Interactive Effects of Molecular, Therapeutic, and Patient Factors on Outcome of Diffuse Low-Grade Glioma.

PURPOSE: In patients with diffuse low-grade glioma (LGG), the extent of surgical tumor resection (EOR) has a controversial role, in part because a randomized clinical trial with different levels of EOR is not feasible. METHODS: In a 20-year retrospective cohort of 392 patients with IDH-mutant grade 2 glioma, we analyzed the combined effects of volumetric EOR and molecular and clinical factors on overall survival (OS) and progression-free survival by recursive partitioning analysis. The OS results were validated in two external cohorts (n = 365). Propensity score analysis of the combined cohorts (n = 757) was used to mimic a randomized clinical trial with varying levels of EOR. RESULTS: Recursive partitioning analysis identified three survival risk groups. Median OS was shortest in two subsets of patients with astrocytoma: those with postoperative tumor volume (TV) > 4.6 mL and those with preoperative TV > 43.1 mL and postoperative TV ≤ 4.6 mL. Intermediate OS was seen in patients with astrocytoma who had chemotherapy with preoperative TV ≤ 43.1 mL and postoperative TV ≤ 4.6 mL in addition to oligodendroglioma patients with either preoperative TV > 43.1 mL and residual TV ≤ 4.6 mL or postoperative residual volume > 4.6 mL. Longest OS was seen in astrocytoma patients with preoperative TV ≤ 43.1 mL and postoperative TV ≤ 4.6 mL who received no chemotherapy and oligodendroglioma patients with preoperative TV ≤ 43.1 mL and postoperative TV ≤ 4.6 mL. EOR ≥ 75% improved survival outcomes, as shown by propensity score analysis. CONCLUSION: Across both subtypes of LGG, EOR beginning at 75% improves OS while beginning at 80% improves progression-free survival. Nonetheless, maximal resection with preservation of neurological function remains the treatment goal. Our findings have implications for surgical strategies for LGGs, particularly oligodendroglioma.

Deep neural networks allow expert-level brain meningioma segmentation and present potential for improvement of clinical practice.

Accurate brain meningioma segmentation and volumetric assessment are critical for serial patient follow-up, surgical planning and monitoring response to treatment. Current gold standard of manual labeling is a time-consuming process, subject to inter-user variability. Fully-automated algorithms for meningioma segmentation have the potential to bring volumetric analysis into clinical and research workflows by increasing accuracy and efficiency, reducing inter-user variability and saving time. Previous research has focused solely on segmentation tasks without assessment of impact and usability of deep learning solutions in clinical practice. Herein, we demonstrate a three-dimensional convolutional neural network (3D-CNN) that performs expert-level, automated meningioma segmentation and volume estimation on MRI scans. A 3D-CNN was initially trained by segmenting entire brain volumes using a dataset of 10,099 healthy brain MRIs. Using transfer learning, the network was then specifically trained on meningioma segmentation using 806 expert-labeled MRIs. The final model achieved a median performance of 88.2% reaching the spectrum of current inter-expert variability (82.6-91.6%). We demonstrate in a simulated clinical scenario that a deep learning approach to meningioma segmentation is feasible, highly accurate and has the potential to improve current clinical practice.

Preoperative Brain Tumor Imaging: Models and Software for Segmentation and Standardized Reporting.

For patients suffering from brain tumor, prognosis estimation and treatment decisions are made by a multidisciplinary team based on a set of preoperative MR scans. Currently, the lack of standardized and automatic methods for tumor detection and generation of clinical reports, incorporating a wide range of tumor characteristics, represents a major hurdle. In this study, we investigate the most occurring brain tumor types: glioblastomas, lower grade gliomas, meningiomas, and metastases, through four cohorts of up to 4,000 patients. Tumor segmentation models were trained using the AGU-Net architecture with different preprocessing steps and protocols. Segmentation performances were assessed in-depth using a wide-range of voxel and patient-wise metrics covering volume, distance, and probabilistic aspects. Finally, two software solutions have been developed, enabling an easy use of the trained models and standardized generation of clinical reports: Raidionics and Raidionics-Slicer. Segmentation performances were quite homogeneous across the four different brain tumor types, with an average true positive Dice ranging between 80 and 90%, patient-wise recall between 88 and 98%, and patient-wise precision around 95%. In conjunction to Dice, the identified most relevant other metrics were the relative absolute volume difference, the variation of information, and the Hausdorff, Mahalanobis, and object average symmetric surface distances. With our Raidionics software, running on a desktop computer with CPU support, tumor segmentation can be performed in 16-54 s depending on the dimensions of the MRI volume. For the generation of a standardized clinical report, including the tumor segmentation and features computation, 5-15 min are necessary. All trained models have been made open-access together with the source code for both software solutions and validation metrics computation. In the future, a method to convert results from a set of metrics into a final single score would be highly desirable for easier ranking across trained models. In addition, an automatic classification of the brain tumor type would be necessary to replace manual user input. Finally, the inclusion of post-operative segmentation in both software solutions will be key for generating complete post-operative standardized clinical reports.

Short-term outcomes associated with temozolomide or PCV chemotherapy for 1p/19q-codeleted WHO grade 3 oligodendrogliomas: A national evaluation.

Background: The optimal chemotherapy regimen between temozolomide and procarbazine, lomustine, and vincristine (PCV) remains uncertain for WHO grade 3 oligodendroglioma (Olig3) patients. We therefore investigated this question using national data. Methods: Patients diagnosed with radiotherapy-treated 1p/19q-codeleted Olig3 between 2010 and 2018 were identified from the National Cancer Database. The overall survival (OS) associated with first-line single-agent temozolomide vs multi-agent PCV was estimated by Kaplan-Meier techniques and evaluated by multivariable Cox regression. Results: One thousand five hundred ninety-six radiotherapy-treated 1p/19q-codeleted Olig3 patients were identified: 88.6% (n = 1414) treated with temozolomide and 11.4% (n = 182) with PCV (from 5.4% in 2010 to 12.0% in 2018) in the first-line setting. The median follow-up was 35.5 months (interquartile range [IQR] 20.7-60.6 months) with 63.3% of patients alive at the time of analysis. There was a significant difference in unadjusted OS between temozolomide (5-year OS 58.9%, 95%CI: 55.6-62.0) and PCV (5-year OS 65.1%, 95%CI: 54.8-73.5; P = .04). However, a significant OS difference between temozolomide and PCV was not observed in the Cox regression analysis adjusted by age and extent of resection (PCV vs temozolomide HR 0.81, 95%CI: 0.59-1.11, P = .18). PCV was more frequently used for younger Olig3s but otherwise was not associated with patient's insurance status or care setting. Conclusions: In a national analysis of Olig3s, first-line PCV chemotherapy was associated with a slightly improved unadjusted short-term OS compared to temozolomide; but not following adjustment by patient age and extent of resection. There has been an increase in PCV utilization since 2010. These findings provide preliminary data while we await the definitive results from the CODEL trial.

Survival Prediction After Neurosurgical Resection of Brain Metastases: A Machine Learning Approach.

BACKGROUND: Current prognostic models for brain metastases (BMs) have been constructed and validated almost entirely with data from patients receiving up-front radiotherapy, leaving uncertainty about surgical patients. OBJECTIVE: To build and validate a model predicting 6-month survival after BM resection using different machine learning algorithms. METHODS: An institutional database of 1062 patients who underwent resection for BM was split into an 80:20 training and testing set. Seven different machine learning algorithms were trained and assessed for performance; an established prognostic model for patients with BM undergoing radiotherapy, the diagnosis-specific graded prognostic assessment, was also evaluated. Model performance was assessed using area under the curve (AUC) and calibration. RESULTS: The logistic regression showed the best performance with an AUC of 0.71 in the hold-out test set, a calibration slope of 0.76, and a calibration intercept of 0.03. The diagnosis-specific graded prognostic assessment had an AUC of 0.66. Patients were stratified into regular-risk, high-risk and very high-risk groups for death at 6 months; these strata strongly predicted both 6-month and longitudinal overall survival ( P < .0005). The model was implemented into a web application that can be accessed through http://brainmets.morethanml.com . CONCLUSION: We developed and internally validated a prediction model that accurately predicts 6-month survival after neurosurgical resection for BM and allows for meaningful risk stratification. Future efforts should focus on external validation of our model.

Survival outcomes associated with MGMT promoter methylation and temozolomide in gliosarcoma patients.

PURPOSE: Gliosarcoma is an uncommon glioblastoma subtype, for which MGMT promoter methylation's relationship with response to temozolomide chemotherapy is unclear. We therefore examined this question using a national cohort. METHODS:  The National Cancer Database was queried for patients histopathologically diagnosed with gliosarcoma between 2010 and 2019. The associations between MGMT promoter methylation, first-line single-agent chemotherapy-presumed to be temozolomide herein-and overall survival (OS) were examined using log-rank tests and Cox regression, with correction for multiple testing (p < 0.01 was significant). RESULTS: 580 newly-diagnosed gliosarcoma patients with MGMT status were available, among whom 33.6% were MGMT promoter methylated. Median OS for gliosarcoma patients that received standard-of-care temozolomide and radiotherapy was 12.1 months (99% confidence interval [CI] 10.8-15.1) for MGMT promoter unmethylated and 21.4 months (99% CI 15.4-26.2) for MGMT promoter methylated gliosarcomas (p = 0.003). In multivariable analysis of gliosarcoma patients-which included the potential confounders of age, sex, maximal tumor size, extent of resection, and radiotherapy-receipt of temozolomide was associated with improved OS in both MGMT promoter methylated (hazard ratio [HR] 0.23 vs. no temozolomide, 99% CI 0.11-0.47, p < 0.001) and unmethylated (HR 0.50 vs. no temozolomide, 99% CI 0.29-0.89, p = 0.002) gliosarcomas. MGMT promoter methylation was associated with improved OS among temozolomide-treated gliosarcoma patients (p < 0.001), but not in patients who did not receive chemotherapy (p = 0.35). CONCLUSION: In a national analysis of gliosarcoma patients, temozolomide was associated with prolonged OS irrespective of MGMT status. These results provide support for the current practice of trimodal therapy for gliosarcoma.

Surgery vs. Biopsy in the Treatment of Butterfly Glioblastoma: A Systematic Review and Meta-Analysis.

Butterfly glioblastomas (bGBM) are grade IV gliomas that spread to bilateral hemispheres by infiltrating the corpus callosum. Data on the effect of surgery are limited to small case series. The aim of this meta-analysis was to compare resection vs. biopsy in terms of survival outcomes and postoperative complications. A systematic review of the literature was conducted using PubMed, EMBASE, and Cochrane databases through March 2021 in accordance with the PRISMA checklist. Pooled hazard ratios were calculated and meta-analyzed in a random-effects model including assessment of heterogeneity. Out of 3367 articles, seven studies were included with 293 patients. Surgical resection was significantly associated with longer overall survival (HR 0.39, 95%CI 0.2-0.55) than biopsy. Low heterogeneity was observed (I2: 0%). In further analysis, the effect persisted in extent of resection subgroups of both ≥80% and <80%. No statistically significant difference between surgery and biopsy was detected in terms of postoperative complications, although these were numerically larger for surgery. In patients with bGBM, surgical resection was associated with longer survival prospects compared with biopsy.

Neurosurgical resection for locally recurrent brain metastasis.

BACKGROUND: In patients with locally recurrent brain metastases (LRBMs), the role of (repeat) craniotomy is controversial. This study aimed to analyze long-term oncological outcomes in this heterogeneous population. METHODS: Craniotomies for LRBM were identified from a tertiary neuro-oncological institution. First, we assessed overall survival (OS) and intracranial control (ICC) stratified by molecular profile, prognostic indices, and multimodality treatment. Second, we compared LRBMs to propensity score-matched patients who underwent craniotomy for newly diagnosed brain metastases (NDBM). RESULTS: Across 180 patients, median survival after LRBM resection was 13.8 months and varied by molecular profile, with >24 months survival in ALK/EGFR+ lung adenocarcinoma and HER2+ breast cancer. Furthermore, 102 patients (56.7%) experienced intracranial recurrence; median time to recurrence was 5.6 months. Compared to NDBMs (n = 898), LRBM patients were younger, more likely to harbor a targetable mutation and less likely to receive adjuvant radiation (P < 0.05). After 1:3 propensity matching stratified by molecular profile, LRBM patients generally experienced shorter OS (hazard ratio 1.67 and 1.36 for patients with or without a mutation, P < 0.05) but similar ICC (hazard ratio 1.11 in both groups, P > 0.20) compared to NDBM patients with similar baseline. Results across specific molecular subgroups suggested comparable effect directions of varying sizes. CONCLUSIONS: In our data, patients with LRBMs undergoing craniotomy comprised a subgroup of brain metastasis patients with relatively favorable clinical characteristics and good survival outcomes. Recurrent status predicted shorter OS but did not impact ICC. Craniotomy could be considered in selected, prognostically favorable patients.

Improved outcomes associated with maximal extent of resection for butterfly glioblastoma: insights from institutional and national data.

BACKGROUND: Butterfly glioblastomas (bGBMs) are grade IV gliomas that infiltrate the corpus callosum and spread to bilateral cerebral hemispheres. Due to the rarity of cases, there is a dearth of information in existing literature. Herein, we evaluate clinical and genetic characteristics, associated predictors, and survival outcomes in an institutional series and compare them to a national cohort. METHODS: We identified all adult patients with bGBM treated at Brigham & Women's Hospital (2008-2018). The National Cancer Database (NCDB) was also queried for bGBM patients. Survival was analyzed with Kaplan-Meier methods, and Cox models were built to assess for predictive factors. RESULTS: Of 993 glioblastoma patients, 62 cases (6.2%) of bGBM were identified. Craniotomy for resection was attempted in 26 patients (41.9%), with a median volumetric extent of resection (vEOR) of 72.3% (95% confidence interval [95%CI] 58.3-82.1). The IDH1 R132H mutation was detected in two patients (3.2%), and MGMT promoter was methylated in 55.5% of the assessed cases. In multivariable regression, factors predictive of longer OS were increased vEOR, MGMT promoter methylation, and receipt of adjuvant therapy. Median OS for the resected cases was 11.5 months (95%CI 7.7-18.8) vs. 6.3 (95%CI 5.1-8.9) for the biopsied. Of 21,353 GBMs, 719 (3.37%) bGBM patients were identified in the NCDB. Resection was more likely to be pursued in recent years, and GTR was independently associated with prolonged OS (p < 0.01). CONCLUSION: Surgical resection followed by adjuvant chemoradiation is associated with significant survival gains and should be pursued in carefully selected bGBM patients.

Patterns of Interaction Between Diffuse Low-Grade Glioma and Pregnancy: An Institutional Case Series.

BACKGROUND: The occurrence of pregnancy in patients with low-grade glioma (LGG) constitutes a unique therapeutic challenge. Owing to the rarity of cases, there is a dearth of information in existing literature. METHODS: We retrospectively identified all patients with a diagnosis of LGG and pregnancy at some point during their illness. Clinical course and obstetrical outcomes were reviewed. A volumetric analysis of tumor growth rate in association with pregnancy was performed. RESULTS: Of 15 women identified, 13 (86.7%) had a prepregnancy LGG diagnosis. Of the 2 patients in whom LGG was diagnosed during pregnancy, one underwent upfront surgery, and the other had watchful waiting with resection after 60 weeks. Nine patients (60.0%) remained asymptomatic during pregnancy, while 5 (33.3%) experienced recurrence of seizures. There was one case of transformation of an astrocytoma to glioblastoma during the third trimester, which was resected emergently. In 10 cases, progression occurred after pregnancy at a median interval of 24.2 months (interquartile range 6.6-37.5 months), with progression within 6 months of delivery in 2 cases. Mean (SD) growth rate during pregnancy was 7.8 (22.2) mm/year compared with 0.62 (1.12) mm/year before pregnancy and 0.29 (1.18) mm/year after pregnancy; the difference did not reach statistical significance (P = 0.306). CONCLUSIONS: Pregnancy was associated with clinical deterioration in one third of patients. No significant change in growth rate was identified. Time to progression and malignant dedifferentiation were unaffected. Patients with LGG wishing to pursue pregnancy should be counseled regarding the risk of complications, and if pregnancy is pursued, close neurological and obstetrical follow-up is recommended.

Programmed Death Receptor Ligand One Expression May Independently Predict Survival in Patients With Non-Small Cell Lung Carcinoma Brain Metastases Receiving Immunotherapy.

PURPOSE: Programmed death receptor ligand 1 (PD-L1) expression is known to predict response to PD-1/PD-L1 inhibitors in non-small cell lung cancer (NSCLC). However, the predictive role of this biomarker in brain metastases (BMs) is unknown. The aim of this study was to assess whether PD-L1 expression predicts survival in patients with NSCLC BMs treated with PD-1/PD-L1 inhibitors, after adjusting for established prognostic models. METHODS AND MATERIALS: In this multi-institutional retrospective cohort study, we identified patients with NSCLC-BM treated with PD-1/PD-L1 inhibitors after local BM treatment (radiation therapy or neurosurgery) but before intracranial progression. Cox proportional hazards models were used to assess the predictive value of PD-L1 expression for overall survival (OS) and intracranial progression-free survival (IC-PFS). RESULTS: Forty-eight patients with BM with available PD-L1 expression were identified. PD-L1 expression was positive in 33 patients (69%). Median survival was 26 months. In univariable analysis, PD-L1 predicted favorable OS (hazard ratio [HR], 0.44; 95% confidence interval [CI], 0.19-1.02; P = .055). This effect persisted after correcting for lung-graded prognostic assessment and other identified potential confounders (HR, 0.24; 95% CI, 0.10-0.61; P = .002). Moreover, when modeled as a continuous variable, there appeared to be a proportional relationship between percentage of PD-L1 expression and survival (HR, 0.86 per 10% expression; 95% CI, 0.77-0.98; P = .02). In contrast, PD-L1 expression did not predict IC-PFS in uni- or multivariable analysis (adjusted HR, 0.54; 95% CI, 0.26-1.14; P = .11). CONCLUSIONS: In patients with NSCLC-BMs treated with PD-1/PD-L1 checkpoint inhibitors and local treatment, PD-L1 expression may predict OS independent of lung-graded prognostic assessment. IC-PFS did not show association with PD-L1 expression, although the present analysis may lack power to assess this. Larger studies are required to validate these findings.

Subtype switching in breast cancer brain metastases: a multicenter analysis.

BACKGROUND: Breast cancer (BC) brain metastases (BM) can have discordant hormonal or human epidermal growth factor receptor 2 (HER2) expression compared with corresponding primary tumors. This study aimed to describe incidence, predictors, and survival outcomes of discordant receptors and associated subtype switching in BM. METHODS: BCBM patients seen at 4 tertiary institutions who had undergone BM resection or biopsy were included. Surgical pathology reports were retrospectively assessed to determine discordance between the primary tumor and the BCBM. In discordant cases, expression in extracranial metastases was also assessed. RESULTS: In BM from 219 patients, prevalence of any discordance was 36.3%; receptor-specific discordance was 16.7% for estrogen, 25.2% for progesterone, and 10.4% for HER2. Because estrogen and progesterone were considered together for hormonal status, 50 (22.8%) patients switched subtype as a result; 20 of these switches were HER2 based. Baseline subtype predicted switching, which occurred in up to 37.5% of primary HR+ patients. Moreover, 14.8% of initially HER2-negative patients gained HER2 in the BM. Most (63.6%) discordant patients with extracranial metastases also had discordance between BM and extracranial subtype. Loss of receptor expression was generally associated with worse survival, which appeared to be driven by estrogen loss (hazard ratio = 1.80, P = 0.03). Patients gaining HER2 status (n = 8) showed a nonsignificant tendency toward improved survival (hazard ratio = 0.64, P = 0.17). CONCLUSIONS: In this multicenter study, we report incidence and predictors of subtype switching, the risk of which varies considerably by baseline subtype. Switches can have clinical implications for prognosis and treatment choice.

Predictors and early survival outcomes of maximal resection in WHO grade II 1p/19q-codeleted oligodendrogliomas.

BACKGROUND: Although surgery plays a crucial diagnostic role in World Health Organization (WHO) grade II 1p/19q-codeleted oligodendrogliomas, the role of maximal tumor surgical resection remains unclear, with early retrospective series limited by lack of molecular classification or appropriate control groups. METHODS: The characteristics, management, and overall survival (OS) of patients ≥20 years old presenting with histology-proven WHO grade II 1p/19q-codeleted oligodendrogliomas during 2010-2016 were evaluated using the National Cancer Database and validated using multi-institutional data. Patients were stratified by watchful waiting (biopsy only) versus surgical resection. OS was analyzed using Kaplan-Meier methods and risk-adjusted proportional hazards. RESULTS: Five hundred ninety adults met inclusion criteria, of whom 79.0% (n = 466) underwent surgical resection. Of patient and tumor characteristics, younger patients were more likely to be resected. Achieving gross total resection (GTR; n = 320) was significantly associated with smaller tumors, management at integrated network cancer programs (vs community cancer programs), and Medicare insurance (as compared with no, private, or Medicaid/other government insurance) and independent of other patient or tumor characteristics. In risk-adjusted analyses, GTR, but not subtotal resection (STR), demonstrated improved OS (vs biopsy only: hazard ratio 0.28, 95% CI: 0.09-0.85, P = 0.02). CONCLUSIONS: WHO grade II 1p/19q-codeleted oligodendrogliomas amenable to resection demonstrated improved OS with GTR, but not STR, compared with biopsy-only watchful waiting. The OS benefits of GTR were independent of age, tumor size, or tumor location. Medicare-insured and integrated network cancer program patients were significantly more likely to have GTR than other patients, suggesting that insurance status and care setting may play important roles in access to timely diagnosis or innovations that improve maximal resection.

Fibrinolytics and Intraventricular Hemorrhage: A Systematic Review and Meta-analysis.

Intraventricular hemorrhage (IVH) is an independent poor prognostic factor in subarachnoid and intra-parenchymal hemorrhage. The use of intraventricular fibrinolytics (IVF) has long been debated, and its exact effects on outcomes are unknown. A systematic review and meta-analysis were performed in accordance with the PRISMA guidelines to assess the impact of IVF after non-traumatic IVH on mortality, functional outcome, intracranial bleeding, ventriculitis, time until clearance of third and fourth ventricles, obstruction of external ventricular drains (EVD), and shunt dependency. Nineteen studies were included in the meta-analysis, totaling 1020 patients. IVF was associated with lower mortality (relative risk [RR] 0.58; 95% confidence interval [CI] 0.47-0.72), fewer EVD obstructions (RR 0.41; 95% CI 0.22-0.74), and a shorter time until clearance of the ventricles (median difference [MD] - 4.05 days; 95% CI - 5.52 to - 2.57). There was no difference in good functional outcome, RR 1.41 (95% CI 0.98-2.03), or shunt dependency, RR 0.93 (95% CI 0.70-1.22). Correction for publication bias predicted an increased risk of intracranial bleeding, RR 1.67 (95% CI 1.01-2.74) and a lower risk of ventriculitis, RR 0.68 (95% CI 0.45-1.03) in IVH patients treated with IVF. IVF was associated with improved survival, faster clearance of blood from the ventricles and fewer drain obstructions, but further research is warranted to elucidate the effects on ventriculitis, long-term functional outcomes, and re-hemorrhage.

Contemporary assessment of extent of resection in molecularly defined categories of diffuse low-grade glioma: a volumetric analysis.

OBJECTIVE: While the effect of increased extent of resection (EOR) on survival in diffuse infiltrating low-grade glioma (LGG) patients is well established, there is still uncertainty about the influence of the new WHO molecular subtypes. The authors designed a retrospective analysis to assess the interplay between EOR and molecular classes. METHODS: The authors retrospectively reviewed the records of 326 patients treated surgically for hemispheric WHO grade II LGG at Brigham and Women's Hospital and Massachusetts General Hospital (2000-2017). EOR was calculated volumetrically and Cox proportional hazards models were built to assess for predictive factors of overall survival (OS), progression-free survival (PFS), and malignant progression-free survival (MPFS). RESULTS: There were 43 deaths (13.2%; median follow-up 5.4 years) among 326 LGG patients. Median preoperative tumor volume was 31.2 cm3 (IQR 12.9-66.0), and median postoperative residual tumor volume was 5.8 cm3 (IQR 1.1-20.5). On multivariable Cox regression, increasing postoperative volume was associated with worse OS (HR 1.02 per cm3; 95% CI 1.00-1.03; p = 0.016), PFS (HR 1.01 per cm3; 95% CI 1.00-1.02; p = 0.001), and MPFS (HR 1.01 per cm3; 95% CI 1.00-1.02; p = 0.035). This result was more pronounced in the worse prognosis subtypes of IDH-mutant and IDH-wildtype astrocytoma, for which differences in survival manifested in cases with residual tumor volume of only 1 cm3. In oligodendroglioma patients, postoperative residuals impacted survival when exceeding 8 cm3. Other significant predictors of OS were age at diagnosis, IDH-mutant and IDH-wildtype astrocytoma classes, adjuvant radiotherapy, and increasing preoperative volume. CONCLUSIONS: The results corroborate the role of EOR in survival and malignant transformation across all molecular subtypes of diffuse LGG. IDH-mutant and IDH-wildtype astrocytomas are affected even by minimal postoperative residuals and patients could potentially benefit from a more aggressive surgical approach.

Machine Learning Models can Detect Aneurysm Rupture and Identify Clinical Features Associated with Rupture.

BACKGROUND: Machine learning (ML) has been increasingly used in medicine and neurosurgery. We sought to determine whether ML models can distinguish ruptured from unruptured aneurysms and identify features associated with rupture. METHODS: We performed a retrospective review of patients with intracranial aneurysms detected on vascular imaging at our institution between 2002 and 2018. The dataset was used to train 3 ML models (random forest, linear support vector machine [SVM], and radial basis function kernel SVM). Relative contributions of individual predictors were derived from the linear SVM model. RESULTS: Complete data were available for 845 aneurysms in 615 patients. Ruptured aneurysms (n = 309, 37%) were larger (mean 6.51 mm vs. 5.73 mm; P = 0.02) and more likely to be in the posterior circulation (20% vs. 11%; P < 0.001) than unruptured aneurysms. Area under the receiver operating curve was 0.77 for the linear SVM, 0.78 for the radial basis function kernel SVM models, and 0.81 for the random forest model. Aneurysm location and size were the 2 features that contributed most significantly to the model. Posterior communicating artery, anterior communicating artery, and posterior inferior cerebellar artery locations were most highly associated with rupture, whereas paraclinoid and middle cerebral artery locations had the strongest association with unruptured status. CONCLUSIONS: ML models are capable of accurately distinguishing ruptured from unruptured aneurysms and identifying features associated with rupture. Consistent with prior studies, location and size show the strongest association with aneurysm rupture.

Automatic assessment of glioma burden: a deep learning algorithm for fully automated volumetric and bidimensional measurement.

BACKGROUND: Longitudinal measurement of glioma burden with MRI is the basis for treatment response assessment. In this study, we developed a deep learning algorithm that automatically segments abnormal fluid attenuated inversion recovery (FLAIR) hyperintensity and contrast-enhancing tumor, quantitating tumor volumes as well as the product of maximum bidimensional diameters according to the Response Assessment in Neuro-Oncology (RANO) criteria (AutoRANO). METHODS: Two cohorts of patients were used for this study. One consisted of 843 preoperative MRIs from 843 patients with low- or high-grade gliomas from 4 institutions and the second consisted of 713 longitudinal postoperative MRI visits from 54 patients with newly diagnosed glioblastomas (each with 2 pretreatment "baseline" MRIs) from 1 institution. RESULTS: The automatically generated FLAIR hyperintensity volume, contrast-enhancing tumor volume, and AutoRANO were highly repeatable for the double-baseline visits, with an intraclass correlation coefficient (ICC) of 0.986, 0.991, and 0.977, respectively, on the cohort of postoperative GBM patients. Furthermore, there was high agreement between manually and automatically measured tumor volumes, with ICC values of 0.915, 0.924, and 0.965 for preoperative FLAIR hyperintensity, postoperative FLAIR hyperintensity, and postoperative contrast-enhancing tumor volumes, respectively. Lastly, the ICCs for comparing manually and automatically derived longitudinal changes in tumor burden were 0.917, 0.966, and 0.850 for FLAIR hyperintensity volume, contrast-enhancing tumor volume, and RANO measures, respectively. CONCLUSIONS: Our automated algorithm demonstrates potential utility for evaluating tumor burden in complex posttreatment settings, although further validation in multicenter clinical trials will be needed prior to widespread implementation.

Survival and prognostic factors in surgically treated brain metastases.

PURPOSE: While surgery and radiation remain the mainstays of therapy for all patients with brain metastases (BM), the management is moving to a more individualized approach based on the underlying tumor. We sought to identify prognostic factors of both intracranial progression (IC-PFS) and overall survival (OS) in a surgical cohort. METHODS: We retrospectively reviewed the records of 1015 patients treated surgically for BM at Brigham and Women's Hospital (2007-2017). Kaplan-Meier curves were used for OS and IC-PFS and Cox proportional hazards models were built to assess for predictive factors. RESULTS: Common origins were lung (43.9%), breast (14.4%), and melanoma (13.8%). Median OS for the cohort was 15.4 months (95% confidence interval [95%CI] 14.1-17.1). Breast cancer (22.1 months, 95%CI 17.8-30.3) and colorectal cancer (10.6 months, 95%CI 7.2-15.4) had the longest and shortest OS, respectively. On multivariable Cox regression, significant prognostic factors of shorter OS were age (HR 1.01, 95%CI 1.01-1.02), number of lesions (HR 1.56, 95%CI 1.28-1.89), extracranial spread at BM diagnosis (HR 1.26, 95%CI 1.05-1.52), and KPS (HR 0.98, 95%CI 0.98-0.99). Regarding molecular factors, all driver mutations in lung adenocarcinoma had a favorable effect (EGFR, HR 0.53, 95%CI 0.31-0.89; ALK, HR 0.28, 95%CI 0.12-0.66; KRAS, HR 0.65, 95%CI 0.47-0.92), triple negative status predicted poor prognosis in breast adenocarcinoma (HR 2.04, 95%CI 1.13-3.69), while no effect of BRAF/NRAS mutations was demonstrated in melanoma BMs. CONCLUSIONS: Our results corroborate the role of tumor origin and systemic as well as intracranial spread in OS. Heterogeneity within histologies was further explained by molecular alterations.

Prognostic Value of Brain Metastasis-Free Interval in Patients with Breast Cancer Brain Metastases.

BACKGROUND: In patients with breast cancer brain metastases (BCBM), time between primary tumor diagnosis and appearance of brain metastases varies widely. Despite being a readily available clinical parameter, it remains unclear whether brain metastasis-free interval (BMFI) carries prognostic value among breast cancer patients. The aim of this study was to compare characteristics and overall survival among patients with varying BMFIs and to assess the prognostic role, if any, for BMFI. METHODS: We retrospectively reviewed 3 institutional databases of adult female patients who were treated for BCBM between 1996 and 2017. Cox proportional hazards model and Kaplan-Meier survival curves were used to determine prognostic value of BMFI for survival. RESULTS: A total of 503 patients were included. Median age at first brain metastasis was 52 (interquartile range [IQR]: 45-58) years. Median BMFI was 38 months (IQR: 18-66), and median overall survival was 17 months (IQR: 8-31). In univariate Cox proportional hazards model, younger age at BCBM, estrogen receptor (ER)+/human epidermal growth factor receptor 2 (HER2)+ tumor subtype, and the absence of liver or lung metastases were associated with longer survival. BMFI >3 years was not associated with longer survival (hazard ratio [HR] = 1.13; P = 0.21). In multivariate analysis, only subtype (ER+/HER2+ vs. ER-/HER2-; HR = 0.77; P = 0.02) and liver metastases (HR = 1.36; P = 0.01) were prognostic for survival. There was no significant association between BMFI and overall survival (HR = 0.99; P = 0.91). CONCLUSIONS: In this large, retrospective cohort of breast cancer patients, BMFI was not prognostic for overall survival.

Machine learning reveals multimodal MRI patterns predictive of isocitrate dehydrogenase and 1p/19q status in diffuse low- and high-grade gliomas.

PURPOSE: Isocitrate dehydrogenase (IDH) and 1p19q codeletion status are importantin providing prognostic information as well as prediction of treatment response in gliomas. Accurate determination of the IDH mutation status and 1p19q co-deletion prior to surgery may complement invasive tissue sampling and guide treatment decisions. METHODS: Preoperative MRIs of 538 glioma patients from three institutions were used as a training cohort. Histogram, shape, and texture features were extracted from preoperative MRIs of T1 contrast enhanced and T2-FLAIR sequences. The extracted features were then integrated with age using a random forest algorithm to generate a model predictive of IDH mutation status and 1p19q codeletion. The model was then validated using MRIs from glioma patients in the Cancer Imaging Archive. RESULTS: Our model predictive of IDH achieved an area under the receiver operating characteristic curve (AUC) of 0.921 in the training cohort and 0.919 in the validation cohort. Age offered the highest predictive value, followed by shape features. Based on the top 15 features, the AUC was 0.917 and 0.916 for the training and validation cohort, respectively. The overall accuracy for 3 group prediction (IDH-wild type, IDH-mutant and 1p19q co-deletion, IDH-mutant and 1p19q non-codeletion) was 78.2% (155 correctly predicted out of 198). CONCLUSION: Using machine-learning algorithms, high accuracy was achieved in the prediction of IDH genotype in gliomas and moderate accuracy in a three-group prediction including IDH genotype and 1p19q codeletion.