Dynamic Contrast-enhanced Time Series Research Articles

Model-free prostate cancer segmentation from dynamic contrast-enhanced MRI with recurrent convolutional networks: A feasibility study

Dynamic contrast enhanced (DCE) magnetic resonance imaging (MRI) is a method of temporal imaging that is commonly used to aid in prostate cancer (PCa) diagnosis and staging. Typically, machine learning models designed for the segmentation and detection of PCa will use an engineered scalar image called Ktrans to summarize the information in the DCE time-series images. This work proposes a new model that amalgamates the U-net and the convGRU neural network architectures for the purpose of interpreting DCE time-series in a temporal and spatial basis for segmenting PCa in MR images. Ultimately, experiments show that the proposed model using the DCE time-series images can outperform a baseline U-net segmentation model using Ktrans. However, when other types of scalar MR images are considered by the models, no significant advantage is observed for the proposed model.

A dual-purpose MRI acquisition to combine 4D-MRI and dynamic contrast-enhanced imaging for abdominal radiotherapy planning

For successful abdominal radiotherapy it is crucial to have a clear tumor definition and an accurate characterization of the motion. While dynamic contrast-enhanced (DCE) MRI aids tumor visualization, it is often hampered by motion artifacts. 4D-MRI characterizes this motion, but often lacks the contrast to clearly visualize the tumor. This dual requirement is challenging due to time constraints. Here, we propose combining both into a single acquisition by reconstructing the data in various ways in order to achieve both high spatio-temporal resolution DCE-MRI and accurate 4D-MRI motion estimates.A 5 min T1-weigthed DCE acquisition was collected in five renal-cell carcinoma patients and simulated in a digital phantom. Data were acquired continuously using a 3D golden angle radial stack-of-stars acquisition. This enabled three types of reconstruction; (1) a high spatio-temporal resolution DCE time series, (2) a 5D reconstruction and (3) a contrast-enhanced 4D-MRI for motion characterization. Motion extracted from the 4D- and 5D-MRI was compared with a separately acquired 4D-MRI and additional 2D cine MR imaging.Simulations on XCAT showed that 5D reconstructions severely underestimated motion (), whereas contrast-enhanced 4D-MRI only underestimated motion by . This was confirmed in the in vivo data where motion of the contrast-enhanced 4D-MRI was approximately smaller than the motion in the 2D cine MRI (5.8 mm versus 6.5 mm), but equal to a separately acquired 4D-MRI (5.8 mm versus 5.9 mm). 5D reconstructions underestimated the motion by more than , but minimized respiratory-induced blurring in the contrast enhanced images. DCE time-series demonstrated clear tumor visualization and contrast enhancement throughout the entire field-of-view.DCE- and 4D-MRI can be integrated into a single acquisition that enables different reconstructions with complementary information for abdominal radiotherapy planning and, in an MRI-guided treatment, precise motion information, input for motion models, and rapid feedback on the contrast enhancement.

In Vivo DCE-MRI for the Discrimination Between Glioblastoma and Radiation Necrosis in Rats

In this study, the potential of semiquantitative and quantitative analysis of dynamic contrast-enhanced (DCE) magnetic resonance imaging (MRI) was investigated to differentiate glioblastoma (GB) from radiation necrosis (RN) in rats. F98 GB growth was seen on MRI 8-23days post-inoculation (n=15). RN lesions developed 6-8months post-irradiation (n=10). DCE-MRI was acquired using a fast low-angle shot (FLASH) sequence. Regions of interest (ROIs) encompassed peripheral contrast enhancement in GB (n=15) and RN (n=10) as well as central necrosis within these lesions (GB (n=4), RN (n=3)). Dynamic contrast-enhanced time series, obtained from the DCE-MRI data, were fitted to determine four function variables (amplitude A, offset from zero C, wash-in rate k, and wash-out rate D) as well as maximal intensity (ImaxF) and time to peak (TTPF). Secondly, maps of semiquantitative and quantitative parameters (extended Tofts model) were created using Olea Sphere (O). Semiquantitative DCE-MRI parameters included wash-inO, wash-outO, area under the curve (AUCO), maximal intensity (ImaxO), and time to peak (TTPO). Quantitative parameters included the rate constant plasma to extravascular-extracellular space (EES) (K trans), the rate constant EES to plasma (K ep), plasma volume (V p), and EES volume (V e). All (semi)quantitative parameters were compared between GB and RN using the Mann-Whitney U test. ROC analysis was performed. Wash-in rate (k) and wash-out rate (D) were significantly higher in GB compared to RN using curve fitting (p=0.016 and p=0.014). TTPF and TTPO were significantly lower in GB compared to RN (p=0.001 and p=0.005, respectively). The highest sensitivity (87%) and specificity (80%) were obtained for TTPF by applying a threshold of 581s. K trans, K ep, and V e were not significantly different between GB and RN. A trend towards higher V p values was found in GB compared to RN, indicating angiogenesis in GB (p=0.075). Based on our results, in a rat model of GB and RN, wash-in rate, wash-out rate, and the time to peak extracted from DCE-MRI time series data may be useful to discriminate GB from RN.

Spectral embedding based active contour (SEAC) for lesion segmentation on breast dynamic contrast enhanced magnetic resonance imaging

Segmentation of breast lesions on dynamic contrast enhanced (DCE) magnetic resonance imaging (MRI) is the first step in lesion diagnosis in a computer-aided diagnosis framework. Because manual segmentation of such lesions is both time consuming and highly susceptible to human error and issues of reproducibility, an automated lesion segmentation method is highly desirable. Traditional automated image segmentation methods such as boundary-based active contour (AC) models require a strong gradient at the lesion boundary. Even when region-based terms are introduced to an AC model, grayscale image intensities often do not allow for clear definition of foreground and background region statistics. Thus, there is a need to find alternative image representations that might provide (1) strong gradients at the margin of the object of interest (OOI); and (2) larger separation between intensity distributions and region statistics for the foreground and background, which are necessary to halt evolution of the AC model upon reaching the border of the OOI. In this paper, the authors introduce a spectral embedding (SE) based AC (SEAC) for lesion segmentation on breast DCE-MRI. SE, a nonlinear dimensionality reduction scheme, is applied to the DCE time series in a voxelwise fashion to reduce several time point images to a single parametric image where every voxel is characterized by the three dominant eigenvectors. This parametric eigenvector image (PrEIm) representation allows for better capture of image region statistics and stronger gradients for use with a hybrid AC model, which is driven by both boundary and region information. They compare SEAC to ACs that employ fuzzy c-means (FCM) and principal component analysis (PCA) as alternative image representations. Segmentation performance was evaluated by boundary and region metrics as well as comparing lesion classification using morphological features from SEAC, PCA+AC, and FCM+AC. On a cohort of 50 breast DCE-MRI studies, PrEIm yielded overall better region and boundary-based statistics compared to the original DCE-MR image, FCM, and PCA based image representations. Additionally, SEAC outperformed a hybrid AC applied to both PCA and FCM image representations. Mean dice similarity coefficient (DSC) for SEAC was significantly better (DSC = 0.74 ± 0.21) than FCM+AC (DSC = 0.50 ± 0.32) and similar to PCA+AC (DSC = 0.73 ± 0.22). Boundary-based metrics of mean absolute difference and Hausdorff distance followed the same trends. Of the automated segmentation methods, breast lesion classification based on morphologic features derived from SEAC segmentation using a support vector machine classifier also performed better (AUC = 0.67 ± 0.05; p < 0.05) than FCM+AC (AUC = 0.50 ± 0.07), and PCA+AC (AUC = 0.49 ± 0.07). In this work, we presented SEAC, an accurate, general purpose AC segmentation tool that could be applied to any imaging domain that employs time series data. SE allows for projection of time series data into a PrEIm representation so that every voxel is characterized by the dominant eigenvectors, capturing the global and local time-intensity curve similarities in the data. This PrEIm allows for the calculation of strong tensor gradients and better region statistics than the original image intensities or alternative image representations such as PCA and FCM. The PrEIm also allows for building a more accurate hybrid AC scheme.