ABSTRACT
BACKGROUND: The extended acquisition times required for MRI limit its availability in resource-constrained settings. Consequently, accelerating MRI by undersampling k-space data, which is necessary to reconstruct an image, has been a long-standing but important challenge. We aimed to develop a deep convolutional neural network (dCNN) optimisation method for MRI reconstruction and to reduce scan times and evaluate its effect on image quality and accuracy of oncological imaging biomarkers. METHODS: In this multicentre, retrospective, cohort study, MRI data from patients with glioblastoma treated at Heidelberg University Hospital (775 patients and 775 examinations) and from the phase 2 CORE trial (260 patients, 1083 examinations, and 58 institutions) and the phase 3 CENTRIC trial (505 patients, 3147 examinations, and 139 institutions) were used to develop, train, and test dCNN for reconstructing MRI from highly undersampled single-coil k-space data with various acceleration rates (R=2, 4, 6, 8, 10, and 15). Independent testing was performed with MRIs from the phase 2/3 EORTC-26101 trial (528 patients with glioblastoma, 1974 examinations, and 32 institutions). The similarity between undersampled dCNN-reconstructed and original MRIs was quantified with various image quality metrics, including structural similarity index measure (SSIM) and the accuracy of undersampled dCNN-reconstructed MRI on downstream radiological assessment of imaging biomarkers in oncology (automated artificial intelligence-based quantification of tumour burden and treatment response) was performed in the EORTC-26101 test dataset. The public NYU Langone Health fastMRI brain test dataset (558 patients and 558 examinations) was used to validate the generalisability and robustness of the dCNN for reconstructing MRIs from available multi-coil (parallel imaging) k-space data. FINDINGS: In the EORTC-26101 test dataset, the median SSIM of undersampled dCNN-reconstructed MRI ranged from 0·88 to 0·99 across different acceleration rates, with 0·92 (95% CI 0·92-0·93) for 10-times acceleration (R=10). The 10-times undersampled dCNN-reconstructed MRI yielded excellent agreement with original MRI when assessing volumes of contrast-enhancing tumour (median DICE for spatial agreement of 0·89 [95% CI 0·88 to 0·89]; median volume difference of 0·01 cm3 [95% CI 0·00 to 0·03] equalling 0·21%; p=0·0036 for equivalence) or non-enhancing tumour or oedema (median DICE of 0·94 [95% CI 0·94 to 0·95]; median volume difference of -0·79 cm3 [95% CI -0·87 to -0·72] equalling -1·77%; p=0·023 for equivalence) in the EORTC-26101 test dataset. Automated volumetric tumour response assessment in the EORTC-26101 test dataset yielded an identical median time to progression of 4·27 months (95% CI 4·14 to 4·57) when using 10-times-undersampled dCNN-reconstructed or original MRI (log-rank p=0·80) and agreement in the time to progression in 374 (95·2%) of 393 patients with data. The dCNN generalised well to the fastMRI brain dataset, with significant improvements in the median SSIM when using multi-coil compared with single-coil k-space data (p<0·0001). INTERPRETATION: Deep-learning-based reconstruction of undersampled MRI allows for a substantial reduction of scan times, with a 10-times acceleration demonstrating excellent image quality while preserving the accuracy of derived imaging biomarkers for the assessment of oncological treatment response. Our developments are available as open source software and hold considerable promise for increasing the accessibility to MRI, pending further prospective validation. FUNDING: Deutsche Forschungsgemeinschaft (German Research Foundation) and an Else Kröner Clinician Scientist Endowed Professorship by the Else Kröner Fresenius Foundation.
Subject(s)
Deep Learning , Glioblastoma , Humans , Artificial Intelligence , Biomarkers , Cohort Studies , Glioblastoma/diagnostic imaging , Magnetic Resonance Imaging , Retrospective StudiesABSTRACT
OBJECTIVE: Multiple variables beyond the extent of recanalization can impact the clinical outcome after acute ischemic stroke due to large vessel occlusions. Here, we assessed the influence of small vessel disease and cortical atrophy on clinical outcome using native cranial computed tomography (NCCT) in a large single-center cohort. METHODS: A total of 1103 consecutive patients who underwent endovascular treatment (EVT) due to occlusion of the middle cerebral artery territory were included. NCCT data were visually assessed for established markers of age-related white matter changes (ARWMC) and brain atrophy. All images were evaluated separately by two readers to assess the inter-observer variability. Regression and machine learning models were built to determine the predictive relevance of ARWMC and atrophy in the presence of important baseline clinical and imaging metrics. RESULTS: Patients with favorable outcome presented lower values for all measured metrics of pre-existing brain deterioration (p < 0.001). Both ARWMC (p < 0.05) and cortical atrophy (p < 0.001) were independent predictors of clinical outcome at 90 days when controlled for confounders in both regression analyses and led to a minor improvement of prediction accuracy in machine learning models (p < 0.001), with atrophy among the top-5 predictors. CONCLUSION: NCCT-based cortical atrophy and ARWMC scores on NCCT were strong and independent predictors of clinical outcome after EVT. CLINICAL RELEVANCE STATEMENT: Visual assessment of cortical atrophy and age-related white matter changes on CT could improve the prediction of clinical outcome after thrombectomy in machine learning models which may be integrated into existing clinical routines and facilitate patient selection. KEY POINTS: ⢠Cortical atrophy and age-related white matter changes were quantified using CT-based visual scores. ⢠Atrophy and age-related white matter change scores independently predicted clinical outcome after mechanical thrombectomy and improved machine learning-based prediction models. ⢠Both scores could easily be integrated into existing clinical routines and prediction models.
Subject(s)
Brain Ischemia , Endovascular Procedures , Ischemic Stroke , Stroke , Humans , Stroke/diagnostic imaging , Stroke/surgery , Brain Ischemia/diagnostic imaging , Brain Ischemia/surgery , Tomography, X-Ray Computed/methods , Thrombectomy/methods , Atrophy , Treatment Outcome , Endovascular Procedures/methods , Retrospective StudiesABSTRACT
OBJECTIVES: Radiomic features have demonstrated encouraging results for non-invasive detection of molecular biomarkers, but the lack of guidelines for pre-processing MRI-data has led to poor generalizability. Here, we assessed the influence of different MRI-intensity normalization techniques on the performance of radiomics-based models for predicting molecular glioma subtypes. METHODS: Preoperative MRI-data from n = 615 patients with newly diagnosed glioma and known isocitrate dehydrogenase (IDH) and 1p/19q status were pre-processed using four different methods: no normalization (naive), N4 bias field correction (N4), N4 followed by either WhiteStripe (N4/WS), or z-score normalization (N4/z-score). A total of 377 Image-Biomarker-Standardisation-Initiative-compliant radiomic features were extracted from each normalized data, and 9 different machine-learning algorithms were trained for multiclass prediction of molecular glioma subtypes (IDH-mutant 1p/19q codeleted vs. IDH-mutant 1p/19q non-codeleted vs. IDH wild type). External testing was performed in public glioma datasets from UCSF (n = 410) and TCGA (n = 160). RESULTS: Support vector machine yielded the best performance with macro-average AUCs of 0.84 (naive), 0.84 (N4), 0.87 (N4/WS), and 0.87 (N4/z-score) in the internal test set. Both N4/WS and z-score outperformed the other approaches in the external UCSF and TCGA test sets with macro-average AUCs ranging from 0.85 to 0.87, replicating the performance of the internal test set, in contrast to macro-average AUCs ranging from 0.19 to 0.45 for naive and 0.26 to 0.52 for N4 alone. CONCLUSION: Intensity normalization of MRI data is essential for the generalizability of radiomic-based machine-learning models. Specifically, both N4/WS and N4/z-score approaches allow to preserve the high model performance, yielding generalizable performance when applying the developed radiomic-based machine-learning model in an external heterogeneous, multi-institutional setting. CLINICAL RELEVANCE STATEMENT: Intensity normalization such as N4/WS or N4/z-score can be used to develop reliable radiomics-based machine learning models from heterogeneous multicentre MRI datasets and provide non-invasive prediction of glioma subtypes. KEY POINTS: ⢠MRI-intensity normalization increases the stability of radiomics-based models and leads to better generalizability. ⢠Intensity normalization did not appear relevant when the developed model was applied to homogeneous data from the same institution. ⢠Radiomic-based machine learning algorithms are a promising approach for simultaneous classification of IDH and 1p/19q status of glioma.
Subject(s)
Brain Neoplasms , Glioma , Humans , Brain Neoplasms/diagnostic imaging , Brain Neoplasms/genetics , Radiomics , Glioma/diagnostic imaging , Glioma/genetics , Magnetic Resonance Imaging/methods , Biomarkers , Isocitrate Dehydrogenase/genetics , Mutation , Retrospective StudiesABSTRACT
OBJECTIVE: Anorexia nervosa (AN) and obesity are weight-related disorders with imbalances in energy homeostasis that may be due to hormonal dysregulation. Given the importance of the hypothalamus in hormonal regulation, we aimed to identify morphometric alterations to hypothalamic subregions linked to these conditions and their connection to appetite-regulating hormones. METHODS: Structural magnetic resonance imaging (MRI) was obtained from 78 patients with AN, 27 individuals with obesity and 100 normal-weight healthy controls. Leptin, ghrelin, and insulin blood levels were measured in a subsample of each group. An automated segmentation method was used to segment the hypothalamus and its subregions. Volumes of the hypothalamus and its subregions were compared between groups, and correlational analysis was employed to assess the relationship between morphometric measurements and appetite-regulating hormone levels. RESULTS: While accounting for total brain volume, patients with AN displayed a smaller volume in the inferior-tubular subregion (ITS). Conversely, obesity was associated with a larger volume in the anterior-superior, ITS, posterior subregions (PS), and entire hypothalamus. There were no significant volumetric differences between AN subtypes. Leptin correlated positively with PS volume, whereas ghrelin correlated negatively with the whole hypothalamus volume in the entire cohort. However, appetite-regulating hormone levels did not mediate the effects of body mass index on volumetric measures. CONCLUSION: Our results indicate the importance of regional structural hypothalamic alterations in AN and obesity, extending beyond global changes to brain volume. Furthermore, these alterations may be linked to changes in hormonal appetite regulation. However, given the small sample size in our correlation analysis, further analyses in a larger sample size are warranted. PUBLIC SIGNIFICANCE: Using an automated segmentation method to investigate morphometric alterations of hypothalamic subregions in AN and obesity, this study provides valuable insights into the complex interplay between hypothalamic alterations, hormonal appetite regulation, and body weight, highlighting the need for further research to uncover underlying mechanisms.
Subject(s)
Anorexia Nervosa , Leptin , Humans , Anorexia Nervosa/diagnostic imaging , Appetite/physiology , Ghrelin , Obesity/diagnostic imaging , Hypothalamus/diagnostic imagingABSTRACT
BACKGROUND AND PURPOSE: This study assessed the predictive performance and relative importance of clinical, multimodal imaging, and angiographic characteristics for predicting the clinical outcome of endovascular treatment for acute ischemic stroke. METHODS: A consecutive series of 246 patients with acute ischemic stroke and large vessel occlusion in the anterior circulation who underwent endovascular treatment between April 2014 and January 2018 was analyzed. Clinical, conventional imaging (electronic Alberta Stroke Program Early CT Score, acute ischemic volume, site of vessel occlusion, and collateral score), and advanced imaging characteristics (CT-perfusion with quantification of ischemic penumbra and infarct core volumes) before treatment as well as angiographic (interval groin puncture-recanalization, modified Thrombolysis in Cerebral Infarction score) and postinterventional clinical (National Institutes of Health Stroke Scale score after 24 hours) and imaging characteristics (electronic Alberta Stroke Program Early CT Score, final infarction volume after 18-36 hours) were assessed. The modified Rankin Scale (mRS) score at 90 days (mRS-90) was used to measure patient outcome (favorable outcome: mRS-90 ≤2 versus unfavorable outcome: mRS-90 >2). Machine-learning with gradient boosting classifiers was used to assess the performance and relative importance of the extracted characteristics for predicting mRS-90. RESULTS: Baseline clinical and conventional imaging characteristics predicted mRS-90 with an area under the receiver operating characteristics curve of 0.740 (95% CI, 0.733-0.747) and an accuracy of 0.711 (95% CI, 0.705-0.717). Advanced imaging with CT-perfusion did not improved the predictive performance (area under the receiver operating characteristics curve, 0.747 [95% CI, 0.740-0.755]; accuracy, 0.720 [95% CI, 0.714-0.727]; P=0.150). Further inclusion of angiographic and postinterventional characteristics significantly improved the predictive performance (area under the receiver operating characteristics curve, 0.856 [95% CI, 0.850-0.861]; accuracy, 0.804 [95% CI, 0.799-0.810]; P<0.001). The most important parameters for predicting mRS 90 were National Institutes of Health Stroke Scale score after 24 hours (importance =100%), premorbid mRS score (importance =44%) and final infarction volume on postinterventional CT after 18 to 36 hours (importance =32%). CONCLUSIONS: Integrative assessment of clinical, multimodal imaging, and angiographic characteristics with machine-learning allowed to accurately predict the clinical outcome following endovascular treatment for acute ischemic stroke. Thereby, premorbid mRS was the most important clinical predictor for mRS-90, and the final infarction volume was the most important imaging predictor, while the extent of hemodynamic impairment on CT-perfusion before treatment had limited importance.
Subject(s)
Clinical Decision Rules , Endovascular Procedures , Ischemic Stroke/surgery , Thrombectomy , Aged , Aged, 80 and over , Carotid Artery Thrombosis/diagnostic imaging , Carotid Artery Thrombosis/surgery , Computed Tomography Angiography , Female , Humans , Image Processing, Computer-Assisted , Infarction, Anterior Cerebral Artery/diagnostic imaging , Infarction, Anterior Cerebral Artery/surgery , Infarction, Middle Cerebral Artery/diagnostic imaging , Infarction, Middle Cerebral Artery/surgery , Ischemic Stroke/diagnostic imaging , Ischemic Stroke/physiopathology , Machine Learning , Male , Perfusion Imaging , Retrospective Studies , Tomography, X-Ray ComputedABSTRACT
Glioblastoma (GBM) is the most malignant brain tumor of adults and is characterized by extensive cell dissemination within the brain parenchyma and enhanced angiogenesis. Effective preclinical modeling of these key features suffers from several shortcomings. Aim of this study was to determine whether modulating the expression of extracellular matrix (ECM) modifiers in proneural (PN) and mesenchymal (MES) cancer stem cells (CSCs) and in conventional glioma cell lines (GCLs) might improve tumor invasion and vascularization. To this end, we selected secreted, acidic and rich in cysteine-like 1 (SPARCL1) as a potential mediator of ECM remodeling in GBM. SPARCL1 transcript and protein expression was assessed in PN and MES CSCs as well as GCLs, in their xenografts and in patient-derived specimens by qPCR, WB and IHC. SPARCL1 expression was then enforced in both CSCs and GCLs by lentiviral-based transduction. The effect of SPARCL1 gain-of-function on microvascular proliferation, microglia activation and advanced imaging features was tested in intracranial xenografts by IHC and MRI and validated by chorioallantoic membrane (CAM) assays. SPARCL1 expression significantly enhanced the infiltrative and neoangiogenic features of PN and MES CSC/GCL-induced tumors, with the concomitant activation of inflammatory responses associated with the tumor microenvironment, thus resulting in experimental GBMs that reproduced both the parenchymal infiltration and the increased microvascular density, typical of GBM. Overall, these results indicate that SPARCL1 overexpression might be instrumental for the generation of CSC-derived preclinical models of GBM in which the main pathognomonic hallmarks of GBMs are retrievable, making them suitable for effective preclinical testing of therapeutics.
Subject(s)
Brain Neoplasms/metabolism , Calcium-Binding Proteins/metabolism , Extracellular Matrix Proteins/metabolism , Glioblastoma/metabolism , Neoplastic Stem Cells/metabolism , Neovascularization, Pathologic/metabolism , Animals , Cell Line, Tumor , Disease Models, Animal , Extracellular Matrix/metabolism , Female , Heterografts , Humans , Mice , Microglia/metabolismABSTRACT
Background Relevance of antiangiogenic treatment with bevacizumab in patients with glioblastoma is controversial because progression-free survival benefit did not translate into an overall survival (OS) benefit in randomized phase III trials. Purpose To perform longitudinal characterization of intratumoral angiogenesis and oxygenation by using dynamic susceptibility contrast agent-enhanced (DSC) MRI and evaluate its potential for predicting outcome from administration of bevacizumab. Materials and Methods In this secondary analysis of the prospective randomized phase II/III European Organization for Research and Treatment of Cancer 26101 trial conducted between October 2011 and December 2015 in 596 patients with first recurrence of glioblastoma, the subset of patients with availability of anatomic MRI and DSC MRI at baseline and first follow-up was analyzed. Patients were allocated into those administered bevacizumab (hereafter, the BEV group; either bevacizumab monotherapy or bevacizumab with lomustine) and those not administered bevacizumab (hereafter, the non-BEV group with lomustine monotherapy). Contrast-enhanced tumor volume, noncontrast-enhanced T2 fluid-attenuated inversion recovery (FLAIR) signal abnormality volume, Gaussian-normalized relative cerebral blood volume (nrCBV), Gaussian-normalized relative blood flow (nrCBF), and tumor metabolic rate of oxygen (nTMRO2) was quantified. The predictive ability of these imaging parameters was assessed with multivariable Cox regression and formal interaction testing. Results A total of 254 of 596 patients were evaluated (mean age, 57 years ± 11; 155 men; 161 in the BEV group and 93 in non-BEV group). Progression-free survival was longer in the BEV group (3.7 months; 95% confidence interval [CI]: 3.0, 4.2) compared with the non-BEV group (2.5 months; 95% CI: 1.5, 2.9; P = .01), whereas OS was not different (P = .15). The nrCBV decreased for the BEV group (-16.3%; interquartile range [IQR], -39.5% to 12.0%; P = .01), but not for the non-BEV group (1.2%; IQR, -17.9% to 23.3%; P = .19) between baseline and first follow-up. An identical pattern was observed for both nrCBF and nTMRO2 values. Contrast-enhanced tumor and noncontrast-enhanced T2 FLAIR signal abnormality volumes decreased for the BEV group (-66% [IQR, -83% to -35%] and -33% [IQR, -71% to -5%], respectively; P < .001 for both), whereas they increased for the non-BEV group (30% [IQR, -17% to 98%], P = .001; and 10% [IQR, -13% to 82%], P = .02, respectively) between baseline and first follow-up. None of the assessed MRI parameters were predictive for OS in the BEV group. Conclusion Bevacizumab treatment decreased tumor volumes, angiogenesis, and oxygenation, thereby reflecting its effectiveness for extending progression-free survival; however, these parameters were not predictive of overall survival (OS), which highlighted the challenges of identifying patients that derive an OS benefit from bevacizumab. © RSNA, 2020 Online supplemental material is available for this article. See also the editorial by Dillon in this issue.
Subject(s)
Angiogenesis Inhibitors/therapeutic use , Bevacizumab/therapeutic use , Brain Neoplasms/drug therapy , Glioblastoma/drug therapy , Magnetic Resonance Imaging/methods , Neovascularization, Pathologic/drug therapy , Antineoplastic Agents, Alkylating/therapeutic use , Brain Neoplasms/pathology , Contrast Media , Europe , Female , Glioblastoma/pathology , Humans , Lomustine/therapeutic use , Male , Middle Aged , Neoplasm Recurrence, Local , Prospective Studies , Survival AnalysisABSTRACT
OBJECTIVES: Patients with multiple sclerosis (MS) regularly undergo MRI for assessment of disease burden. However, interpretation may be time consuming and prone to intra- and interobserver variability. Here, we evaluate the potential of artificial neural networks (ANN) for automated volumetric assessment of MS disease burden and activity on MRI. METHODS: A single-institutional dataset with 334 MS patients (334 MRI exams) was used to develop and train an ANN for automated identification and volumetric segmentation of T2/FLAIR-hyperintense and contrast-enhancing (CE) lesions. Independent testing was performed in a single-institutional longitudinal dataset with 82 patients (266 MRI exams). We evaluated lesion detection performance (F1 scores), lesion segmentation agreement (DICE coefficients), and lesion volume agreement (concordance correlation coefficients [CCC]). Independent evaluation was performed on the public ISBI-2015 challenge dataset. RESULTS: The F1 score was maximized in the training set at a detection threshold of 7 mm3 for T2/FLAIR lesions and 14 mm3 for CE lesions. In the training set, mean F1 scores were 0.867 for T2/FLAIR lesions and 0.636 for CE lesions, as compared to 0.878 for T2/FLAIR lesions and 0.715 for CE lesions in the test set. Using these thresholds, the ANN yielded mean DICE coefficients of 0.834 and 0.878 for segmentation of T2/FLAIR and CE lesions in the training set (fivefold cross-validation). Corresponding DICE coefficients in the test set were 0.846 for T2/FLAIR lesions and 0.908 for CE lesions, and the CCC was ≥ 0.960 in each dataset. CONCLUSIONS: Our results highlight the capability of ANN for quantitative state-of-the-art assessment of volumetric lesion load on MRI and potentially enable a more accurate assessment of disease burden in patients with MS. KEY POINTS: ⢠Artificial neural networks (ANN) can accurately detect and segment both T2/FLAIR and contrast-enhancing MS lesions in MRI data. ⢠Performance of the ANN was consistent in a clinically derived dataset, with patients presenting all possible disease stages in MRI scans acquired from standard clinical routine rather than with high-quality research sequences. ⢠Computer-aided evaluation of MS with ANN could streamline both clinical and research procedures in the volumetric assessment of MS disease burden as well as in lesion detection.
Subject(s)
Brain/pathology , Magnetic Resonance Imaging/methods , Multiple Sclerosis/diagnosis , Neural Networks, Computer , Adult , Female , Humans , Male , Middle Aged , Reproducibility of ResultsABSTRACT
BACKGROUND: The Response Assessment in Neuro-Oncology (RANO) criteria and requirements for a uniform protocol have been introduced to standardise assessment of MRI scans in both clinical trials and clinical practice. However, these criteria mainly rely on manual two-dimensional measurements of contrast-enhancing (CE) target lesions and thus restrict both reliability and accurate assessment of tumour burden and treatment response. We aimed to develop a framework relying on artificial neural networks (ANNs) for fully automated quantitative analysis of MRI in neuro-oncology to overcome the inherent limitations of manual assessment of tumour burden. METHODS: In this retrospective study, we compiled a single-institution dataset of MRI data from patients with brain tumours being treated at Heidelberg University Hospital (Heidelberg, Germany; Heidelberg training dataset) to develop and train an ANN for automated identification and volumetric segmentation of CE tumours and non-enhancing T2-signal abnormalities (NEs) on MRI. Independent testing and large-scale application of the ANN for tumour segmentation was done in a single-institution longitudinal testing dataset from the Heidelberg University Hospital and in a multi-institutional longitudinal testing dataset from the prospective randomised phase 2 and 3 European Organisation for Research and Treatment of Cancer (EORTC)-26101 trial (NCT01290939), acquired at 38 institutions across Europe. In both longitudinal datasets, spatial and temporal tumour volume dynamics were automatically quantified to calculate time to progression, which was compared with time to progression determined by RANO, both in terms of reliability and as a surrogate endpoint for predicting overall survival. We integrated this approach for fully automated quantitative analysis of MRI in neuro-oncology within an application-ready software infrastructure and applied it in a simulated clinical environment of patients with brain tumours from the Heidelberg University Hospital (Heidelberg simulation dataset). FINDINGS: For training of the ANN, MRI data were collected from 455 patients with brain tumours (one MRI per patient) being treated at Heidelberg hospital between July 29, 2009, and March 17, 2017 (Heidelberg training dataset). For independent testing of the ANN, an independent longitudinal dataset of 40 patients, with data from 239 MRI scans, was collected at Heidelberg University Hospital in parallel with the training dataset (Heidelberg test dataset), and 2034 MRI scans from 532 patients at 34 institutions collected between Oct 26, 2011, and Dec 3, 2015, in the EORTC-26101 study were of sufficient quality to be included in the EORTC-26101 test dataset. The ANN yielded excellent performance for accurate detection and segmentation of CE tumours and NE volumes in both longitudinal test datasets (median DICE coefficient for CE tumours 0·89 [95% CI 0·86-0·90], and for NEs 0·93 [0·92-0·94] in the Heidelberg test dataset; CE tumours 0·91 [0·90-0·92], NEs 0·93 [0·93-0·94] in the EORTC-26101 test dataset). Time to progression from quantitative ANN-based assessment of tumour response was a significantly better surrogate endpoint than central RANO assessment for predicting overall survival in the EORTC-26101 test dataset (hazard ratios ANN 2·59 [95% CI 1·86-3·60] vs central RANO 2·07 [1·46-2·92]; p<0·0001) and also yielded a 36% margin over RANO (p<0·0001) when comparing reliability values (ie, agreement in the quantitative volumetrically defined time to progression [based on radiologist ground truth vs automated assessment with ANN] of 87% [266 of 306 with sufficient data] compared with 51% [155 of 306] with local vs independent central RANO assessment). In the Heidelberg simulation dataset, which comprised 466 patients with brain tumours, with 595 MRI scans obtained between April 27, and Sept 17, 2018, automated on-demand processing of MRI scans and quantitative tumour response assessment within the simulated clinical environment required 10 min of computation time (average per scan). INTERPRETATION: Overall, we found that ANN enabled objective and automated assessment of tumour response in neuro-oncology at high throughput and could ultimately serve as a blueprint for the application of ANN in radiology to improve clinical decision making. Future research should focus on prospective validation within clinical trials and application for automated high-throughput imaging biomarker discovery and extension to other diseases. FUNDING: Medical Faculty Heidelberg Postdoc-Program, Else Kröner-Fresenius Foundation.
Subject(s)
Brain Neoplasms/diagnostic imaging , Brain Neoplasms/therapy , Diagnosis, Computer-Assisted , Image Interpretation, Computer-Assisted , Magnetic Resonance Imaging , Neural Networks, Computer , Automation , Brain Neoplasms/pathology , Clinical Trials, Phase II as Topic , Clinical Trials, Phase III as Topic , Databases, Factual , Disease Progression , Female , Germany , Humans , Male , Multicenter Studies as Topic , Predictive Value of Tests , Randomized Controlled Trials as Topic , Reproducibility of Results , Retrospective Studies , Time Factors , Treatment Outcome , Tumor Burden , WorkflowABSTRACT
Brain extraction is a critical preprocessing step in the analysis of neuroimaging studies conducted with magnetic resonance imaging (MRI) and influences the accuracy of downstream analyses. The majority of brain extraction algorithms are, however, optimized for processing healthy brains and thus frequently fail in the presence of pathologically altered brain or when applied to heterogeneous MRI datasets. Here we introduce a new, rigorously validated algorithm (termed HD-BET) relying on artificial neural networks that aim to overcome these limitations. We demonstrate that HD-BET outperforms six popular, publicly available brain extraction algorithms in several large-scale neuroimaging datasets, including one from a prospective multicentric trial in neuro-oncology, yielding state-of-the-art performance with median improvements of +1.16 to +2.50 points for the Dice coefficient and -0.66 to -2.51 mm for the Hausdorff distance. Importantly, the HD-BET algorithm, which shows robust performance in the presence of pathology or treatment-induced tissue alterations, is applicable to a broad range of MRI sequence types and is not influenced by variations in MRI hardware and acquisition parameters encountered in both research and clinical practice. For broader accessibility, the HD-BET prediction algorithm is made freely available (www.neuroAI-HD.org) and may become an essential component for robust, automated, high-throughput processing of MRI neuroimaging data.
Subject(s)
Brain/diagnostic imaging , Image Processing, Computer-Assisted/methods , Magnetic Resonance Imaging , Neural Networks, Computer , Algorithms , Humans , Neuroimaging/methodsABSTRACT
Adrenomyeloneuropathy is the late-onset form of X-linked adrenoleukodystrophy, and is considered the most frequent metabolic hereditary spastic paraplegia. In adrenomyeloneuropathy the spinal cord is the main site of pathology. Differently from quantitative magnetic resonance imaging of the brain, little is known about the feasibility and utility of advanced neuroimaging in quantifying the spinal cord abnormalities in hereditary diseases. Moreover, little is known about the subtle pathological changes that can characterize the brain of adrenomyeloneuropathy subjects in the early stages of the disease. We performed a cross-sectional study on 13 patients with adrenomyeloneuropathy and 12 age-matched healthy control subjects who underwent quantitative magnetic resonance imaging to assess the structural changes of the upper spinal cord and brain. Total cord areas from C2-3 to T2-3 level were measured, and diffusion tensor imaging metrics, i.e. fractional anisotropy, mean, axial and radial diffusivity values were calculated in both grey and white matter of spinal cord. In the brain, grey matter regions were parcellated with Freesurfer and average volume and thickness, and mean diffusivity and fractional anisotropy from co-registered diffusion maps were calculated in each region. Brain white matter diffusion tensor imaging metrics were assessed using whole-brain tract-based spatial statistics, and tractography-based analysis on corticospinal tracts. Correlations among clinical, structural and diffusion tensor imaging measures were calculated. In patients total cord area was reduced by 26.3% to 40.2% at all tested levels (P < 0.0001). A mean 16% reduction of spinal cord white matter fractional anisotropy (P ≤ 0.0003) with a concomitant 9.7% axial diffusivity reduction (P < 0.009) and 34.5% radial diffusivity increase (P < 0.009) was observed, suggesting co-presence of axonal degeneration and demyelination. Brain tract-based spatial statistics showed a marked reduction of fractional anisotropy, increase of radial diffusivity (P < 0.001) and no axial diffusivity changes in several white matter tracts, including corticospinal tracts and optic radiations, indicating predominant demyelination. Tractography-based analysis confirmed the results within corticospinal tracts. No significant cortical volume and thickness reduction or grey matter diffusion tensor imaging values alterations were observed in patients. A correlation between radial diffusivity and disease duration along the corticospinal tracts (r = 0.806, P < 0.01) was found. In conclusion, in adrenomyeloneuropathy patients quantitative magnetic resonance imaging-derived measures identify and quantify structural changes in the upper spinal cord and brain which agree with the expected histopathology, and suggest that the disease could be primarily caused by a demyelination rather than a primitive axonal damage. The results of this study may also encourage the employment of quantitative magnetic resonance imaging in other hereditary diseases with spinal cord involvement.
Subject(s)
Adrenoleukodystrophy/diagnostic imaging , Adrenoleukodystrophy/pathology , Brain/pathology , Spinal Cord/pathology , Adult , Anisotropy , Case-Control Studies , Cross-Sectional Studies , Diffusion Tensor Imaging/methods , Diffusion Tensor Imaging/statistics & numerical data , Gray Matter/pathology , Humans , Male , Neuroimaging/statistics & numerical data , White Matter/pathology , Young AdultSubject(s)
Brain Ischemia , Stroke , Brain Ischemia/diagnosis , Brain Ischemia/therapy , Humans , Medical Futility , Stroke/diagnosis , Stroke/therapyABSTRACT
Purpose To develop a fully automated device- and sequence-independent convolutional neural network (CNN) for reliable and high-throughput labeling of heterogeneous, unstructured MRI data. Materials and Methods Retrospective, multicentric brain MRI data (2179 patients with glioblastoma, 8544 examinations, 63 327 sequences) from 249 hospitals and 29 scanner types were used to develop a network based on ResNet-18 architecture to differentiate nine MRI sequence types, including T1-weighted, postcontrast T1-weighted, T2-weighted, fluid-attenuated inversion recovery, susceptibility-weighted, apparent diffusion coefficient, diffusion-weighted (low and high b value), and gradient-recalled echo T2*-weighted and dynamic susceptibility contrast-related images. The two-dimensional-midsection images from each sequence were allocated to training or validation (approximately 80%) and testing (approximately 20%) using a stratified split to ensure balanced groups across institutions, patients, and MRI sequence types. The prediction accuracy was quantified for each sequence type, and subgroup comparison of model performance was performed using χ2 tests. Results On the test set, the overall accuracy of the CNN (ResNet-18) ensemble model among all sequence types was 97.9% (95% CI: 97.6, 98.1), ranging from 84.2% for susceptibility-weighted images (95% CI: 81.8, 86.6) to 99.8% for T2-weighted images (95% CI: 99.7, 99.9). The ResNet-18 model achieved significantly better accuracy compared with ResNet-50 despite its simpler architecture (97.9% vs 97.1%; P ≤ .001). The accuracy of the ResNet-18 model was not affected by the presence versus absence of tumor on the two-dimensional-midsection images for any sequence type (P > .05). Conclusion The developed CNN (www.github.com/neuroAI-HD/HD-SEQ-ID) reliably differentiates nine types of MRI sequences within multicenter and large-scale population neuroimaging data and may enhance the speed, accuracy, and efficiency of clinical and research neuroradiologic workflows. Keywords: MR-Imaging, Neural Networks, CNS, Brain/Brain Stem, Computer Applications-General (Informatics), Convolutional Neural Network (CNN), Deep Learning Algorithms, Machine Learning Algorithms Supplemental material is available for this article. © RSNA, 2023.
Subject(s)
Deep Learning , Humans , Brain/diagnostic imaging , Magnetic Resonance Imaging/methods , Neuroimaging , Retrospective Studies , Multicenter Studies as TopicABSTRACT
Background: This study investigates the influence of diffusion-weighted Magnetic Resonance Imaging (DWI-MRI) on radiomic-based prediction of glioma types according to molecular status and assesses the impact of DWI intensity normalization on model generalizability. Methods: Radiomic features, compliant with image biomarker standardization initiative standards, were extracted from preoperative MRI of 549 patients with diffuse glioma, known IDH, and 1p19q-status. Anatomical sequences (T1, T1c, T2, FLAIR) underwent N4-Bias Field Correction (N4) and WhiteStripe normalization (N4/WS). Apparent diffusion coefficient (ADC) maps were normalized using N4 or N4/z-score. Nine machine-learning algorithms were trained for multiclass prediction of glioma types (IDH-mutant 1p/19q codeleted, IDH-mutant 1p/19q non-codeleted, IDH-wild type). Four approaches were compared: Anatomical, anatomicalâ +â ADC naive, anatomicalâ +â ADC N4, and anatomicalâ +â ADC N4/z-score. The University of California San Francisco (UCSF)-glioma dataset (nâ =â 409) was used for external validation. Results: Naïve-Bayes algorithms yielded overall the best performance on the internal test set. Adding ADC radiomics significantly improved AUC from 0.79 to 0.86 (Pâ =â .011) for the IDH-wild-type subgroup, but not for the other 2 glioma subgroups (Pâ >â .05). In the external UCSF dataset, the addition of ADC radiomics yielded a significantly higher AUC for the IDH-wild-type subgroup (Pâ ≤â .001): 0.80 (N4/WS anatomical alone), 0.81 (anatomicalâ +â ADC naive), 0.81 (anatomicalâ +â ADC N4), and 0.88 (anatomicalâ +â ADC N4/z-score) as well as for the IDH-mutant 1p/19q non-codeleted subgroup (Pâ <â .012 each). Conclusions: ADC radiomics can enhance the performance of conventional MRI-based radiomic models, particularly for IDH-wild-type glioma. The benefit of intensity normalization of ADC maps depends on the type and context of the used data.
ABSTRACT
PURPOSE: To propose a novel recursive partitioning analysis (RPA) classification model in patients with IDH-wildtype glioblastomas that incorporates the recently expanded conception of the extent of resection (EOR) in terms of both supramaximal and total resections. EXPERIMENTAL DESIGN: This multicenter cohort study included a developmental cohort of 622 patients with IDH-wildtype glioblastomas from a single institution (Severance Hospital) and validation cohorts of 536 patients from three institutions (Seoul National University Hospital, Asan Medical Center, and Heidelberg University Hospital). All patients completed standard treatment including concurrent chemoradiotherapy and underwent testing to determine their IDH mutation and MGMTp methylation status. EORs were categorized into either supramaximal, total, or non-total resections. A novel RPA model was then developed and compared to a previous RTOG RPA model. RESULTS: In the developmental cohort, the RPA model included age, MGMTp methylation status, KPS, and EOR. Younger patients with MGMTp methylation and supramaximal resections showed a more favorable prognosis (class I: median overall survival [OS] 57.3 months), while low-performing patients with non-total resections and without MGMTp methylation showed the worst prognosis (class IV: median OS 14.3 months). The prognostic significance of the RPA was subsequently confirmed in the validation cohorts, which revealed a greater separation between prognostic classes for all cohorts compared to the previous RTOG RPA model. CONCLUSIONS: The proposed RPA model highlights the impact of supramaximal versus total resections and incorporates clinical and molecular factors into survival stratification. The RPA model may improve the accuracy of assessing prognostic groups.
ABSTRACT
BACKGROUND: Quantitative and automated volumetric evaluation of early ischemic changes on non-contrast CT (NCCT) has recently been proposed as a new tool to improve prognostic performance in patients undergoing endovascular therapy (EVT) for acute ischemic stroke (AIS). We aimed to test its clinical value compared with the Alberta Stroke Program Early CT Score (ASPECTS) in a large single-institutional patient cohort. METHODS: A total of 1103 patients with AIS due to large vessel occlusion in the M1 or proximal M2 segments who underwent NCCT and EVT between January 2013 and November 2019 were retrospectively enrolled. Acute ischemic volumes (AIV) and ASPECTS were generated from the baseline NCCT through e-ASPECTS (Brainomix). Correlations were tested using Spearman's coefficient. The predictive capabilities of AIV for a favorable outcome (modified Rankin Scale score at 90 days ≤2) were tested using multivariable logistic regression as well as machine-learning models. Performance of the models was assessed using receiver operating characteristic (ROC) curves and differences were tested using DeLong's test. RESULTS: Patients with a favorable outcome had a significantly lower AIV (median 12.0 mL (IQR 5.7-21.7) vs 18.8 mL (IQR 9.4-33.9), p<0.001). AIV was highly correlated with ASPECTS (rho=0.78, p<0.001) and weakly correlated with the National Institutes of Health Stroke Scale score at baseline (rho=0.22, p<0.001), and was an independent predictor of an unfavorable clinical outcome (adjusted OR 0.97, 95% CI 0.96 to 0.98). No significant difference was found between machine-learning models using either AIV or ASPECTS or both metrics for predicting a good clinical outcome (p>0.05). CONCLUSION: AIV is an independent predictor of clinical outcome and presented a non-inferior performance compared with ASPECTS, without clear advantages for prognostic modelling.
Subject(s)
Brain Ischemia , Endovascular Procedures , Ischemic Stroke , Stroke , Humans , Brain Ischemia/diagnostic imaging , Brain Ischemia/surgery , Tomography, X-Ray Computed , Retrospective Studies , Ischemic Stroke/diagnostic imaging , Ischemic Stroke/surgery , Stroke/diagnostic imaging , Stroke/surgery , Treatment Outcome , ThrombectomyABSTRACT
Background: Pseudoprogression (PsPD) is a major diagnostic challenge in the follow-up of patients with glioblastoma (GB) after chemoradiotherapy (CRT). Conventional imaging signs and parameters derived from diffusion and perfusion-MRI have yet to prove their reliability in clinical practice for an accurate differential diagnosis. Here, we tested these parameters and combined them with radiomic features (RFs), clinical data, and MGMT promoter methylation status using machine- and deep-learning (DL) models to distinguish PsPD from Progressive disease. Methods: In a single-center analysis, 105 patients with GB who developed a suspected imaging PsPD in the first 7 months after standard CRT were identified retrospectively. Imaging data included standard MRI anatomical sequences, apparent diffusion coefficient (ADC), and normalized relative cerebral blood volume (nrCBV) maps. Median values (ADC, nrCBV) and RFs (all sequences) were calculated from DL-based tumor segmentations. Generalized linear models with LASSO feature-selection and DL models were built integrating clinical data, MGMT methylation status, median ADC and nrCBV values and RFs. Results: A model based on clinical data and MGMT methylation status yielded an areas under the receiver operating characteristic curve (AUC) = 0.69 (95% CI 0.55-0.83) for detecting PsPD, and the addition of median ADC and nrCBV values resulted in a nonsignificant increase in performance (AUC = 0.71, 95% CI 0.57-0.85, P = .416). Combining clinical/MGMT information with RFs derived from ADC, nrCBV, and from all available sequences both resulted in significantly (both P < .005) lower model performances, with AUC = 0.52 (0.38-0.66) and AUC = 0.54 (0.40-0.68), respectively. DL imaging models resulted in AUCs ≤ 0.56. Conclusion: Currently available imaging biomarkers could not reliably differentiate PsPD from true tumor progression in patients with glioblastoma; larger collaborative efforts are needed to build more reliable models.
ABSTRACT
BACKGROUND: To assess whether artificial intelligence (AI)-based decision support allows more reproducible and standardized assessment of treatment response on MRI in neuro-oncology as compared to manual 2-dimensional measurements of tumor burden using the Response Assessment in Neuro-Oncology (RANO) criteria. METHODS: A series of 30 patients (15 lower-grade gliomas, 15 glioblastoma) with availability of consecutive MRI scans was selected. The time to progression (TTP) on MRI was separately evaluated for each patient by 15 investigators over two rounds. In the first round the TTP was evaluated based on the RANO criteria, whereas in the second round the TTP was evaluated by incorporating additional information from AI-enhanced MRI sequences depicting the longitudinal changes in tumor volumes. The agreement of the TTP measurements between investigators was evaluated using concordance correlation coefficients (CCC) with confidence intervals (CI) and P-values obtained using bootstrap resampling. RESULTS: The CCC of TTP-measurements between investigators was 0.77 (95% CI = 0.69,0.88) with RANO alone and increased to 0.91 (95% CI = 0.82,0.95) with AI-based decision support (P = .005). This effect was significantly greater (P = .008) for patients with lower-grade gliomas (CCC = 0.70 [95% CI = 0.56,0.85] without vs. 0.90 [95% CI = 0.76,0.95] with AI-based decision support) as compared to glioblastoma (CCC = 0.83 [95% CI = 0.75,0.92] without vs. 0.86 [95% CI = 0.78,0.93] with AI-based decision support). Investigators with less years of experience judged the AI-based decision as more helpful (P = .02). CONCLUSIONS: AI-based decision support has the potential to yield more reproducible and standardized assessment of treatment response in neuro-oncology as compared to manual 2-dimensional measurements of tumor burden, particularly in patients with lower-grade gliomas. A fully-functional version of this AI-based processing pipeline is provided as open-source (https://github.com/NeuroAI-HD/HD-GLIO-XNAT).
Subject(s)
Brain Neoplasms , Glioblastoma , Glioma , Humans , Glioblastoma/pathology , Brain Neoplasms/diagnostic imaging , Brain Neoplasms/therapy , Brain Neoplasms/pathology , Artificial Intelligence , Reproducibility of Results , Glioma/diagnostic imaging , Glioma/therapy , Glioma/pathologyABSTRACT
Glioblastoma, the most common and aggressive primary brain tumor type, is considered an immunologically "cold" tumor with sparse infiltration by adaptive immune cells. Immunosuppressive tumor-associated myeloid cells are drivers of tumor progression. Therefore, targeting and reprogramming intratumoral myeloid cells is an appealing therapeutic strategy. Here, we investigate a ß-cyclodextrin nanoparticle (CDNP) formulation encapsulating the Toll-like receptor 7 and 8 (TLR7/8) agonist R848 (CDNP-R848) to reprogram myeloid cells in the glioma microenvironment. We show that intravenous monotherapy with CDNP-R848 induces regression of established syngeneic experimental glioma, resulting in increased survival rates compared with unloaded CDNP controls. Mechanistically, CDNP-R848 treatment reshapes the immunosuppressive tumor microenvironment and orchestrates tumor clearing by pro-inflammatory tumor-associated myeloid cells, independently of T cells and NK cells. Using serial magnetic resonance imaging, we identify a radiomic signature in response to CDNP-R848 treatment and ultrasmall superparamagnetic iron oxide (USPIO) imaging reveals that immunosuppressive macrophage recruitment is reduced by CDNP-R848. In conclusion, CDNP-R848 induces tumor regression in experimental glioma by targeting blood-borne macrophages without requiring adaptive immunity.
Subject(s)
Glioma , Nanoparticles , Toll-Like Receptor 7 , Toll-Like Receptor 8 , Humans , Adjuvants, Immunologic , Glioma/drug therapy , Macrophages , T-Lymphocytes , Toll-Like Receptor 7/agonists , Tumor Microenvironment , Toll-Like Receptor 8/agonistsABSTRACT
Swift diagnosis and treatment play a decisive role in the clinical outcome of patients with acute ischemic stroke (AIS), and computer-aided diagnosis (CAD) systems can accelerate the underlying diagnostic processes. Here, we developed an artificial neural network (ANN) which allows automated detection of abnormal vessel findings without any a-priori restrictions and in <2 minutes. Pseudo-prospective external validation was performed in consecutive patients with suspected AIS from 4 different hospitals during a 6-month timeframe and demonstrated high sensitivity (≥87%) and negative predictive value (≥93%). Benchmarking against two CE- and FDA-approved software solutions showed significantly higher performance for our ANN with improvements of 25-45% for sensitivity and 4-11% for NPV (p ≤ 0.003 each). We provide an imaging platform ( https://stroke.neuroAI-HD.org ) for online processing of medical imaging data with the developed ANN, including provisions for data crowdsourcing, which will allow continuous refinements and serve as a blueprint to build robust and generalizable AI algorithms.