Axial3D Insight

Axial3D Insight is a cloud-based software service that processes DICOM medical images using segmentation techniques, including machine learning for initial outlines, to create digital files for 3D printing patient-specific anatomical models. These models, validated for accuracy and intended use through clinical studies and phantom testing, assist clinicians in treatment planning and diagnostic procedures for trauma, orthopedic, maxillofacial, and cardiovascular conditions, supplementing expert clinical judgment.

Axial3D Insight is a secure, cloud-based medical image management and processing system designed for image segmentation and 3D modeling. It processes DICOM imaging information from medical scanners, specifically CT and CTA modalities, to generate an output file. This digital file can then be utilized to fabricate patient-specific physical replicas via additive manufacturing methods, or the digital file itself can be used directly. The system employs a combination of automated tools and smart editing tools for segmentation, including machine learning models (AxialML) for initial segmentation, which are subsequently reviewed and finalized by trained personnel. The primary purpose of the Axial3D Insight output, whether as a digital file or a 3D printed physical model, is to aid in treatment planning and for diagnostic purposes across several medical fields: trauma, orthopedic, maxillofacial, and cardiovascular applications. It is intended to be used in conjunction with other diagnostic tools and expert clinical judgment. The system's performance has been validated through several studies, including a Clinical Segmentation Performance study using the RADPEER framework, which confirmed segmentation accuracy with all cases scored within acceptance criteria, and an Intended Use Validation study, which affirmed that the 3D models meet end-user needs. Additionally, phantom testing was conducted to verify the geometric accuracy (to an acceptance criterion of ±0.3mm) of models produced by newly added compatible 3D printers like the Stratasys Origin One and J850.

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. Regarding the AxialML machine learning component, what metrics define its performance for initial segmentation, and how does this performance vary with different image qualities or anatomical complexities before human refinement?; What specific quantitative measures, beyond dimensional accuracy from phantom tests (e.g., volumetric fidelity, surface smoothness), are used to validate the anatomical correctness of the 3D printed models against the original patient DICOM data for clinical use?; Are there comparative studies evaluating the impact of using Axial3D Insight models on surgical planning time, intraoperative efficiency, or patient outcomes versus traditional methods or other 3D visualization tools?; The Indications for Use state the device should be used with "expert clinical judgment." What are the recommended qualifications or training for users to effectively integrate Axial3D Insight outputs into their diagnostic and treatment planning workflows?; Given its application in cardiovascular cases, what are the documented capabilities or limitations of Axial3D Insight in handling complex dynamic cardiac imaging or cases where CT/CTA data might be suboptimal compared to MRI, particularly for soft tissue delineation?