Background: Skull base chordoma is a rare tumor with low-grade malignancy and a high recurrence rate, the factors affecting the prognosis of patients need to be further studied. For that, we investigated prognostic factors of skull base chordoma through the database of the Surveillance, Epidemiology, and End Results (SEER) program, and validated in an independent data set from the Xiangya Hospital.Methods: Six hundred and forty-three patients diagnosed with skull base chordoma were obtained from the SEER database (606 patients) and the Xiangya Hospital (37 patients). Categorical variables were selected by Chi-square test with a statistical difference. Survival curves were constructed by Kaplan–Meier analysis and compared by log-rank test. Univariate and multivariate Cox regression analyses were used to explore the prognostic factors. Propensity score matching (PSM) analysis was undertaken to reduce the substantial bias between gross total resection (GTR) and subtotal resection (STR) groups. Furthermore, clinical data of 37 patients from the Xiangya Hospital were used as validation cohorts to check the survival impacts of the extent of resection and adjuvant radiotherapy on prognosis.Results: We found that age at diagnosis, primary site, disease stage, surgical treatment, and tumor size was significantly associated with the prognosis of skull base chordoma. PSM analysis revealed that there was no significant difference in the OS between GTR and STR (p = 0.157). Independent data set from the Xiangya Hospital proved no statistical difference in OS between GTR and STR groups (p = 0.16), but the GTR group was superior to the STR group for progression-free survival (PFS) (p = 0.048). Postoperative radiotherapy does not improve OS (p = 0.28), but it can prolong PFS (p = 0.0037). Nomograms predicting 5- and 10-year OS and DSS were constructed based on statistically significant factors identified by multivariate Cox analysis. Age, primary site, tumor size, surgical treatment, and disease stage were included as prognostic predictors in the nomograms with good performance.Conclusions: We identified age, tumor size, surgery, primary site, and tumor stage as main factors affecting the prognosis of the skull base chordoma. Resection of the tumor as much as possible while ensuring safety, combined with postoperative radiotherapy may be the optimum treatment for skull base chordoma.

Despite a generally better prognosis than high-grade glioma (HGG), recurrence and malignant progression are the main causes for the poor prognosis and difficulties in the treatment of low-grade glioma (LGG). It is of great importance to learn about the risk factors and underlying mechanisms of LGG recurrence and progression. In this study, the transcriptome characteristics of four groups, namely, normal brain tissue and recurrent LGG (rLGG), normal brain tissue and secondary glioblastoma (sGBM), primary LGG (pLGG) and rLGG, and pLGG and sGBM, were compared using Chinese Glioma Genome Atlas (CGGA) and Genotype-Tissue Expression Project (GTEx) databases. In this study, 296 downregulated and 396 upregulated differentially expressed genes (DEGs) with high consensus were screened out. Univariate Cox regression analysis of data from The Cancer Genome Atlas (TCGA) yielded 86 prognostically relevant DEGs; a prognostic prediction model based on five key genes (HOXA1, KIF18A, FAM133A, HGF, and MN1) was established using the least absolute shrinkage and selection operator (LASSO) regression dimensionality reduction and multivariate Cox regression analysis. LGG was divided into high- and low-risk groups using this prediction model. Gene Set Enrichment Analysis (GSEA) revealed that signaling pathway differences in the high- and low-risk groups were mainly seen in tumor immune regulation and DNA damage-related cell cycle checkpoints. Furthermore, the infiltration of immune cells in the high- and low-risk groups was analyzed, which indicated a stronger infiltration of immune cells in the high-risk group than that in the low-risk group, suggesting that an immune microenvironment more conducive to tumor growth emerged due to the interaction between tumor and immune cells. The tumor mutational burden and tumor methylation burden in the high- and low-risk groups were also analyzed, which indicated higher gene mutation burden and lower DNA methylation level in the high-risk group, suggesting that with the accumulation of genomic mutations and epigenetic changes, tumor cells continued to evolve and led to the progression of LGG to HGG. Finally, the value of potential therapeutic targets for the five key genes was analyzed, and findings demonstrated that KIF18A was the gene most likely to be a potential therapeutic target. In conclusion, the prediction model based on these five key genes can better identify the high- and low-risk groups of LGG and lay a solid foundation for evaluating the risk of LGG recurrence and malignant progression.

BackgroundGerm cell tumors (GCTs) are neoplasms derived from reproductive cells, mostly occurring in children and adolescents at 10 to 19 years of age. Intracranial GCTs are classified histologically into germinomas and non-germinomatous germ cell tumors. Germinomas of the basal ganglia are difficult to distinguish based on symptoms or routine MRI images from gliomas, even for experienced neurosurgeons or radiologists. Meanwhile, intracranial germinoma has a lower incidence rate than glioma in children and adults. Therefore, we established a model based on pre-trained ResNet18 with transfer learning to better identify germinomas of the basal ganglia.MethodsThis retrospective study enrolled 73 patients diagnosed with germinoma or glioma of the basal ganglia. Brain lesions were manually segmented based on both T1C and T2 FLAIR sequences. The T1C sequence was used to build the tumor classification model. A 2D convolutional architecture and transfer learning were implemented. ResNet18 from ImageNet was retrained on the MRI images of our cohort. Class activation mapping was applied for the model visualization.ResultsThe model was trained using five-fold cross-validation, achieving a mean AUC of 0.88. By analyzing the class activation map, we found that the model’s attention was focused on the peri-tumoral edema region of gliomas and tumor bulk for germinomas, indicating that differences in these regions may help discriminate these tumors.ConclusionsThis study showed that the T1C-based transfer learning model could accurately distinguish germinomas from gliomas of the basal ganglia preoperatively.