Preoperative relative cerebral blood volume analysis in gliomas predicts survival and mitigates risk of biopsy sampling error.

Appropriate management of adult gliomas requires an accurate histopathological diagnosis. However, the heterogeneity of gliomas can lead to misdiagnosis and undergrading, especially with biopsy. We evaluated the role of preoperative relative cerebral blood volume (rCBV) analysis in conjunction with histopathological analysis as a predictor of overall survival and risk of undergrading. We retrospectively identified 146 patients with newly diagnosed gliomas (WHO grade II-IV) that had undergone preoperative MRI with rCBV analysis. We compared overall survival by histopathologically determined WHO tumor grade and by rCBV using Kaplan-Meier survival curves and the Cox proportional hazards model. We also compared preoperative imaging findings and initial histopathological diagnosis in 13 patients who underwent biopsy followed by subsequent resection. Survival curves by WHO grade and rCBV tier similarly separated patients into low, intermediate, and high-risk groups with shorter survival corresponding to higher grade or rCBV tier. The hazard ratio for WHO grade III versus II was 3.91 (p = 0.018) and for grade IV versus II was 11.26 (p < 0.0001) and the hazard ratio for each increase in 1.0 rCBV units was 1.12 (p < 0.002). Additionally, 3 of 13 (23%) patients initially diagnosed by biopsy were upgraded on subsequent resection. Preoperative rCBV was elevated at least one standard deviation above the mean in the 3 upgraded patients, suggestive of undergrading, but not in the ten concordant diagnoses. In conclusion, rCBV can predict overall survival similarly to pathologically determined WHO grade in patients with gliomas. Discordant rCBV analysis and histopathology may help identify patients at higher risk for undergrading.

Biological and therapeutic implications of multisector sequencing in newly diagnosed glioblastoma.

Background: Diagnostic workflows for glioblastoma (GBM) patients increasingly include DNA sequencing-based analysis of a single tumor site following biopsy or resection. We hypothesized that sequencing of multiple sectors within a given tumor would provide a more comprehensive representation of the molecular landscape and potentially inform therapeutic strategies. Methods: Ten newly diagnosed, isocitrate dehydrogenase 1 (IDH1) wildtype GBM tumor samples were obtained from 2 (n = 9) or 4 (n = 1) spatially distinct tumor regions. Tumor and matched blood DNA samples underwent whole-exome sequencing. Results: Across all 10 tumors, 51% of mutations were clonal and 3% were subclonal and shared in different sectors, whereas 46% of mutations were subclonal and private. Two of the 10 tumors exhibited a regional hypermutator state despite being treatment naïve, and remarkably, the high mutational load was predominantly limited to one sector in each tumor. Among the canonical cancer-associated genes, only telomerase reverse transcriptase (TERT) promoter mutations were observed in the founding clone in all tumors. Reconstruction of the clonal architecture in different sectors revealed regionally divergent evolution, and integration of data from 2 sectors increased the resolution of inferred clonal architecture in a given tumor. Predicted therapeutic mutations differed in presence and frequency between tumor regions. Similarly, different sectors exhibited significant divergence in the predicted neoantigen landscape. Conclusions: The substantial spatial heterogeneity observed in different GBM tumor sectors, especially in spatially restricted hypermutator cases, raises important caveats to our current dependence on single-sector molecular information to guide either targeted or immune-based treatments.

Dynamic susceptibility contrast and dynamic contrast-enhanced MRI characteristics to distinguish microcystic meningiomas from traditional Grade I meningiomas and high-grade gliomas.

OBJECTIVE Microcystic meningioma (MM) is a meningioma variant with a multicystic appearance that may mimic intrinsic primary brain tumors and other nonmeningiomatous tumor types. Dynamic susceptibility contrast (DSC) and dynamic contrast-enhanced (DCE) MRI techniques provide imaging parameters that can differentiate these tumors according to hemodynamic and permeability characteristics with the potential to aid in preoperative identification of tumor type. METHODS The medical data of 18 patients with a histopathological diagnosis of MM were identified through a retrospective review of procedures performed between 2008 and 2012; DSC imaging data were available for 12 patients and DCE imaging data for 6. A subcohort of 12 patients with Grade I meningiomas (i.e., of meningoepithelial subtype) and 54 patients with Grade IV primary gliomas (i.e., astrocytomas) was also included, and all preoperative imaging sequences were analyzed. Clinical variables including patient sex, age, and surgical blood loss were also included in the analysis. Images were acquired at both 1.5 and 3.0 T. The DSC images were acquired at a temporal resolution of either 1500 msec (3.0 T) or 2000 msec (1.5 T). In all cases, parameters including normalized cerebral blood volume (CBV) and transfer coefficient (kTrans) were calculated with region-of-interest analysis of enhancing tumor volume. The normalized CBV and kTrans data from the patient groups were analyzed with 1-way ANOVA, and post hoc statistical comparisons among groups were conducted with the Bonferroni adjustment. RESULTS Preoperative DSC imaging indicated mean (± SD) normalized CBVs of 5.7 ± 2.2 ml for WHO Grade I meningiomas of the meningoepithelial subtype (n = 12), 4.8 ± 1.8 ml for Grade IV astrocytomas (n = 54), and 12.3 ± 3.8 ml for Grade I meningiomas of the MM subtype (n = 12). The normalized CBV measured within the enhancing portion of the tumor was significantly higher in the MM subtype than in typical meningiomas and Grade IV astrocytomas (p < 0.001 for both). Preoperative DCE imaging indicated mean kTrans values of 0.49 ± 0.20 min-1 in Grade I meningiomas of the meningoepithelial subtype (n = 12), 0.27 ± 0.12 min-1 for Grade IV astrocytomas (n = 54), and 1.35 ± 0.74 min-1 for Grade I meningiomas of the MM subtype (n = 6). The kTrans was significantly higher in the MM variants than in the corresponding nonmicrocystic Grade 1 meningiomas and Grade IV astrocytomas (p < 0.001 for both). Intraoperative blood loss tended to increase with increased normalized CBV (R = 0.45, p = 0.085). CONCLUSIONS An enhancing cystic lesion with a normalized CBV greater than 10.3 ml or a kTrans greater than 0.88 min-1 should prompt radiologists and surgeons to consider the diagnosis of MM rather than traditional Grade I meningioma or high-grade glioma in planning surgical care. Higher normalized CBVs tend to be associated with increased intraoperative blood loss.

Attrition rates in neurosurgery residency: analysis of 1361 consecutive residents matched from 1990 to 1999.

OBJECT: The objective of this study is to determine neurosurgery residency attrition rates by sex of matched applicant and by type and rank of medical school attended. METHODS: The study follows a cohort of 1361 individuals who matched into a neurosurgery residency program through the SF Match Fellowship and Residency Matching Service from 1990 to 1999. The main outcome measure was achievement of board certification as documented in the American Board of Neurological Surgery Directory of Diplomats. A secondary outcome measure was documentation of practicing medicine as verified by the American Medical Association DoctorFinder and National Provider Identifier websites. Overall, 10.7% (n=146) of these individuals were women. Twenty percent (n=266) graduated from a top 10 medical school (24% of women [35/146] and 19% of men [232/1215], p=0.19). Forty-five percent (n=618) were graduates of a public medical school, 50% (n=680) of a private medical school, and 5% (n=63) of an international medical school. At the end of the study, 0.2% of subjects (n=3) were deceased and 0.3% (n=4) were lost to follow-up. RESULTS: The total residency completion rate was 86.0% (n=1171) overall, with 76.0% (n=111/146) of women and 87.2% (n=1059/1215) of men completing residency. Board certification was obtained by 79.4% (n=1081) of all individuals matching into residency between 1990 and 1999. Overall, 63.0% (92/146) of women and 81.3% (989/1215) of men were board certified. Women were found to be significantly more at risk (p<0.005) of not completing residency or becoming board certified than men. Public medical school alumni had significantly higher board certification rates than private and international alumni (82.2% for public [508/618]; 77.1% for private [524/680]; 77.8% for international [49/63]; p<0.05). There was no significant difference in attrition for graduates of top 10-ranked institutions versus other institutions. There was no difference in number of years to achieve neurosurgical board certification for men versus women. CONCLUSIONS: Overall, neurosurgery training attrition rates are low. Women have had greater attrition than men during and after neurosurgery residency training. International and private medical school alumni had higher attrition than public medical school alumni.

Predicting a multi-parametric probability map of active tumor extent using random forests.

Glioblastoma Mulitforme is highly infiltrative, making precise delineation of tumor margin difficult. Multimodality or multi-parametric MR imaging sequences promise an advantage over anatomic sequences such as post contrast enhancement as methods for determining the spatial extent of tumor involvement. In considering multi-parametric imaging sequences however, manual image segmentation and classification is time-consuming and prone to error. As a preliminary step toward integration of multi-parametric imaging into clinical assessments of primary brain tumors, we propose a machine-learning based multi-parametric approach that uses radiologist generated labels to train a classifier that is able to classify tissue on a voxel-wise basis and automatically generate a tumor segmentation. A random forests classifier was trained using a leave-one-out experimental paradigm. A simple linear classifier was also trained for comparison. The random forests classifier accurately predicted radiologist generated segmentations and tumor extent.