Comparative analysis of three-dimensional volume rendering and maximum intensity projection for preoperative planning in liver cancer

Streamlining neuroradiology workflow with AI for improved cerebrovascular structure monitoring

Radiological imaging to examine intracranial blood vessels is critical for preoperative planning and postoperative follow-up. Automated segmentation of cerebrovascular anatomy from Time-Of-Flight Magnetic Resonance Angiography (TOF-MRA) can provide radiologists with a more detailed and precise view of these vessels. This paper introduces a domain generalized artificial intelligence (AI) solution for volumetric monitoring of cerebrovascular structures from multi-center MRAs. Our approach utilizes a multi-task deep convolutional neural network (CNN) with a topology-aware loss function to learn voxel-wise segmentation of the cerebrovascular tree. We use Decorrelation Loss to achieve domain regularization for the encoder network and auxiliary tasks to provide additional regularization and enable the encoder to learn higher-level intermediate representations for improved performance. We compare our method to six state-of-the-art 3D vessel segmentation methods using retrospective TOF-MRA datasets from multiple private and public data sources scanned at six hospitals, with and without vascular pathologies. The proposed model achieved the best scores in all the qualitative performance measures. Furthermore, we have developed an AI-assisted Graphical User Interface (GUI) based on our research to assist radiologists in their daily work and establish a more efficient work process that saves time.

VT3D: a visualization toolbox for 3D transcriptomic data

Data visualization empowers researchers to communicate their results that support scientific reasoning in an intuitive way. Three-dimension (3D) spatially resolved transcriptomic atlases constructed from multi-view and high-dimensional data have rapidly emerged as a powerful tool to unravel spatial gene expression patterns and cell type distribution in biological samples, revolutionizing the understanding of gene regulatory interactions and cell niches. However, limited accessible tools for data visualization impede the potential impact and application of this technology. Here we introduce VT3D, a visualization toolbox that allows users to explore 3D transcriptomic data, enabling gene expression projection to any 2D plane of interest, 2D virtual slice creation and visualization, and interactive 3D data browsing with surface model plots. In addition, it can either work on personal devices in standalone mode or be hosted as a web-based server. We apply VT3D to multiple datasets produced by the most popular techniques, including both sequencing-based approaches (Stereo-seq, spatial transcriptomics, and Slide-seq) and imaging-based approaches (MERFISH and STARMap), and successfully build a 3D atlas database that allows interactive data browsing. We demonstrate that VT3D bridges the gap between researchers and spatially resolved transcriptomics, thus accelerating related studies such as embryogenesis and organogenesis processes. The source code of VT3D is available at https://github.com/BGI-Qingdao/VT3D, and the modeled atlas database is available at http://www.bgiocean.com/vt3d_example.

Procedural Software Toolkit in the Armamentarium of Interventional Therapies: A Review of Additive Usefulness and Current Evidence

Interventional radiology is a fast-paced specialty that uses many advanced and emerging technological solutions. Several procedural hardware and software products are available commercially. Image-guided procedural software helps save time and effort in interventionist practice and adds precision to the intraoperative decisions made by the end user. Interventional radiologists, including interventional oncologists, have access to a wide range of commercially available procedural software that can be integrated into their workflow. However, the resources and real-world evidence related to such software are limited. Thus, we performed a detailed review of the current resources available, such as software-related publications, vendors' multimedia materials (e.g., user guides), and each software's functions and features, to compile a resource for interventional therapies. We also reviewed previous studies that have verified the use of such software in angiographic suites. Procedural software products will continue to increase in number and usage; these will likely be advanced further with deep learning, artificial intelligence, and new add-ins. Therefore, classifying procedural product software can improve our understanding of these entities. This review significantly contributes to the existing literature because it highlights the lack of studies on procedural product software.

Anatomical Description of Loggerhead Turtle (Caretta caretta) and Green Iguana (Iguana iguana) Skull by Three-Dimensional Computed Tomography Reconstruction and Maximum Intensity Projection Images

The growing interest in reptiles has posed a challenge to veterinary clinicians due to the lack of a standardized system to perform anatomical studies similar to those used for dogs and cats. In this paper, we have attempted to describe, employing computed tomography and subsequent three-dimensional reconstructions, the normal anatomical features that comprise the skulls of two species of reptiles: the loggerhead turtle (Caretta caretta) and the green iguana (Iguana iguana). Computed tomography (CT) and subsequent image processing allowed the identification of the bony structures that comprise the head of these species. As a result, and based on previous articles, we propose the most significant anatomical differences and similarities between these species.

Role of CT Imaging With Three-Dimensional Maximum Intensity Projection Reconstruction in the Evaluation of Portal Vein Variants at a Tertiary Care Hospital

INTRODUCTION

Portal vein (PV) is the principal blood vessel transporting blood from the alimentary tract and spleen to the liver. The aim of this study is to determine the prevalence of PV anatomical variations in our population using multidetector CT with maximum intensity projection (MIP) technique at a tertiary care hospital.

METHODS

This cross-sectional study was prospectively conducted from November 2018 to June 2019 in the Department of Radiology at a tertiary care hospital in Karachi. After informed consent, all the patients with no known hepatic pathology undergoing routine abdomen CT were included in this study. Patients with previous hepatic resection surgeries, undiagnosed large hepatic tumors/metastasis, and those with PV thrombosis were excluded.

RESULTS

A total of 500 patients (256 males and 244 females) were included in the study; the mean age of female patients was relatively higher as compared to the male patients (53.80 ± 18.44 vs. 44.15 ± 19.94 years; p = 0.000). Standard PV anatomy (type 1) was found in 438 patients (87.6%). Trifurcation (type 2) occurred in 18 patients (3.6%). Right posterior portal vein as the first branch of main PV (type 3) was found in 22 patients (4.4%). A separate branch of the right portal vein (RPV) to segment VII (type 4) and separate branch of the RPV to segment VI (type 5) were found in 6 (1.2%) and 16 (3.2%) patients, respectively.

CONCLUSION

Our study displayed a relatively higher frequency of standard PV anatomy (type 1) compared to previous studies. We highlight the role of MIP in the analysis of hepatic venous anatomy with its utility demonstrating improved detection of variations.