RESUMO
Standard of care diagnostic procedure for suspected skin cancer is microscopic examination of hematoxylin & eosin stained tissue by a pathologist. Areas of high inter-pathologist discordance and rising biopsy rates necessitate higher efficiency and diagnostic reproducibility. We present and validate a deep learning system which classifies digitized dermatopathology slides into 4 categories. The system is developed using 5,070 images from a single lab, and tested on an uncurated set of 13,537 images from 3 test labs, using whole slide scanners manufactured by 3 different vendors. The system's use of deep-learning-based confidence scoring as a criterion to consider the result as accurate yields an accuracy of up to 98%, and makes it adoptable in a real-world setting. Without confidence scoring, the system achieved an accuracy of 78%. We anticipate that our deep learning system will serve as a foundation enabling faster diagnosis of skin cancer, identification of cases for specialist review, and targeted diagnostic classifications.
Assuntos
Processamento de Imagem Assistida por Computador/métodos , Patologia/métodos , Reconhecimento Automatizado de Padrão , Neoplasias Cutâneas/diagnóstico por imagem , Algoritmos , Calibragem , Proliferação de Células , Simulação por Computador , Aprendizado Profundo , Humanos , Interpretação de Imagem Assistida por Computador/métodos , Melanócitos/citologia , Redes Neurais de Computação , Estudos Prospectivos , Curva ROC , Reprodutibilidade dos Testes , Carga de TrabalhoRESUMO
PURPOSE: To obviate online slice-by-slice RF shim optimization and reduce B1+ mapping requirements for patient-specific RF shimming in high-field magnetic resonance imaging. THEORY AND METHODS: RF Shim Prediction by Iteratively Projected Ridge Regression (PIPRR) predicts patient-specific, SAR-efficient RF shims with a machine learning approach that merges learning with training shim design. To evaluate it, a set of B1+ maps was simulated for 100 human heads for a 24-element coil at 7T. Features were derived from tissue masks and the DC Fourier coefficients of the coils' B1+ maps in each slice, which were used for kernelized ridge regression prediction of SAR-efficient RF shim weights. Predicted shims were compared to directly designed shims, circularly polarized mode, and nearest-neighbor shims predicted using the same features. RESULTS: PIPRR predictions had 87% and 13% lower B1+ coefficients of variation compared to circularly polarized mode and nearest-neighbor shims, respectively, and achieved homogeneity and SAR similar to that of directly designed shims. Predictions were calculated in 4.92 ms on average. CONCLUSION: PIPRR predicted uniform, SAR-efficient RF shims, and could save a large amount of B1+ mapping and computation time in RF-shimmed ultra-high field magnetic resonance imaging.
Assuntos
Processamento de Imagem Assistida por Computador/métodos , Imageamento por Ressonância Magnética/métodos , Aprendizado de Máquina Supervisionado , Algoritmos , Encéfalo/diagnóstico por imagem , Simulação por Computador , Bases de Dados Factuais , Cabeça/diagnóstico por imagem , Humanos , Imagens de FantasmasRESUMO
PURPOSE: To correct line-to-line delays and phase errors in echo-planar imaging (EPI). THEORY AND METHODS: EPI-trajectory auto-corrected image reconstruction (EPI-TrACR) is an iterative maximum-likelihood technique that exploits data redundancy provided by multiple receive coils between nearby lines of k-space to determine and correct line-to-line trajectory delays and phase errors that cause ghosting artifacts. EPI-TrACR was efficiently implemented using a segmented FFT and was applied to in vivo brain data acquired at 7 T across acceleration (1×-4×) and multishot factors (1-4 shots), and in a time series. RESULTS: EPI-TrACR reduced ghosting across all acceleration factors and multishot factors, compared to conventional calibrated reconstructions and the PAGE method. It also achieved consistently lower ghosting in the time series. Averaged over all cases, EPI-TrACR reduced root-mean-square ghosted signal outside the brain by 27% compared to calibrated reconstruction, and by 40% compared to PAGE. CONCLUSION: EPI-TrACR automatically corrects line-to-line delays and phase errors in multishot, accelerated, and dynamic EPI. While the method benefits from additional calibration data for initialization, it was not a requirement for most reconstructions. Magn Reson Med 79:3114-3121, 2018. © 2017 International Society for Magnetic Resonance in Medicine.
Assuntos
Imagem Ecoplanar/métodos , Processamento de Imagem Assistida por Computador/métodos , Algoritmos , Bases de Dados Factuais , Humanos , Imagens de FantasmasRESUMO
PURPOSE: To estimate k-space trajectory errors in non-Cartesian acquisitions and reconstruct distortion-free images, without trajectory measurements or gradient calibrations. THEORY AND METHODS: The Trajectory Auto-Corrected image Reconstruction method jointly estimates k-space trajectory errors and images, based on SENSE and SPIRiT parallel imaging reconstruction. The underlying idea is that parallel imaging and oversampling in the center of k-space provides data redundancy that can be exploited to simultaneously reconstruct images and correct trajectory errors. Trajectory errors are represented as weighted sums of trajectory-dependent error basis functions, the coefficients of which are estimated using gradient-based optimization. RESULTS: Trajectory Auto-Corrected image Reconstruction was applied to reconstruct images and errors in golden angle radial, center-out radial, and spiral in vivo 7 Tesla brain acquisitions in five subjects. Compared to reconstructions using nominal trajectories, Trajectory auto-corrected image reconstructions contained considerably less blurring and streaking and were of similar quality to images reconstructed using measured k-space trajectories in the center-out radial and spiral cases. Reconstruction cost function reductions and improvements in normalized image gradient squared were also similar to those for images reconstructed using measured trajectories. CONCLUSION: Trajectory Auto-Corrected image Reconstruction enables non-Cartesian image reconstructions free from trajectory errors without the need for separate gradient calibrations or trajectory measurements. Magn Reson Med 76:757-768, 2016. © 2015 Wiley Periodicals, Inc.
Assuntos
Algoritmos , Artefatos , Encéfalo/anatomia & histologia , Aumento da Imagem/métodos , Interpretação de Imagem Assistida por Computador/métodos , Imageamento por Ressonância Magnética/métodos , Reconhecimento Automatizado de Padrão/métodos , Humanos , Reprodutibilidade dos Testes , Sensibilidade e EspecificidadeRESUMO
The temporal relationship between different stages of cognitive processing is long debated. This debate is ongoing, primarily because it is often difficult to measure the time course of multiple cognitive processes simultaneously. We employed a manipulation that allowed us to isolate ERP components related to perceptual processing, working memory, and response preparation, and then examined the temporal relationship between these components while observers performed a visual search task. We found that, when response speed and accuracy were equally stressed, our index of perceptual processing ended before both the transfer of information into working memory and response preparation began. However, when we stressed speed over accuracy, response preparation began before the completion of perceptual processing or transfer of information into working memory on trials with the fastest reaction times. These findings show that individuals can control the flow of information transmission between stages, either waiting for perceptual processing to be completed before preparing a response or configuring these stages to overlap in time.