3DXR

The 3DXR is a post-processing software designed for three-dimensional reconstruction of X-ray interventional images, enhancing their utility for diagnosis and procedural planning by mitigating common artifacts. This is achieved through the integration of the "CleaRecon DL" option, a Deep-Learning based algorithm specifically developed to remove streak artifacts, which has been validated through rigorous non-clinical and retrospective clinical evaluations to maintain image quality and safety.

The 3DXR is a sophisticated post-processing software application by GE Medical Systems SCS, engineered to perform three-dimensional (3D) reconstruction of images sourced from X-ray interventional systems. Operating on an Advantage Workstation, its core function is to generate 3D volumes to assist physicians in critical areas such as diagnosis, surgical planning, interventional procedures, and treatment follow-up, particularly for detailed imaging of vessels, bones, soft tissues, and other internal body structures. A significant innovation integrated into this system is the "CleaRecon DL" option, which employs a Deep-Learning based algorithm specifically designed to effectively remove streak artifacts, especially those originating from the pulsatile nature of blood flow around contrasted objects or medical devices. Rigorous non-clinical and clinical testing phases were conducted to comprehensively evaluate the modified 3DXR system. Engineering bench tests conclusively demonstrated the CleaRecon DL algorithm's proficiency in reducing streak artifacts without any detrimental impact on overall image quality. Further validation came from a retrospective clinical image quality evaluation, which involved 110 patient examinations reviewed by 13 clinicians, where an impressive 98% of reconstructed images were assessed as clearer than those produced by conventional methods, with no instances of inadvertently hidden or artificially created structures. These comprehensive evaluations, coupled with adherence to established quality system processes, affirm that the device possesses technological characteristics identical or substantially equivalent to its predecessor, ensuring its continued safety and effectiveness for its intended clinical applications, especially for the adult population regarding the CleaRecon DL feature.

AI Regulatory Concerns

Powered by

Jev reads the full FDA letter and returns calibrated concern probabilities per regulatory category. These are AI estimates, not regulatory advice.

• Q1. How does the CleaRecon DL deep-learning model specifically differentiate between pulsatile blood flow artifacts and actual anatomical structures, and what was the nature of its training dataset?; What were the quantitative metrics used in the engineering bench tests to assess the extent of streak artifact reduction and the preservation of crucial image details?; Were there any specific types of patient pathologies or anatomical regions where the CleaRecon DL algorithm demonstrated more or less significant improvement, or where its performance varied?; What are the computational demands and typical processing times for the CleaRecon DL option compared to conventional 3DXR reconstruction methods, especially relevant for real-time interventional procedures?; Are there any plans for prospective clinical trials to further evaluate the impact of CleaRecon DL on diagnostic accuracy, treatment planning efficacy, or ultimate patient outcomes?