An efficient and fast GPU-based algorithm for visualizing large volume of 4D data from virtual heart simulations

MR and ultrasound cardiac image dynamic visualization and synchronization over Internet for distributed heart function diagnosis

Dual-modality 4D cardiac data visualization can convey a significant amount of complementary image information from various sources into a single and meaningful display. Even though there are existing publications on combining multiple medical images into a unique representation, there has been no work on rendering a series of cardiac image sequences, acquired from multiple sources, using web browsers and synchronizing the result over the Internet in real time. The ability to display multi-modality beating heart images using Web-based technology is hampered by the lack of efficient algorithms for fusing and visualizing constantly updated multi-source images and streaming the rendering results using internet protocols. To address this practical issue, in this paper we introduce a new Internet-based algorithm and a software platform running on a Node.js server, where a series of registered cardiac images from both magnetic resonance (MR) and ultrasound are employed to display dynamic fused cardiac structures in web browsers. Taking advantage of the bidirectional WebSocket protocol and WebGL-based graphics acceleration, internal cardiac structures are dynamically displayed, and the results of rendering and data exploration are synchronized among all the connected client computers. The presented research and software have the potential to provide clinicians with comprehensive information and intuitive feedback relating to cardiac behavior and anatomy and could impact areas such as distributed diagnosis of cardiac function and collaborative treatment planning for various heart diseases.

A CT-guided robotic needle puncture method for lung tumours with respiratory motion

PURPOSE

Percutaneous interventions rely on needle puncture to deliver medical devices into lesions. For lung tumours, respiratory motion makes effective puncture procedures difficult to achieve. To address this issue, a needle puncture method considering respiration is proposed to improve the accuracy of lung puncture.

METHODS

The accuracy of puncture is ensured by visualization and needle guidance. Dynamic visualization of the respiratory motion is developed for needle path planning based on four-dimensional computed tomography (4DCT) images. The rendered image is synchronized with the actual breathing by using respiratory signals. A robotic needle insertion strategy for velocity adjustment based on these respiratory signals is designed to guide the needle towards the moving tumour.

RESULTS

The dynamic visualization was tested on multiple 4DCT datasets and achieved a frame rate of over 32 frames per second (FPS). A computer simulation was carried out to verify the feasibility of the needle insertion strategy. Needle puncture was performed on a phantom, and a mean accuracy of 1.34±0.18 mm was achieved.

CONCLUSIONS

In this paper, an efficient and robust method is proposed to improve the visualization and targeting of lung puncture, which reduces the impact of respiratory motion on the accuracy.