GE HealthCare Launches CleaRecon DL, Introducing AI-Based 3D Reconstruction

AI-driven solution advances image quality, encouraging cone-beam computed tomography.

CleaRecon DL versus conventional imaging.
CleaRecon DL versus conventional imaging.
GE Healthcare

GE HealthCare has announced the launch of CleaRecon DL, technology powered by a deep-learning algorithm, to improve the quality of cone-beam computed tomography (CBCT) images.

The artificial intelligence (AI)-driven solution is designed to remove streak artifacts caused by the pulsatile nature of blood flow in the arteries. The solution additionally addresses changes in the distribution of contrast during CBCT acquisitions in liver, prostate, neuro, and endovascular aortic repair procedures. CleaRecon DL recently received U.S. FDA 510(k) clearance and CE mark and will be available for use on the Allia platform.

CBCT is used in interventional suites to provide cross-sectional imaging during procedures. However, the quality of CBCT reconstructed images may be diminished due to artifacts resulting from vessels’ pulsatility, which can reduce image clarity and accuracy. These limitations can impact the confidence in CBCT image interpretation and its adoption in routine clinical practice. Despite these challenges, CBCT remains crucial in interventional procedures for its ability to provide comprehensive visualization of anatomical structures and may enhance procedural accuracy.

“The introduction of CleaRecon DL represents a leap forward in the interventional suite and for the advancement of CBCT. By improving image quality and reducing artifacts, this technology can empower clinicians to perform procedures with greater precision and confidence,” said Arnaud Marie, General Manager, Interventional Solutions at GE HealthCare. “This solution builds on our portfolio of tools aimed at improving the user experience and workflow efficiency, enabling clinicians to deliver more accurate and effective interventions for enhanced patient outcomes.”

Deep learning is an AI technology that has become a machine learning technique for image processing and is trained to output data and perform specific tasks. It is based on population representative data collection and thorough tests with clinical domain experts. During clinical validation testing, a recent survey noted that in 98% of cases, CBCT images reconstructed with CleaRecon DL are clearer than conventional CBCT images. This technology was also shown to improve CBCT image interpretation confidence in 94% of cases.

More in Imaging