Cerebral perivascular spaces as predictors of dementia risk and accelerated brain atrophy.

Cerebral small vessel disease, an important risk factor for dementia, lacks robust, in vivo measurement methods. Perivascular spaces (PVS) on brain MRI are surrogates for small parenchymal blood vessels and their perivascular compartment, and may relate to brain health. We developed a novel, robust algorithm to automatically assess PVS count and size on MRI, and investigated their relationship with dementia risk and brain atrophy. We analyzed 46,478 clinical measurements of cognitive functioning and 20,845 brain MRI scans from 10,004 participants (71.1±9.7 years-old, 56.6% women). Fewer PVS and larger PVS diameter at baseline were associated with higher dementia risk and accelerated brain atrophy. Longitudinal trajectories of PVS markers were significantly different in non-demented individuals who converted to dementia compared with non-converters. In simulated placebo-controlled trials for treatments targeting cognitive decline, screening out participants less likely to develop dementia based on our PVS markers enhanced the power of the trial. These novel radiographic cerebrovascular markers may improve risk-stratification of individuals, potentially reducing cost and increasing throughput of clinical trials to combat dementia.

Quantification of cerebrospinal fluid tumor DNA in lung cancer patients with suspected leptomeningeal carcinomatosis.

Cerebrospinal fluid tumor-derived DNA (CSF-tDNA) analysis is a promising approach for monitoring the neoplastic processes of the central nervous system. We applied a lung cancer-specific sequencing panel (CAPP-Seq) to 81 CSF, blood, and tissue samples from 24 lung cancer patients who underwent lumbar puncture (LP) for suspected leptomeningeal disease (LMD). A subset of the cohort (N = 12) participated in a prospective trial of osimertinib for refractory LMD in which serial LPs were performed before and during treatment. CSF-tDNA variant allele fractions (VAFs) were significantly higher than plasma circulating tumor DNA (ctDNA) VAFs (median CSF-tDNA, 32.7%; median plasma ctDNA, 1.8%; P < 0.0001). Concentrations of tumor DNA in CSF and plasma were positively correlated (Spearman's ρ, 0.45; P = 0.03). For LMD diagnosis, cytology was 81.8% sensitive and CSF-tDNA was 91.7% sensitive. CSF-tDNA was also strongly prognostic for overall survival (HR = 7.1; P = 0.02). Among patients with progression on targeted therapy, resistance mutations, such as EGFR T790M and MET amplification, were common in peripheral blood but were rare in time-matched CSF, indicating differences in resistance mechanisms based on the anatomic compartment. In the osimertinib cohort, patients with CNS progression had increased CSF-tDNA VAFs at follow-up LP. Post-osimertinib CSF-tDNA VAF was strongly prognostic for CNS progression (HR = 6.2, P = 0.009). Detection of CSF-tDNA in lung cancer patients with suspected LMD is feasible and may have clinical utility. CSF-tDNA improves the sensitivity of LMD diagnosis, enables improved prognostication, and drives therapeutic strategies that account for spatial heterogeneity in resistance mechanisms.

Detecting High-Dose Methotrexate-Induced Brain Changes in Pediatric and Young Adult Cancer Survivors Using [18F]FDG PET/MRI: A Pilot Study.

Significant improvements in treatments for children with cancer have resulted in a growing population of childhood cancer survivors who may face long-term adverse outcomes. Here, we aimed to diagnose high-dose methotrexate-induced brain injury on [18F]FDG PET/MRI and correlate the results with cognitive impairment identified by neurocognitive testing in pediatric cancer survivors. Methods: In this prospective, single-center pilot study, 10 children and young adults with sarcoma (n = 5), lymphoma (n = 4), or leukemia (n = 1) underwent dedicated brain [18F]FDG PET/MRI and a 2-h expert neuropsychologic evaluation on the same day, including the Wechsler Abbreviated Scale of Intelligence, second edition, for intellectual functioning; Delis-Kaplan Executive Function System (DKEFS) for executive functioning; and Wide Range Assessment of Memory and Learning, second edition (WRAML), for verbal and visual memory. Using PMOD software, we measured the SUVmean, cortical thickness, mean cerebral blood flow (CBFmean), and mean apparent diffusion coefficient of 3 different cortical regions (prefrontal cortex, cingulate gyrus, and hippocampus) that are routinely involved during the above-specified neurocognitive testing. Standardized scores of different measures were converted to z scores. Pairs of multivariable regression models (one for z scores < 0 and one for z scores > 0) were fitted for each brain region, imaging measure, and test score. Heteroscedasticity regression models were used to account for heterogeneity in variances between brain regions and to adjust for clustering within patients. Results: The regression analysis showed a significant correlation between the SUVmean of the prefrontal cortex and cingulum and DKEFS-sequential tracking (DKEFS-TM4) z scores (P = 0.003 and P = 0.012, respectively). The SUVmean of the hippocampus did not correlate with DKEFS-TM4 z scores (P = 0.111). The SUVmean for any evaluated brain regions did not correlate significantly with WRAML-visual memory (WRAML-VIS) z scores. CBFmean showed a positive correlation with SUVmean (r = 0.56, P = 0.01). The CBFmean of the cingulum, hippocampus, and prefrontal cortex correlated significantly with DKEFS-TM4 (all P < 0.001). In addition, the hippocampal CBFmean correlated significantly with negative WRAML-VIS z scores (P = 0.003). Conclusion: High-dose methotrexate-induced brain injury can manifest as a reduction in glucose metabolism and blood flow in specific brain areas, which can be detected with [18F]FDG PET/MRI. The SUVmean and CBFmean of the prefrontal cortex and cingulum can serve as quantitative measures for detecting executive functioning problems. Hippocampal CBFmean could also be useful for monitoring memory problems.

Emergence of Boltzmann Subspaces in Open Quantum Systems Far from Equilibrium.

Single molecule junctions are important examples of complex out-of-equilibrium many-body quantum systems. We identify a nontrivial clustering of steady state populations into distinctive subspaces with Boltzmann-like statistics, which persist far from equilibrium. Such Boltzmann subspaces significantly reduce the information needed to describe the steady state, enabling modeling of high-dimensional systems that are otherwise beyond the reach of current computations. The emergence of Boltzmann subspaces is demonstrated analytically and numerically for fermionic transport systems of increasing complexity.

Arterial Spin-Labeling and DSC Perfusion Metrics Improve Agreement in Neuroradiologists' Clinical Interpretations of Posttreatment High-Grade Glioma Surveillance MR Imaging-An Institutional Experience.

BACKGROUND AND PURPOSE: MR perfusion has shown value in the evaluation of posttreatment high-grade gliomas, but few studies have shown its impact on the consistency and confidence of neuroradiologists' interpretation in routine clinical practice. We evaluated the impact of adding MR perfusion metrics to conventional contrast-enhanced MR imaging in posttreatment high-grade glioma surveillance imaging. MATERIALS AND METHODS: This retrospective study included 45 adults with high-grade gliomas who had posttreatment perfusion MR imaging. Four neuroradiologists assigned Brain Tumor Reporting and Data System scores for each examination on the basis of the interpretation of contrast-enhanced MR imaging and then after the addition of arterial spin-labeling-CBF, DSC-relative CBV, and DSC-fractional tumor burden. Interrater agreement and rater agreement with a multidisciplinary consensus group were assessed with κ statistics. Raters used a 5-point Likert scale to report confidence scores. The frequency of clinically meaningful score changes resulting from the addition of each perfusion metric was determined. RESULTS: Interrater agreement was moderate for contrast-enhanced MR imaging alone (κ = 0.63) and higher with perfusion metrics (arterial spin-labeling-CBF, κ = 0.67; DSC-relative CBV, κ = 0.66; DSC-fractional tumor burden, κ = 0.70). Agreement between raters and consensus was highest with DSC-fractional tumor burden (κ = 0.66-0.80). Confidence scores were highest with DSC-fractional tumor burden. Across all raters, the addition of perfusion resulted in clinically meaningful interpretation changes in 2%-20% of patients compared with contrast-enhanced MR imaging alone. CONCLUSIONS: Adding perfusion to contrast-enhanced MR imaging improved interrater agreement, rater agreement with consensus, and rater confidence in the interpretation of posttreatment high-grade glioma MR imaging, with the highest agreement and confidence scores seen with DSC-fractional tumor burden. Perfusion MR imaging also resulted in interpretation changes that could change therapeutic management in up to 20% of patients.

A rare non-gadolinium enhancing sarcoma brain metastasis with microenvironment dominated by tumor-associated macrophages.

Brain metastases occur in 1% of sarcoma cases and are associated with a median overall survival of 6 months. We report a rare case of a brain metastasis with unique radiologic and histopathologic features in a patient with low grade fibromyxoid sarcoma (LGFMS) previously treated with immune checkpoint inhibitor (ICI) therapy. The lone metastasis progressed in the midbrain tegmentum over 15 months as a non-enhancing, T2-hyperintense lesion with peripheral diffusion restriction, mimicking a demyelinating lesion. Histopathology of the lesion at autopsy revealed a rich infiltrate of tumor-associated macrophages (TAMs) with highest density at the leading edge of the metastasis, whereas there was a paucity of lymphocytes, suggestive of an immunologically cold environment. Given the important immunosuppressive and tumor-promoting functions of TAMs in gliomas and carcinoma/melanoma brain metastases, this unusual case provides an interesting example of a dense TAM infiltrate in a much rarer sarcoma brain metastasis.

Tumor treating fields increases blood-brain barrier permeability and relative cerebral blood volume in patients with glioblastoma.

BACKGROUND AND OBJECTIVE: 200 kHz tumor treating fields (TTFields) is clinically approved for newly-diagnosed glioblastoma (nGBM). Because its effects on conventional surveillance MRI brain scans are equivocal, we investigated its effects on perfusion MRI (pMRI) brain scans. METHODS: Each patient underwent institutional standard pMRI: dynamic contrast-enhanced (DCE) and dynamic susceptibility contrast (DSC) pMRI at three time points: baseline, 2-, and 6-months on-adjuvant therapy. At each timepoint, the difference between T1 pre- versus post-contrast tumor volume (ΔT1) and these pMRI metrics were evaluated: normalized and standardized relative cerebral blood volume (nRCBV, sRCBV); fractional plasma volume (Vp), volume of extravascular extracellular space (EES) per volume of tissue (Ve), blood-brain barrier (BBB) permeability (Ktrans), and time constant for gadolinium reflux from EES back into the vascular system (Kep). Between-group comparisons were performed using rank-sum analysis, and bootstrapping evaluated likely reproducibility of the results. RESULTS: Among 13 pMRI datasets (11 nGBM, 2 recurrent GBM), therapies included temozolomide-only (n = 9) and temozolomide + TTFields (n = 4). No significant differences were found in patient or tumor characteristics. Compared to temozolomide-only, temozolomide + TTFields did not significantly affect the percent-change in pMRI metrics from baseline to 2 months. But during the 2- to 6-month period, temozolomide + TTFields significantly increased the percent-change in nRCBV (+26.9% [interquartile range 55.1%] vs -39.1% [37.0%], p = 0.049), sRCBV (+9.5% [39.7%] vs -30.5% [39.4%], p = 0.049), Ktrans (+54.6% [1768.4%] vs -26.9% [61.2%], p = 0.024), Ve (+111.0% [518.1%] vs -13.0% [22.5%], p = 0.048), and Vp (+98.8% [2172.4%] vs -24.6% [53.3%], p = 0.024) compared to temozolomide-only. CONCLUSION: Using pMRI, we provide initial in-human validation of pre-clinical studies regarding the effects of TTFields on tumor blood volume and BBB permeability in GBM.

Investigating the value of radiomics stemming from DSC quantitative biomarkers in IDH mutation prediction in gliomas.

Objective: This study aims to assess the value of biomarker based radiomics to predict IDH mutation in gliomas. The patient cohort consists of 160 patients histopathologicaly proven of primary glioma (WHO grades 2-4) from 3 different centers. Methods: To quantify the DSC perfusion signal two different mathematical modeling methods were used (Gamma fitting, leakage correction algorithms) considering the assumptions about the compartments contributing in the blood flow between the extra- and intra vascular space. Results: The Mean slope of increase (MSI) and the K1 parameter of the bidirectional exchange model exhibited the highest performance with (ACC 74.3% AUROC 74.2%) and (ACC 75% AUROC 70.5%) respectively. Conclusion: The proposed framework on DSC-MRI radiogenomics in gliomas has the potential of becoming a reliable diagnostic support tool exploiting the mathematical modeling of the DSC signal to characterize IDH mutation status through a more reproducible and standardized signal analysis scheme for facilitating clinical translation.

Brain Tumor Imaging: Review of Conventional and Advanced Techniques.

Approaches to central nervous system (CNS) tumor classification and evaluation have undergone multiple iterations over the past few decades, in large part due to our growing understanding of the influence of genetics on tumor behavior and our refinement of brain tumor imaging techniques. Computed tomography and magnetic resonance imaging (MRI) both play a critical role in the diagnosis and monitoring of brain tumors, although MRI has become especially important due to its superior soft tissue resolution. The purpose of this article will be to briefly review the fundamentals of conventional and advanced techniques used in brain tumor imaging. We will also highlight the applications of these imaging tools in the context of commonly encountered tumors based on the most recently updated 2021 World Health Organization (WHO) classification of CNS tumors framework.

Retinal artery and vein occlusion in calciphylaxis.

Purpose: To report a case of branch retinal artery occlusion (BRAO) followed by branch retinal vein occlusion (BRVO) and paracentral acute middle maculopathy (PAMM) in a patient with confirmed calciphylaxis. Observations: A 52-year-old female with a history of BRAO in the right eye one-year prior presented with decreased vision and a new inferotemporal scotoma. Computed tomography angiography of the head and neck demonstrated vascular calcifications at the origin of both ophthalmic arteries, which were otherwise poorly visualized. Ophthalmic examination demonstrated retinal whitening superiorly with intraretinal hemorrhages inferiorly. Optical coherence tomography (OCT) demonstrated middle retinal hyperreflectivity and a mild epiretinal membrane. Fluorescein angiography (FFA) demonstrated delayed perfusion of superior retinal arcade. On further questioning, patient was found to have a history of IgA nephropathy with end-stage renal disease, secondary hyperparathyroidism and calciphylaxis. Calciphylaxis is a systemic disease, characterized by high levels of calcium and progressive calcification of the vascular medial layer leading to ischemia. Anterior ischemic optic neuropathy (AION) and crystalline retinopathy have been reported as ocular manifestations of calciphylaxis, however, there are very few reports on ophthalmic manifestations of calciphylaxis. Conclusion and importance: Clinical manifestations of calciphylaxis are variable and a detailed clinical history is important to suspect calciphylaxis. Calciphylaxis should be considered in the differential diagnosis of BRAO, BRVO, PAMM or any ophthalmic vascular manifestation in patients with end-stage renal disease.

Conventional and Advanced Imaging Techniques in Post-treatment Glioma Imaging.

Despite decades of advancement in the diagnosis and therapy of gliomas, the most malignant primary brain tumors, the overall survival rate is still dismal, and their post-treatment imaging appearance remains very challenging to interpret. Since the limitations of conventional magnetic resonance imaging (MRI) in the distinction between recurrence and treatment effect have been recognized, a variety of advanced MR and functional imaging techniques including diffusion-weighted imaging (DWI), diffusion tensor imaging (DTI), perfusion-weighted imaging (PWI), MR spectroscopy (MRS), as well as a variety of radiotracers for single photon emission computed tomography (SPECT) and positron emission tomography (PET) have been investigated for this indication along with voxel-based and more quantitative analytical methods in recent years. Machine learning and radiomics approaches in recent years have shown promise in distinguishing between recurrence and treatment effect as well as improving prognostication in a malignancy with a very short life expectancy. This review provides a comprehensive overview of the conventional and advanced imaging techniques with the potential to differentiate recurrence from treatment effect and includes updates in the state-of-the-art in advanced imaging with a brief overview of emerging experimental techniques. A series of representative cases are provided to illustrate the synthesis of conventional and advanced imaging with the clinical context which informs the radiologic evaluation of gliomas in the post-treatment setting.

How to stop using gadolinium chelates for magnetic resonance imaging: clinical-translational experiences with ferumoxytol.

Gadolinium chelates have been used as standard contrast agents for clinical MRI for several decades. However, several investigators recently reported that rare Earth metals such as gadolinium are deposited in the brain for months or years. This is particularly concerning for children, whose developing brain is more vulnerable to exogenous toxins compared to adults. Therefore, a search is under way for alternative MR imaging biomarkers. The United States Food and Drug Administration (FDA)-approved iron supplement ferumoxytol can solve this unmet clinical need: ferumoxytol consists of iron oxide nanoparticles that can be detected with MRI and provide significant T1- and T2-signal enhancement of vessels and soft tissues. Several investigators including our research group have started to use ferumoxytol off-label as a new contrast agent for MRI. This article reviews the existing literature on the biodistribution of ferumoxytol in children and compares the diagnostic accuracy of ferumoxytol- and gadolinium-chelate-enhanced MRI. Iron oxide nanoparticles represent a promising new class of contrast agents for pediatric MRI that can be metabolized and are not deposited in the brain.

2.5D and 3D segmentation of brain metastases with deep learning on multinational MRI data.

Introduction: Management of patients with brain metastases is often based on manual lesion detection and segmentation by an expert reader. This is a time- and labor-intensive process, and to that end, this work proposes an end-to-end deep learning segmentation network for a varying number of available MRI available sequences. Methods: We adapt and evaluate a 2.5D and a 3D convolution neural network trained and tested on a retrospective multinational study from two independent centers, in addition, nnU-Net was adapted as a comparative benchmark. Segmentation and detection performance was evaluated by: (1) the dice similarity coefficient, (2) a per-metastases and the average detection sensitivity, and (3) the number of false positives. Results: The 2.5D and 3D models achieved similar results, albeit the 2.5D model had better detection rate, whereas the 3D model had fewer false positive predictions, and nnU-Net had fewest false positives, but with the lowest detection rate. On MRI data from center 1, the 2.5D, 3D, and nnU-Net detected 79%, 71%, and 65% of all metastases; had an average per patient sensitivity of 0.88, 0.84, and 0.76; and had on average 6.2, 3.2, and 1.7 false positive predictions per patient, respectively. For center 2, the 2.5D, 3D, and nnU-Net detected 88%, 86%, and 78% of all metastases; had an average per patient sensitivity of 0.92, 0.91, and 0.85; and had on average 1.0, 0.4, and 0.1 false positive predictions per patient, respectively. Discussion/Conclusion: Our results show that deep learning can yield highly accurate segmentations of brain metastases with few false positives in multinational data, but the accuracy degrades for metastases with an area smaller than 0.4 cm2.

MRI pulse sequence integration for deep-learning-based brain metastases segmentation.

PURPOSE: Magnetic resonance (MR) imaging is an essential diagnostic tool in clinical medicine. Recently, a variety of deep-learning methods have been applied to segmentation tasks in medical images, with promising results for computer-aided diagnosis. For MR images, effectively integrating different pulse sequences is important to optimize performance. However, the best way to integrate different pulse sequences remains unclear. In addition, networks trained with a certain subset of pulse sequences as input are unable to perform when given a subset of those pulse sequences. In this study, we evaluate multiple architectural features and characterize their effects in the task of metastasis segmentation while creating a method to robustly train a network to be able to work given any strict subset of the pulse sequences available during training. METHODS: We use a 2.5D DeepLabv3 segmentation network to segment metastases lesions on brain MR's with four pulse sequence inputs. To study how we can best integrate MR pulse sequences for this task, we consider (1) different pulse sequence integration schemas, combining our features at early, middle, and late points within a deep network, (2) different modes of weight sharing for parallel network branches, and (3) a novel integration level dropout layer, which will allow the networks to be robust to performing inference on input with only a subset of pulse sequences available at the training. RESULTS: We find that levels of integration and modes of weight sharing that favor low variance work best in our regime of small amounts of training data (n = 100). By adding an input-level dropout layer, we could preserve the overall performance of these networks while allowing for inference on inputs with missing pulse sequences. We illustrate not only the generalizability of the network but also the utility of this robustness when applying the trained model to data from a different center, which does not use the same pulse sequences. Finally, we apply network visualization methods to better understand which input features are most important for network performance. CONCLUSIONS: Together, these results provide a framework for building networks with enhanced robustness to missing data while maintaining comparable performance in medical imaging applications.

Multicenter DSC-MRI-Based Radiomics Predict IDH Mutation in Gliomas.

To address the current lack of dynamic susceptibility contrast magnetic resonance imaging (DSC-MRI)-based radiomics to predict isocitrate dehydrogenase (IDH) mutations in gliomas, we present a multicenter study that featured an independent exploratory set for radiomics model development and external validation using two independent cohorts. The maximum performance of the IDH mutation status prediction on the validation set had an accuracy of 0.544 (Cohen's kappa: 0.145, F1-score: 0.415, area under the curve-AUC: 0.639, sensitivity: 0.733, specificity: 0.491), which significantly improved to an accuracy of 0.706 (Cohen's kappa: 0.282, F1-score: 0.474, AUC: 0.667, sensitivity: 0.6, specificity: 0.736) when dynamic-based standardization of the images was performed prior to the radiomics. Model explainability using local interpretable model-agnostic explanations (LIME) and Shapley additive explanations (SHAP) revealed potential intuitive correlations between the IDH-wildtype increased heterogeneity and the texture complexity. These results strengthened our hypothesis that DSC-MRI radiogenomics in gliomas hold the potential to provide increased predictive performance from models that generalize well and provide understandable patterns between IDH mutation status and the extracted features toward enabling the clinical translation of radiogenomics in neuro-oncology.

EGFR-targeted intraoperative fluorescence imaging detects high-grade glioma with panitumumab-IRDye800 in a phase 1 clinical trial.

Rationale: First-line therapy for high-grade gliomas (HGGs) includes maximal safe surgical resection. The extent of resection predicts overall survival, but current neuroimaging approaches lack tumor specificity. The epidermal growth factor receptor (EGFR) is a highly expressed HGG biomarker. We evaluated the safety and feasibility of an anti-EGFR antibody, panitumuab-IRDye800, at subtherapeutic doses as an imaging agent for HGG. Methods: Eleven patients with contrast-enhancing HGGs were systemically infused with panitumumab-IRDye800 at a low (50 mg) or high (100 mg) dose 1-5 days before surgery. Near-infrared fluorescence imaging was performed intraoperatively and ex vivo, to identify the optimal tumor-to-background ratio by comparing mean fluorescence intensities of tumor and histologically uninvolved tissue. Fluorescence was correlated with preoperative T1 contrast, tumor size, EGFR expression and other biomarkers. Results: No adverse events were attributed to panitumumab-IRDye800. Tumor fragments as small as 5 mg could be detected ex vivo and detection threshold was dose dependent. In tissue sections, panitumumab-IRDye800 was highly sensitive (95%) and specific (96%) for pathology confirmed tumor containing tissue. Cellular delivery of panitumumab-IRDye800 was correlated to EGFR overexpression and compromised blood-brain barrier in HGG, while normal brain tissue showed minimal fluorescence. Intraoperative fluorescence improved optical contrast in tumor tissue within and beyond the T1 contrast-enhancing margin, with contrast-to-noise ratios of 9.5 ± 2.1 and 3.6 ± 1.1, respectively. Conclusions: Panitumumab-IRDye800 provided excellent tumor contrast and was safe at both doses. Smaller fragments of tumor could be detected at the 100 mg dose and thus more suitable for intraoperative imaging.

Clinical Review of Computed Tomography and MR Perfusion Imaging in Neuro-Oncology.

Neuroimaging plays an essential role in the initial diagnosis and continued surveillance of intracranial neoplasms. The advent of perfusion techniques with computed tomography and MR imaging have proven useful in neuro-oncology, offering enhanced approaches for tumor grading, guiding stereotactic biopsies, and monitoring treatment efficacy. Perfusion imaging can help to identify treatment-related processes, such as radiation necrosis, pseudoprogression, and pseudoregression, and can help to inform treatment-related decision making. Perfusion imaging is useful to differentiate between tumor types and between tumor and nonneoplastic conditions. This article reviews the clinical relevance and implications of perfusion imaging in neuro-oncology and highlights promising perfusion biomarkers.

Handling missing MRI sequences in deep learning segmentation of brain metastases: a multicenter study.

The purpose of this study was to assess the clinical value of a deep learning (DL) model for automatic detection and segmentation of brain metastases, in which a neural network is trained on four distinct MRI sequences using an input-level dropout layer, thus simulating the scenario of missing MRI sequences by training on the full set and all possible subsets of the input data. This retrospective, multicenter study, evaluated 165 patients with brain metastases. The proposed input-level dropout (ILD) model was trained on multisequence MRI from 100 patients and validated/tested on 10/55 patients, in which the test set was missing one of the four MRI sequences used for training. The segmentation results were compared with the performance of a state-of-the-art DeepLab V3 model. The MR sequences in the training set included pre-gadolinium and post-gadolinium (Gd) T1-weighted 3D fast spin echo, post-Gd T1-weighted inversion recovery (IR) prepped fast spoiled gradient echo, and 3D fluid attenuated inversion recovery (FLAIR), whereas the test set did not include the IR prepped image-series. The ground truth segmentations were established by experienced neuroradiologists. The results were evaluated using precision, recall, Intersection over union (IoU)-score and Dice score, and receiver operating characteristics (ROC) curve statistics, while the Wilcoxon rank sum test was used to compare the performance of the two neural networks. The area under the ROC curve (AUC), averaged across all test cases, was 0.989 ± 0.029 for the ILD-model and 0.989 ± 0.023 for the DeepLab V3 model (p = 0.62). The ILD-model showed a significantly higher Dice score (0.795 ± 0.104 vs. 0.774 ± 0.104, p = 0.017), and IoU-score (0.561 ± 0.225 vs. 0.492 ± 0.186, p < 0.001) compared to the DeepLab V3 model, and a significantly lower average false positive rate of 3.6/patient vs. 7.0/patient (p < 0.001) using a 10 mm3 lesion-size limit. The ILD-model, trained on all possible combinations of four MRI sequences, may facilitate accurate detection and segmentation of brain metastases on a multicenter basis, even when the test cohort is missing input MRI sequences.