in a randomized order by 8 clinicians. The clinicians showed significant increases in sensitivity, accuracy, and interrater agreement when augmented with neural network model–generated segmentations. Meaning This study suggests that the performance of clinicians in the detection of intracranial aneurysms can be improved by augmentation using deep learning segmentation models.

Hemangioblastomas are rare, benign, vascular tumors of the central nervous system (CNS), often associated with von-hippel lindau (VHL) disease. Current therapeutic options include microsurgical resection or stereotactic radiosurgery (SRS). With no randomized controlled studies and minimal data beyond single-institution reviews, the optimal management approach for patients with CNS hemangioblastomas is unclear. We completed a Pubmed/SCOPUS literature search from January 1990 to January 2017 for eligible studies on SRS for CNS hemangioblastomas. Relevant articles were identified and reviewed in accordance to the Preferred Reporting Items for Systematic Review and Meta-Analysis (PRISMA) guidelines. 26 studies met eligibility criteria for qualitative synthesis, representing 596 subjects and 1535 tumors. The Gamma Knife was the most published SRS method for CNS hemangioblastomas. After critical study appraisal for intra-study bias, 14 studies were used for quantitative meta-analysis of 5-year progression free survival (PFS). The pooled 5-year PFS across all eligible studies was 88.4%. No difference was observed between spine versus intracranial studies. Individual patient data (IPD) was extracted from 14 studies, representing 322 tumors. Univariate analysis of IPD revealed that VHL patients were younger, and had smaller tumors compared to those with sporadic disease. Adverse events were associated with increasing marginal dose, independent of tumor volume. VHL status, sex, radiosurgical method, tumor location, and tumor volume were not found to be significantly associated with tumor progression. Multiple studies show excellent tumor control at 5-year follow up, however, the long-term efficacy of SRS for CNS hemangioblastomas still needs to be investigated, and the studies exploring the role of SRS for early treatment of asymptomatic lesions is wanting.

Background Cerebellar mutism syndrome (CMS) is a common complication following resection of posterior fossa tumors, most commonly after surgery for medulloblastoma. Medulloblastoma subgroups have historically been treated as a single entity when assessing CMS risk; however, recent studies highlighting their clinical heterogeneity suggest the need for subgroup-specific analysis. Here, we examine a large international multicenter cohort of molecularly characterized medulloblastoma patients to assess predictors of CMS. Methods We assembled a cohort of 370 molecularly characterized medulloblastoma subjects with available neuroimaging from 5 sites globally, including Great Ormond Street Hospital, Christian Medical College and Hospital, the Hospital for Sick Children, King Hussein Cancer Center, and Lucile Packard Children’s Hospital. Age at diagnosis, sex, tumor volume, and CMS development were assessed in addition to molecular subgroup. Results Overall, 23.8% of patients developed CMS. CMS patients were younger (mean difference −2.05 years ± 0.50, P = 0.0218) and had larger tumors (mean difference 10.25 cm3 ± 4.60, P = 0.0010) that were more often midline (odds ratio [OR] = 5.72, P < 0.0001). In a multivariable analysis adjusting for age, sex, midline location, and tumor volume, Wingless (adjusted OR = 4.91, P = 0.0063), Group 3 (adjusted OR = 5.56, P = 0.0022), and Group 4 (adjusted OR = 8.57 P = 9.1 × 10−5) tumors were found to be independently associated with higher risk of CMS compared with sonic hedgehog tumors. Conclusions Medulloblastoma subgroup is a very strong predictor of CMS development, independent of tumor volume and midline location. These findings have significant implications for management of both the tumor and CMS.

Background Surgical resection is a mainstay in the treatment of pediatric brain tumors to achieve tissue diagnosis and tumor debulking. While maximal safe resection of tumors is desired, it can be challenging to differentiate normal brain from neoplastic tissue using only microscopic visualization, intraoperative navigation, and tactile feedback. Here, we investigate the potential for Raman spectroscopy (RS) to accurately diagnose pediatric brain tumors intraoperatively. Methods Using a rapid acquisition RS device, we intraoperatively imaged fresh ex vivo brain tissue samples from 29 pediatric patients at the Lucile Packard Children’s Hospital between October 2018 and March 2020 in a prospective fashion. Small tissue samples measuring 2-4 mm per dimension were obtained with each individual tissue sample undergoing multiple unique Raman spectra acquisitions. All tissue samples from which Raman spectra were acquired underwent individual histopathology review. A labeled dataset of 684 unique Raman spectra gathered from 156 samples was then used to develop a machine learning model capable of 1) differentiating normal brain from tumor tissue and 2) normal brain from low grade glioma (LGG) tissue. Results Trained logistic regression model classifiers were developed using our labeled dataset. Model performance was evaluated using leave-one-patient-out cross-validation. The area-under the-curve (AUC) of the receiver-operating-characteristic (ROC) curve for our tumor versus normal brain model was 0.94. The AUC of the ROC curve for LGG versus normal brain was 0.91. Conclusions Our work suggests that RS can be used to develop a machine learning based classifier to differentiate tumor versus non-tumor tissue during resection of pediatric brain tumors.