Objective. The purpose of this study was to investigate the feasibility of applying handcrafted radiomics (HCR) and deep learning-based radiomics (DLR) for the accurate preoperative classification of glioblastoma (GBM) and solitary brain metastasis (BM). Methods. A retrospective analysis of the magnetic resonance imaging (MRI) data of 140 patients (110 in the training dataset and 30 in the test dataset) with GBM and 128 patients (98 in the training dataset and 30 in the test dataset) with BM confirmed by surgical pathology was performed. The regions of interest (ROIs) on T1-weighted imaging (T1WI), T2-weighted imaging (T2WI), and contrast-enhanced T1WI (T1CE) were drawn manually, and then, HCR and DLR analyses were performed. On this basis, different machine learning algorithms were implemented and compared to find the optimal modeling method. The final classifiers were identified and validated for different MRI modalities using HCR features and HCR + DLR features. By analyzing the receiver operating characteristic (ROC) curve, the area under the curve (AUC), accuracy, sensitivity, and specificity were calculated to evaluate the predictive efficacy of different methods. Results. In multiclassifier modeling, random forest modeling showed the best distinguishing performance among all MRI modalities. HCR models already showed good results for distinguishing between the two types of brain tumors in the test dataset (T1WI, AUC = 0.86; T2WI, AUC = 0.76; T1CE, AUC = 0.93). By adding DLR features, all AUCs showed significant improvement (T1WI, AUC = 0.87; T2WI, AUC = 0.80; T1CE, AUC = 0.97; ). The T1CE-based radiomic model showed the best classification performance (AUC = 0.99 in the training dataset and AUC = 0.97 in the test dataset), surpassing the other MRI modalities ( ). The multimodality radiomic model also showed robust performance (AUC = 1 in the training dataset and AUC = 0.84 in the test dataset). Conclusion. Machine learning models using MRI radiomic features can help distinguish GBM from BM effectively, especially the combination of HCR and DLR features.
Alternative splicing (AS) of pre-mRNA has been widely reported to be associated with the progression of malignant tumors. However, a systematic investigation into the prognostic value of AS events in glioblastoma (GBM) is urgently required. The gene expression profile and matched AS events data of GBM patients were obtained from The Cancer Genome Atlas Project (TCGA) and TCGA SpliceSeq database, respectively. 775 AS events were identified as prognostic factors using univariate Cox regression analysis. The least absolute shrinkage and selection operator (LASSO) cox model was performed to narrow down candidate AS events, and a risk score model based on several AS events were developed subsequently. The risk score-based signature was proved as an efficient predictor of overall survival and was closely related to the tumor purity and immunosuppression in GBM. Combined similarity network fusion and consensus clustering (SNF-CC) analysis revealed two distinct GBM subtypes based on the prognostic AS events, and the associations between this novel molecular classification and clinicopathological factors, immune cell infiltration, as well as immunogenic features were further explored. We also constructed a regulatory network to depict the potential mechanisms that how prognostic splicing factors (SFs) regulate splicing patterns in GBM. Finally, a nomogram incorporating AS events signature and other clinical-relevant covariates was built for clinical application. This comprehensive analysis highlights the potential implications for predicting prognosis and clinical management in GBM.
Background. IDH mutation is the most common in diffuse LGGs, correlated with a favorable prognosis. However, the IDH-mutant LGGs patients with poor prognoses need to be identified, and the potential mechanism leading to a worse outcome and treatment options needs to be investigated. Methods. A six-gene immune-related prognostic signature in IDH-mutant LGGs was constructed based on two public datasets and univariate, multivariate, and LASSO Cox regression analysis. Patients were divided into low- and high-risk groups based on the median risk score in the training and validation sets. We analyzed enriched pathways and immune cell infiltration, applying the GSEA and the immune evaluation algorithms. Results. Stratification and multivariate Cox analysis unveiled that the six-gene signature was an independent prognostic factor. The signature (0.806/0.795/0.822) showed a remarkable prognostic performance, with 1-, 3-, and 5-year time-dependent AUC, higher than for grade (0.612/0.638/0.649) and 1p19q codeletion status (0.606/0.658/0.676). High-risk patients had higher infiltrating immune cells. However, the specific immune escape was observed in the high-risk group after immune activation, owing to increasing immunosuppressive cells, inhibitory cytokines, and immune checkpoint molecules. Moreover, a novel nomogram model was developed to evaluate the survival in IDH-mutant LGGs patients. Conclusion. The six-gene signature could be a promising prognostic biomarker, which is promising to promote individual therapy and improve the clinical outcomes of IDH-mutant gliomas. The study also refined the current classification system of IDH-mutant gliomas, classifying patients into two subtypes with distinct immunophenotypes and overall survival.
It is important to observe the metallic vapor dissipation after interrupting a vacuum arc, for getting a deeper understanding of dielectric recovery processes, especially for DC interruptions. The object of this paper is to obtain two dimensional density distribution of copper vapor in a forced vacuum arc extinction. The method of laser-induced fluorescence (LIF) using a pulse laser with a central wavelength at 324.8 nm was adopted here. Meanwhile, an ICCD was used to capture the 510.6 nm fluorescence signal with an exposure time of 50 ns. The fluorescence can be distinguished from 125μs (the moment of arc extinction) to 375μs with a time interval of 25μs. As a result, we could obtain the relative density distribution of ground-state copper metal vapor after the arc extinction. The results show that on the condition of CuCr25 contacts the metal vapor was evenly distributed in the gap and rapidly spread over time. The density of the copper vapor between the contacts dropped to less than 10% of the initial stage 200μs after arc extinction. On the contrary, on the condition of CuCr50, higher copper metal vapor density at the moment of arc extinction was observed. It seems that the higher Cr content in the Cu-Cr contacts will lead to producing more metal vapor around current zero.