RESUMO
Segmenting deep brain structures from magnetic resonance images is important for patient diagnosis, surgical planning, and research. Most current state-of-the-art solutions follow a segmentation-by-registration approach, where subject magnetic resonance imaging (MRIs) are mapped to a template with well-defined segmentations. However, registration-based pipelines are time-consuming, thus, limiting their clinical use. This paper uses deep learning to provide a one-step, robust, and efficient deep brain segmentation solution directly in the native space. The method consists of a preprocessing step to conform all MRI images to the same orientation, followed by a convolutional neural network using the nnU-Net framework. We use a total of 14 datasets from both research and clinical collections. Of these, seven were used for training and validation and seven were retained for testing. We trained the network to segment 30 deep brain structures, as well as a brain mask, using labels generated from a registration-based approach. We evaluated the generalizability of the network by performing a leave-one-dataset-out cross-validation, and independent testing on unseen datasets. Furthermore, we assessed cross-domain transportability by evaluating the results separately on different domains. We achieved an average dice score similarity of 0.89 ± 0.04 on the test datasets when compared to the registration-based gold standard. On our test system, the computation time decreased from 43 min for a reference registration-based pipeline to 1.3 min. Our proposed method is fast, robust, and generalizes with high reliability. It can be extended to the segmentation of other brain structures. It is publicly available on GitHub, and as a pip package for convenient usage.
Assuntos
Encéfalo , Processamento de Imagem Assistida por Computador , Humanos , Processamento de Imagem Assistida por Computador/métodos , Reprodutibilidade dos Testes , Encéfalo/diagnóstico por imagem , Redes Neurais de Computação , Imageamento por Ressonância Magnética/métodosRESUMO
Deep brain stimulation (DBS) is a surgical therapy to alleviate symptoms of certain brain disorders by electrically modulating neural tissues. Computational models predicting electric fields and volumes of tissue activated are key for efficient parameter tuning and network analysis. Currently, we lack efficient and flexible software implementations supporting complex electrode geometries and stimulation settings. Available tools are either too slow (e.g. finite element method-FEM), or too simple, with limited applicability to basic use-cases. This paper introduces FastField, an efficient open-source toolbox for DBS electric field and VTA approximations. It computes scalable electric field approximations based on the principle of superposition, and VTA activation models from pulse width and axon diameter. In benchmarks and case studies, FastField is solved in about 0.2 s, ~ 1000 times faster than using FEM. Moreover, it is almost as accurate as using FEM: average Dice overlap of 92%, which is around typical noise levels found in clinical data. Hence, FastField has the potential to foster efficient optimization studies and to support clinical applications.
Assuntos
Encéfalo/fisiologia , Estimulação Encefálica Profunda , Fenômenos Eletromagnéticos , Axônios/fisiologia , Estimulação Encefálica Profunda/instrumentação , Eletrodos Implantados , Fenômenos Eletrofisiológicos , Humanos , Modelos Neurológicos , SoftwareRESUMO
Anorexia nervosa (AN) severely impacts individual's mental and physical health as well as quality of life. In 21% of cases no durable response to conservative treatment can be obtained. The serious course of the disease in the most severely affected patients justifies invasive treatment options. One of the treatment methods increasingly used in recent years is deep brain stimulation (DBS). A 42-year-old woman suffering from chronic AN of the bulimic subtype shows a 46.9% weight gain and a subjective increase in quality of life, 12 months after bilateral nucleus accumbens (NAcc) DBS implantation. No improvement in comorbid depression could be achieved. DBS of the NAcc is a treatment option to be considered in severe AN when conventional treatment modalities recommended by evidence-based guidelines have not been able to bring lasting relief to the patient's suffering.