Accurate and non-invasive monitoring of hematopoietic stem cell transplantation (HSCT) patients is crucial but challenging due to factors including the limitations of traditional biopsies and the intensive manual analysis required for emerging comprehensive imaging techniques like FLT PET/CT. Furthermore, low-dose CT resolution in this vulnerable patient population often hinders precise vertebral body segmentation, a critical step for comprehensive marrow compartment assessment. This paper presents a novel approach for per-vertebral body instance segmentation in HSCT FLT PET/CT scans, addressing the inherent difficulties of low-resolution data and limited annotated training cases. Our method leverages an attention-gated U-Net architecture, significantly enhanced by a novel data augmentation strategy involving downsampled high-resolution VerSe dataset images. We demonstrate, for the first time, accurate vertebral body segmentation on this challenging low-resolution dataset. Our approach integrates an attention-based U-Net model and is compared against TotalSegmentator as a baseline, showing superior segmentation performance, particularly in the anatomically complex upper spine where TotalSegmentator exhibits suboptimal results. To the best of our knowledge, this work reports fully automated high-quality instance segmentation results for individual vertebral bodies in CT volumes of HSCT FLT PET/CT patients for the first time, promising to facilitate automation for critical quantitative assessments like SUV measurement and ultimately improve long-term patient management and outcomes.

We introduce a new, fully automatic vertebral instance segmentation method to facilitate the extraction of standard uptake values (SUV) from the medullary cavities of individual vertebral bodies in joint 18 F-fluorothymidine (FLT) PET/CT scans acquired from hematopoietic stem cell transplantation (HSCT) patients at 28 days after transplant. Due to dosing considerations, the CT voxels in these scans are characterized by a large 5 mm axial slice thickness which significantly complicates the vertebral body segmentation problem. The key ideas of our method are to first apply an ensemble of U-Nets to obtain a binary mask for the aggregated collection of vertebral bodies as a single object without estimating the intervertebral boundaries, and then leverage the relatively better 4 mm axial slice spacing in the PET data to estimate a “best fit” axial coordinate to approximate the break between each pair of vertebrae. This PET-CT fusion approach results in an approximate vertebral body segmentation where each estimated intervertebral boundary is, by construction, restricted to lie in a single axial plane. However, because the FLT uptake is well localized within the medullary cavities, our results show that this approximate segmentation is sufficiently accurate to enable FLT SUV data to be isolated for individual vertebral bodies. Compared to traditional methods for assessing engraftment based on single aspirate biopsies, this new technique has potential to facilitate a significantly more comprehensive assessment of the medullary compartment by providing fully automated SUV data for a plurality of individual bones.

OBJECTIVE: To demonstrate enabling multi-institutional training without centralizing or sharing the underlying physical data via federated learning (FL). MATERIALS AND METHODS: Deep learning models were trained at each participating institution using local clinical data, and an additional model was trained using FL across all of the institutions. RESULTS: We found that the FL model exhibited superior performance and generalizability to the models trained at single institutions, with an overall performance level that was significantly better than that of any of the institutional models alone when evaluated on held-out test sets from each institution and an outside challenge dataset. DISCUSSION: The power of FL was successfully demonstrated across 3 academic institutions while avoiding the privacy risk associated with the transfer and pooling of patient data. CONCLUSION: Federated learning is an effective methodology that merits further study to enable accelerated development of models across institutions, enabling greater generalizability in clinical use.

Multi-domain data are widely leveraged in vision applications taking advantage of complementary information from different modalities, e.g., brain tumor segmentation from multi-parametric magnetic resonance imaging (MRI). However, due to possible data corruption and different imaging protocols, the availability of images for each domain could vary amongst multiple data sources in practice, which makes it challenging to build a universal model with a varied set of input data. To tackle this problem, we propose a general approach to complete the random missing domain(s) data in real applications. Specifically, we develop a novel multi-domain image completion method that utilizes a generative adversarial network (GAN) with a representational disentanglement scheme to extract shared content encoding and separate style encoding across multiple domains. We further illustrate that the learned representation in multi-domain image completion could be leveraged for high-level tasks, e.g., segmentation, by introducing a unified framework consisting of image completion and segmentation with a shared content encoder. The experiments demonstrate consistent performance improvement on three datasets for brain tumor segmentation, prostate segmentation, and facial expression image completion respectively.

Temporal enhanced ultrasound (TeUS), comprising the analysis of variations in backscattered signals from a tissue over a sequence of ultrasound frames, has been previously proposed as a new paradigm for tissue characterization. In this paper, we propose to use deep recurrent neural networks (RNN) to explicitly model the temporal information in TeUS. By investigating several RNN models, we demonstrate that long short-term memory (LSTM) networks achieve the highest accuracy in separating cancer from benign tissue in the prostate. We also present algorithms for in-depth analysis of LSTM networks. Our in vivo study includes data from 255 prostate biopsy cores of 157 patients. We achieve area under the curve, sensitivity, specificity, and accuracy of 0.96, 0.76, 0.98, and 0.93, respectively. Our result suggests that temporal modeling of TeUS using RNN can significantly improve cancer detection accuracy over previously presented works.

Prostate segmentation from Magnetic Resonance (MR) images plays an important role in image guided intervention. However, the lack of clear boundary specifically at the apex and base, and huge variation of shape and texture between the images from different patients make the task very challenging. To overcome these problems, in this paper, we propose a deeply supervised convolutional neural network (CNN) utilizing the convolutional information to accurately segment the prostate from MR images. The proposed model can effectively detect the prostate region with additional deeply supervised layers compared with other approaches. Since some information will be abandoned after convolution, it is necessary to pass the features extracted from early stages to later stages. The experimental results show that significant segmentation accuracy improvement has been achieved by our proposed method compared to other reported approaches.

Clinical assessment of bone marrow is limited by an inability to evaluate the marrow space comprehensively and dynamically and there is no current method for automatically assessing hematopoietic activity within the medullary space. Evaluating the hematopoietic space in its entirety could be applicable in blood disorders, malignancies, infections, and medication toxicity. In this paper, we introduce a CT/PET 3D automatic framework for measurement of the hematopoietic compartment proliferation within osseous sites. We first perform a full-body bone structure segmentation using 3D graph-cut on the CT volume. The vertebrae are segmented by detecting the discs between adjacent vertebrae. Finally, we register the bone marrow CT volume with its corresponding PET volume and capture the spinal bone marrow volume. The proposed framework was tested on 17 patients, achieving an average accuracy of 86.37% and a worst case accuracy of 82.3% in automatically extracting the aggregate volume of the spinal marrow cavities.

Multiatlas based method is commonly used in medical image segmentation. In multiatlas based image segmentation, atlas selection and combination are considered as two key factors affecting the performance. Recently, manifold learning based atlas selection methods have emerged as very promising methods. However, due to the complexity of prostate structures in raw images, it is difficult to get accurate atlas selection results by only measuring the distance between raw images on the manifolds. Although the distance between the regions to be segmented across images can be readily obtained by the label images, it is infeasible to directly compute the distance between the test image (gray) and the label images (binary). This paper tries to address this problem by proposing a label image constrained atlas selection method, which exploits the label images to constrain the manifold projection of raw images. Analyzing the data point distribution of the selected atlases in the manifold subspace, a novel weight computation method for atlas combination is proposed. Compared with other related existing methods, the experimental results on prostate segmentation from T2w MRI showed that the selected atlases are closer to the target structure and more accurate segmentation were obtained by using our proposed method.

SUMMARY: In vivo dynamic contrast-enhanced imaging tools provide non-invasive methods for analyzing various functional changes associated with disease initiation, progression and responses to therapy. The quantitative application of these tools has been hindered by its inability to accurately resolve and characterize targeted tissues due to spatially mixed tissue heterogeneity. Convex Analysis of Mixtures - Compartment Modeling (CAM-CM) signal deconvolution tool has been developed to automatically identify pure-volume pixels located at the corners of the clustered pixel time series scatter simplex and subsequently estimate tissue-specific pharmacokinetic parameters. CAM-CM can dissect complex tissues into regions with differential tracer kinetics at pixel-wise resolution and provide a systems biology tool for defining imaging signatures predictive of phenotypes. AVAILABILITY: The MATLAB source code can be downloaded at the authors' website www.cbil.ece.vt.edu/software.htm CONTACT: [email protected] SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) provides a noninvasive method for evaluating tumor vasculature patterns based on contrast accumulation and washout. However, due to limited imaging resolution and tumor tissue heterogeneity, tracer concentrations at many pixels often represent a mixture of more than one distinct compartment. This pixel-wise partial volume effect (PVE) would have profound impact on the accuracy of pharmacokinetics studies using existing compartmental modeling (CM) methods. We, therefore, propose a convex analysis of mixtures (CAM) algorithm to explicitly mitigate PVE by expressing the kinetics in each pixel as a nonnegative combination of underlying compartments and subsequently identifying pure volume pixels at the corners of the clustered pixel time series scatter plot simplex. The algorithm is supported theoretically by a well-grounded mathematical framework and practically by plug-in noise filtering and normalization preprocessing. We demonstrate the principle and feasibility of the CAM-CM approach on realistic synthetic data involving two functional tissue compartments, and compare the accuracy of parameter estimates obtained with and without PVE elimination using CAM or other relevant techniques. Experimental results show that CAM-CM achieves a significant improvement in the accuracy of kinetic parameter estimation. We apply the algorithm to real DCE-MRI breast cancer data and observe improved pharmacokinetic parameter estimation, separating tumor tissue into regions with differential tracer kinetics on a pixel-by-pixel basis and revealing biologically plausible tumor tissue heterogeneity patterns. This method combines the advantages of multivariate clustering, convex geometry analysis, and compartmental modeling approaches. The open-source MATLAB software of CAM-CM is publicly available from the Web.

Dynamic contrast enhanced MRI (DCE-MRI) is often employed as an indicator of drug activity in clinical trials of angiogenic inhibitors. The data obtained with DCE-MRI is reported semi quantitatively using parameters derived from pharmacokinetic models. Most MRI kinetic models were developed from nuclear medicine quantitative studies but the limitations of MRI dictated specific modifications. A number of these MRI models are in current use. In this work, we review several pharmacokinetic models used in DCE-MRI in order to determine the most appropriate model to assess tumor angiogenesis. These models are compared with respect to their physiological appropriateness; compartmental analysis; initial and boundary conditions; and clinically relevant output parameters.

Vascular disease produces changes in lumenal shape evident in magnetic resonance angiography (MRA). However, quantification of vascular shape from MRA is problematic due to image artifacts. Prior deformable models for vascular surface reconstruction primarily resolve problems of initialization of the surface mesh. However, initialization can be obtained in a trivial manner for MRA using isosurfaces. We propose a methodology for deforming the isosurface to conform to the boundaries of objects in the image with minimal a priori assumptions of object shape. As in conventional methods, external forces attract the surface toward edges in the image. However, smoothing is produced by a moment that aligns the normals of adjacent surface triangles. Notably, the moment produces no translational motion of surface triangles. The deformable isosurface was applied to a digital phantom of a stenotic artery, to MRA of three renal arteries with atherosclerotic disease and MRA of one carotid artery with atherosclerotic disease. Results of the surface reconstruction from the deformable model were compared with conventional X-ray angiography for the renal arteries. Measurement of the degree of stenosis of the renal arteries was within 12% +/- 6%. The deformable model provided improvements over the isosurface in all cases in terms of measurement of the degree of stenosis or improving the surface smoothness.

Radiofrequency ablation (RFA) is a minimally-invasive image-guided method for the local destruction of tumors. Successful ablation, or burning, of tumors, is impeded by blood flow in the vicinity of the tumor that tends to cool the tissue. We have developed methods for visualizing the tumors and their spatial relation to blood vessels for the purpose of treatment planning. We apply these methods to hepatic tumors. The visualization method employs contrast-enhanced (Gd-DTPA) magnetic resonance angiography (MRA) and magnetic resonance venography (MRV). The arteries and veins are delineated using the ordered region-growing (ORG) skeletonization algorithm. Tumors are contoured manually. A shaded surface display is generated that includes arteries, veins and tumors. This 3D map is to be used to optimize treatment planning and to better limit the effects of perfusion on tumor ablation. A better understanding of the relationship of blood vessel location, size and flow to thermal lesions could facilitate improved patient outcomes.

Renovascular disease is an important cause of hypertension. For assessing treatment options for renovascular disease, such as angioplasty or nephrectomy, it is important to characterize the renal tissue. Magnetic resonance (MR) renography is becoming a viable method for the characterization of the renal tissue. However, the analysis of MR renography is hampered by tissue motion. We investigate two automated image registration methods for minimizing the effects of tissue motion. The first is semi-automated registration using contours. The second is an adaptation of the automated image registration (AIR) algorithm that accommodates large-scale motion and tissue enhancement from a contrast agent. We compared the results of these methods with manual registration using image overlays. Semi-automated registration using contours accurately registered a 2D MR renography data set of 140 time frames with obvious errors in only seven slices. With correction in those slices, semi-automatic registration had equivalent quality to manual registration. The adaptation of the AIR algorithm produced better results on 3D MR renography in healthy kidneys than manual registration, but worse results in a diseased kidney. We conclude that automated registration of 2D and 3D MR renography is feasible.

In the setting of von Hippel-Lindau disease, accurate quantitation of kidney lesions is important for genetic research. Unfortunately, fully automated quantitation is difficult because the lesion boundaries are complex. Therefore, we developed a method to semi-automate the quantitation of these renal lesions. We studied helical CT scans of 10 kidneys from 8 patients with von Hippel-Lindau disease. The kidneys were segmented from surrounding structures using an interactive marker-controlled watershed algorithm. Renal lesions (cysts and solid tumors) were identified using thresholding and then characterized by size using mathematical morphology and granulometry. There were 50 cysts and 16 solid lesions. The mean (/spl plusmn/ sd) numbers of interior and exterior manually placed contours required to perform a complete watershed segmentation of the kidneys were 2.2 /spl plusmn/1.2 and 1.2 /spl plusmn/0.6, respectively. The mean difference between the watershed and manual methods of computing renal volume was 13 /spl plusmn/18 mL (5 /spl plusmn/2% of total renal volume) and is not clinically significant. There was no significant difference between volumes of renal lesions measured manually and using the semi-automated method (p > 0.3).

Three-dimensional (3-D) angiographic methods are gaining acceptance for evaluation of atherosclerotic disease. However, measurement of vessel stenosis from 3-D angiographic methods can be problematic due to limited image resolution and contrast. We present a method for reconstructing vessel surfaces from 3-D angiographic methods that allows for objective measurement of vessel stenosis. The method is a deformable model that employs a tubular coordinate system. Vertex merging is incorporated into the coordinate system to maintain even vertex spacing and to avoid problems of self-intersection of the surface. The deformable model was evaluated on clinical magnetic resonance (MR) images of the carotid (n = 6) and renal (n = 2) arteries, on an MR image of a physical vascular phantom and on a digital vascular phantom. Only one gross error occurred for all clinical images. All reconstructed surfaces had a realistic, smooth appearance. For all segments of the physical vascular phantom, vessel radii from the surface reconstruction had an error of less than 0.2 of the average voxel dimension. Variability of manual initialization of the deformable model had negligible effect on the measurement of the degree of stenosis of the digital vascular phantom.

Interpretation of magnetic resonance angiography (MRA) is problematic due to complexities of vascular shape and to artifacts such as the partial volume effect. We present new methods to assist in the interpretation of MRA. These include methods for detection of vessel paths and for determination of branching patterns of vascular trees. They are based on the ordered region growing (ORG) algorithm that represents the image as an acyclic graph, which can be reduced to a skeleton by specifying vessel endpoints or by a pruning process. Ambiguities in the vessel branching due to vessel overlap are effectively resolved by heuristic methods that incorporate a priori knowledge of bifurcation spacing. Vessel paths are detected at interactive speeds on a 500-MHz processor using vessel endpoints. These methods apply best to smaller vessels where the image intensity peaks at the center of the lumen which, for the abdominal MRA, includes vessels whose diameter is less than 1 cm.

In this paper we discuss a method for volumetric segmentation that is used by a system we have developed for body tissue (e.g., bones, organs, internal organ structures, and tumours and cysts) segmentation and visualization in high resolution computerized tomographic (CT) images of the lower torso. The system is designed for application to diagnosis and surgical planning and is especially useful for applications involving von Hippel Lindau kidney disease. We focus on the system's segmentation technique and demonstrate the capabilities of the technique on a number of CT images.