RESUMO
PURPOSE: Skull base features, including increased foramen ovale (FO) cross-sectional area, are associated with lateral skull base spontaneous cerebrospinal fluid (sCSF) leak and encephalocele. Manual measurement requires skill in interpreting imaging studies and is time consuming. The goal of this study was to develop a fully automated deep learning method for FO segmentation and to determine the predictive value in identifying patients with sCSF leak or encephalocele. METHODS: A retrospective cohort study at a tertiary care academic hospital of 34 adults with lateral skull base sCSF leak or encephalocele were compared with 815 control patients from 2013-2021. A convolutional neural network (CNN) was constructed for image segmentation of axial computed tomography (CT) studies. Predicted FO segmentations were compared to manual segmentations, and receiver operating characteristic (ROC) curves were constructed. RESULTS: 295 CTs were used for training and validation of the CNN. A separate dataset of 554 control CTs was matched 5:1 on age and sex with the sCSF leak/encephalocele group. The mean Dice score was 0.81. The sCSF leak/encephalocele group had greater mean (SD) FO cross-sectional area compared to the control group, 29.0 (7.7) mm2 versus 24.3 (7.6) mm2 (P = .002, 95% confidence interval 0.02-0.08). The area under the ROC curve was 0.69. CONCLUSION: CNNs can be used to segment the cross-sectional area of the FO accurately and efficiently. Used together with other predictors, this method could be used as part of a clinical tool to predict the risk of sCSF leak or encephalocele.
RESUMO
Clinical monitoring of metastatic disease to the brain can be a laborious and timeconsuming process, especially in cases involving multiple metastases when the assessment is performed manually. The Response Assessment in Neuro-Oncology Brain Metastases (RANO-BM) guideline, which utilizes the unidimensional longest diameter, is commonly used in clinical and research settings to evaluate response to therapy in patients with brain metastases. However, accurate volumetric assessment of the lesion and surrounding peri-lesional edema holds significant importance in clinical decision-making and can greatly enhance outcome prediction. The unique challenge in performing segmentations of brain metastases lies in their common occurrence as small lesions. Detection and segmentation of lesions that are smaller than 10 mm in size has not demonstrated high accuracy in prior publications. The brain metastases challenge sets itself apart from previously conducted MICCAI challenges on glioma segmentation due to the significant variability in lesion size. Unlike gliomas, which tend to be larger on presentation scans, brain metastases exhibit a wide range of sizes and tend to include small lesions. We hope that the BraTS-METS dataset and challenge will advance the field of automated brain metastasis detection and segmentation.
RESUMO
Worldwide, there are nearly 10 million new cases of dementia annually, of which Alzheimer's disease (AD) is the most common. New measures are needed to improve the diagnosis of individuals with cognitive impairment due to various etiologies. Here, we report a deep learning framework that accomplishes multiple diagnostic steps in successive fashion to identify persons with normal cognition (NC), mild cognitive impairment (MCI), AD, and non-AD dementias (nADD). We demonstrate a range of models capable of accepting flexible combinations of routinely collected clinical information, including demographics, medical history, neuropsychological testing, neuroimaging, and functional assessments. We then show that these frameworks compare favorably with the diagnostic accuracy of practicing neurologists and neuroradiologists. Lastly, we apply interpretability methods in computer vision to show that disease-specific patterns detected by our models track distinct patterns of degenerative changes throughout the brain and correspond closely with the presence of neuropathological lesions on autopsy. Our work demonstrates methodologies for validating computational predictions with established standards of medical diagnosis.
Assuntos
Doença de Alzheimer , Disfunção Cognitiva , Aprendizado Profundo , Doença de Alzheimer/diagnóstico por imagem , Doença de Alzheimer/psicologia , Disfunção Cognitiva/diagnóstico , Disfunção Cognitiva/patologia , Progressão da Doença , Humanos , Neuroimagem/métodosRESUMO
Background: Acetabular fractures account for 10% of pelvis injuries, which are especially difficult to treat in developing countries with less access to resources. 3D printing has previously been shown to be a beneficial method of surgical planning, however the steep initial costs associated with purchasing a 3D printer may prevent some facilities form utilizing this technique. The purpose of this study was to develop 3D printed models for acetabular surgery using methodologies of varying cost to determine differences in model accuracy and overall quality. Methods: Five acetabular fracture models were developed from de-identified CT data using (I) proprietary and open-source segmentation software and (II) fused deposition modeling (FDM) and stereolithography (SLA) 3D printing methods. The distance between the posterior inferior iliac spine (PIIS) and the ischial spine as well as a unique fracture fragment for each model was compared between the different printing methodologies. The models were then given to 5 physicians and assessed on their overall accuracy compared to traditional 2D images. Results: Printing methodology did not affect the distance from PIIS to ischial spine (P=0.263). However, fracture fragment representation differed across 3D printed models, with the most accurate model produced by the high-end resin-based printer (P=0.007). The survey analysis showed that the low-cost printing methods produced models that were not as accurate in their representation of the fractured region (P=0.008). Conclusions: The differences between models developed using traditional methods and low-cost methods have slight differences but may still provide useful information when developing a surgical plan.
RESUMO
The purpose of this study was to investigate the usage of an anatomical model to improve surgical planning of a complex schwannoma resection. As advancements in additive manufacturing continue to prosper, new applications of this valuable technology are being implemented in the medical field. One of the most recent applications has been in the development of patient-specific anatomical models for unique clinical education as well as for preoperative planning. In this case, a multidisciplinary team with expertise in research, three-dimensional printing, and medicine was formed to develop a three-dimensional printed model that could be used to help plan the reduction of a tumor from the cervical spine of a pediatric patient. Image segmentation and stereolithography creation were accomplished using Mimics and 3-matic, respectively. Models were developed on two different printer types to view different aspects of the region of interest. Reports from the operating surgeon indicated that the model was instrumental in the planning procedures of the operation and reducing operation time.
RESUMO
OBJECTIVE. Primary CNS posttransplant lymphoproliferative disorder (PTLD) may present as multiple contrast-enhancing intraaxial lesions, often with central necrosis and surrounding edema. This imaging appearance is similar to the pattern seen in brain metastases. The purpose of this study was to find differences in the radiologic features of primary CNS PTLD lesions and brain metastases. MATERIALS AND METHODS. We retrospectively reviewed the radiologic findings of 51 primary CNS PTLD lesions in 10 patients and 156 metastatic brain lesions in 25 patients. Lesion size, multifocality, location, necrosis, hemorrhage, perilesional vasogenic edema, contrast enhancement, and diffusion and perfusion features were evaluated. We used the chi-square test or Fisher exact test when appropriate to compare the findings between primary CNS PTLD lesions and brain metastases. RESULTS. Primary CNS PTLD lesions occur in the deep gray matter and periventricular locations more frequently than brain metastases (p < 0.0001) and are not present at the gray and white matter junctions and vascular border zones as commonly as brain metastases are (p < 0.0001). Primary CNS PTLD tends to have less frequent hemorrhage (p < 0.0001), more restricted diffusion (p = 0.001), and lower perfusion (p = 0.002) than brain metastases. We did not find statistically significant differences between primary CNS PTLD and brain metastases for lesion size, multifocality, necrosis, and perilesional edema. CONCLUSION. The imaging characteristics of primary CNS PTLD overlap considerably with those of brain metastases, but there are significant differences between primary CNS PTLD lesions and brain metastases in lesion location, diffusion and perfusion features, and tendency to hemorrhage.
Assuntos
Encefalopatias/diagnóstico por imagem , Neoplasias Encefálicas/diagnóstico por imagem , Neoplasias Encefálicas/secundário , Transtornos Linfoproliferativos/diagnóstico por imagem , Neuroimagem/métodos , Transplante de Órgãos , Complicações Pós-Operatórias/diagnóstico por imagem , Adolescente , Adulto , Idoso , Criança , Pré-Escolar , Meios de Contraste , Diagnóstico Diferencial , Feminino , Humanos , Lactente , Imageamento por Ressonância Magnética/métodos , Masculino , Pessoa de Meia-Idade , Complicações Pós-Operatórias/patologia , Estudos RetrospectivosRESUMO
PURPOSE: Hydrocephalus is a clinically significant condition which can have devastating consequences if left untreated. Currently available methods for quantifying this condition using CT imaging are unreliable and prone to error. The purpose of this study is to investigate the clinical utility of using convolutional neural networks to calculate ventricular volume and explore limitations. METHODS: A two-dimensional convolutional neural network was designed to perform fully automated ventricular segmentation on CT images. A total of 300 head CTs were collected and used in this exploration. Two hundred were used to train the network, 50 were used for validation, and 50 were used for testing. RESULTS: Dice scores for the left lateral, right lateral, and third ventricle segmentations were 0.92, 0.92, and 0.79, respectively; the coefficients of determination were r2 = 0.991, r2 = 0.994, and r2 = 0.976; the average volume differences between manual and automated segmentation were 0.821 ml, 0.587 ml, and 0.099 ml. CONCLUSION: Two-dimensional convolutional neural network architectures can be used to accurately segment and quantify intracranial ventricle volume. While further refinements are necessary, it is likely these networks could be used as a clinical tool to quantify hydrocephalus accurately and efficiently.
Assuntos
Ventrículos Cerebrais/diagnóstico por imagem , Redes Neurais de Computação , Tomografia Computadorizada por Raios X/métodos , Humanos , Tamanho do Órgão , Estudos RetrospectivosRESUMO
Spine anatomy can be difficult to master and is essential for performing spine procedures. We sought to utilize the rapidly expanding field of 3D technology to create freely available, interactive educational materials for spine procedures. Our secondary goal was to convey lessons learned about 3D modeling and printing. This project involved two parallel processes: the creation of 3D-printed physical models and interactive digital models. We segmented illustrative CT studies of the lumbar and cervical spine to create 3D models and then printed them using a consumer 3D printer and a professional 3D printing service. We also included downloadable versions of the models in an interactive eBook and platform-independent web viewer. We then provided these educational materials to residents with a pretest and posttest to assess efficacy. The "Spine Procedures in 3D" eBook has been downloaded 71 times as of October 5, 2016. All models used in the book are available for download and printing. Regarding test results, the mean exam score improved from 70 to 86%, with the most dramatic improvement seen in the least experienced trainees. Participants reported increased confidence in performing lumbar punctures after exposure to the material. We demonstrate the value of 3D models, both digital and printed, in learning spine procedures. Moreover, 3D printing and modeling is a rapidly expanding field with a large potential role for radiologists. We have detailed our process for creating and sharing 3D educational materials in the hopes of motivating and enabling similar projects.
Assuntos
Recursos Audiovisuais , Imageamento Tridimensional , Modelos Anatômicos , Impressão Tridimensional , Coluna Vertebral/anatomia & histologia , Vértebras Cervicais/anatomia & histologia , Vértebras Cervicais/diagnóstico por imagem , Humanos , Vértebras Lombares/anatomia & histologia , Vértebras Lombares/diagnóstico por imagem , Software , Punção Espinal , Coluna Vertebral/diagnóstico por imagemRESUMO
OBJECTIVES: Application of fMRI connectivity metrics as diagnostic biomarkers at the individual level will require reliability, sensitivity and specificity to longitudinal changes in development, aging, neurocognitive, and behavioral performance and pathologies. Such metrics have not been well characterized for recent advances in BOLD acquisition. EXPERIMENTAL DESIGN: Analysis of multiband BOLD data from the HCP 500 Subjects Release was performed with FIX ICA and with WM, CSF and motion parameter regression. Analysis with ROIs covering the gray matter at 5 mm resolution was performed to assess functional connectivity. ROIs in key areas were used to demonstrate statistical differences between specific connections. Reproducibility of group-mean functional connectivity and for single connections for individuals was evaluated for both resting state and task acquisitions. PRINCIPAL OBSERVATIONS: Systematic differences in group-mean connectivity were demonstrated during task and rest and during different tasks, although individual differences in connectivity were maintained. Reproducibility of a single connection for a subject and across subjects for resting and task acquisition was demonstrated to be a linear function of the square root of imaging time. Randomly removing up to 50% of time points had little effect on reliability, while truncating an acquisition was associated with decreased reliability. Reliability was highest within the cortex, and lowest for deep gray nuclei, gray-white junction, and near large sulci. CONCLUSIONS: This study found systematic differences in group-mean connectivity acquired during task and rest acquitisions and preserved individual differences in connectivity due to intrinsic differences in an individual's brain activity and structural brain architecture. We also show that longer scan times are needed to acquire data on single subjects for information on connections between specific ROIs. Longer scans may be facilitated by acquisition during task paradigms, which will systematically affect functional connectivity but may preserve individual differences in connectivity on top of task modulations.