BACKGROUND AND PURPOSE: The World Health Organization has recently placed new emphasis on the integration of genetic information for gliomas. While tissue sampling remains the criterion standard, noninvasive imaging techniques may provide complimentary insight into clinically relevant genetic mutations. Our aim was to train a convolutional neural network to independently predict underlying molecular genetic mutation status in gliomas with high accuracy and identify the most predictive imaging features for each mutation. MATERIALS AND METHODS: MR imaging data and molecular information were retrospectively obtained from The Cancer Imaging Archives for 259 patients with either low- or high-grade gliomas. A convolutional neural network was trained to classify isocitrate dehydrogenase 1 (IDH1) mutation status, 1p/19q codeletion, and O6-methylguanine-DNA methyltransferase (MGMT) promotor methylation status. Principal component analysis of the final convolutional neural network layer was used to extract the key imaging features critical for successful classification. RESULTS: Classification had high accuracy: IDH1 mutation status, 94%; 1p/19q codeletion, 92%; and MGMT promotor methylation status, 83%. Each genetic category was also associated with distinctive imaging features such as definition of tumor margins, T1 and FLAIR suppression, extent of edema, extent of necrosis, and textural features. CONCLUSIONS: Our results indicate that for The Cancer Imaging Archives dataset, machine-learning approaches allow classification of individual genetic mutations of both low- and high-grade gliomas. We show that relevant MR imaging features acquired from an added dimensionality-reduction technique demonstrate that neural networks are capable of learning key imaging components without prior feature selection or human-directed training.

Chemotherapy-related cognitive impairment (CRCI) is commonly reported following the administration of chemotherapeutic agents and comprises a wide variety of neurological problems. No effective treatments for CRCI are currently available. Here we examined the mechanisms involving cisplatin-induced hippocampal damage following cisplatin administration in a rat model and in cultured rat hippocampal neurons and neural stem/progenitor cells (NSCs). We also assessed the protective effects of the antioxidant, N-acetylcysteine in mitigating these damages. Adult male rats received 6mg/kg cisplatin in the acute studies. In chronic studies, rats received 5mg/kg cisplatin or saline injections once per week for 4 weeks. N-acetylcysteine (250mg/kg/day) or saline was administered for five consecutive days during cisplatin treatment. Cognitive testing was performed 5 weeks after treatment cessation. Cisplatin-treated cultured hippocampal neurons and NSCs were examined for changes in mitochondrial function, oxidative stress production, caspase-9 activation, and neuronal dendritic spine density. Acute cisplatin treatment reduced dendritic branching and spine density, and induced mitochondrial degradation. Rats receiving the chronic cisplatin regimen showed impaired performance in contextual fear conditioning, context object discrimination, and novel object recognition tasks compared to controls. Cisplatin induced mitochondrial DNA damage, impaired respiratory activity, increased oxidative stress, and activated caspase-9 in cultured hippocampal neurons and NSCs. N-acetylcysteine treatment prevented free radical production, ameliorated apoptotic cellular death and dendritic spine loss, and partially reversed the cisplatin-induced cognitive impairments. Our results suggest that mitochondrial dysfunction and increased oxidative stress are involved in cisplatin-induced cognitive impairments. Therapeutic agents, such as N-acetylcysteine, may be effective in mitigating the deleterious effects of cisplatin.

The proteasome plays a vital role in the physiology of glioblastoma (GBM), and proteasome inhibition can be used as a strategy for treating GBM. Marizomib is a second-generation, irreversible proteasome inhibitor with a more lipophilic structure that suggests the potential for penetrating the blood-brain barrier. While bortezomib and carfilzomib, the 2 proteasome inhibitors approved for treatment of multiple myeloma, have little activity against malignant gliomas in vivo, marizomib could be a novel therapeutic strategy for primary brain tumors. The in-vitro antitumor activity of marizomib was studied in glioma cell lines U-251 and D-54. The ability of marizomib to cross the blood-brain barrier and regulate proteasome activities was evaluated in cynomolgus monkeys and rats. The antitumor effect of marizomib in vivo was tested in an orthotopic xenograft model of human GBM. Marizomib inhibited the proteasome activity, proliferation, and invasion of glioma cells. Meanwhile, free radical production and apoptosis induced by marizomib could be blocked by antioxidant N-acetyl cysteine. In animal studies, marizomib distributed into the brain at 30% of blood levels in rats and significantly inhibited (>30%) baseline chymotrypsin-like proteasome activity in brain tissue of monkeys. Encouragingly, the immunocompromised mice, intracranially implanted with glioma xenografts, survived significantly longer than the control animals (P < .05) when treated with marizomib. These preclinical studies demonstrated that marizomib can cross the blood-brain barrier and inhibit proteasome activity in rodent and nonhuman primate brain and elicit a significant antitumor effect in a rodent intracranial model of malignant glioma.

Objectives: New data suggest that glioma stem-like cells (GSCs) and neural stem/progenitor cells (NSCs) may share common origins. GSCs drive tumor proliferation and appear to be resistant to classic chemotherapy, while the effects of chemotherapy on NSCs are not well studied. As the role of NSCs in learning and memory is increasingly recognized, we need to identify drugs that reduce neurotoxicity but are still effective against glial tumors. Methods: We treated 3 human NSC cultures and multiple low- and high-grade GSC cultures with the commonly used agents temozolomide (TMZ) and cisplatin (CIS), and with 2 newer, promising drugs: the proteasome inhibitor bortezomib (BTZ) and the epidermal growth factor receptor tyrosine kinase inhibitor erlotinib (ERL). We measured cell survival, proliferation, cell death induction, and drug resistance markers. Results: TMZ decreased NSC viability, while minimally affecting GSCs. TMZ induced NSC death, which was partially compensated for by increased proliferation. CIS had similar effects. The NSC9s sensitivity to TMZ and CIS correlated with low expression of the multidrug resistance gene ABCG2, but not of MGMT or MSH1/MLH2. BTZ caused an 80%decrease in GSCs, while minimally affecting NSCs. GSCs had lower proteasome levels and activity after BTZ treatment. ERL treatment also decreased GSC numbers, but not NSC viability, which correlated with low EGFR expression in NSCs compared to GSCs. Conclusions: Newer chemotherapy agents ERL and BTZ are effective against GSCs yet produce minimal effects on NSCs, while the older drugs TMZ and CIS are more toxic for NSCs than for GSCs. The identification and testing of more selective drugs is clearly warranted.