VisAble.IO

VisAble.IO is a CT image processing software designed to assist physicians in planning and confirming liver ablation procedures by providing tools for anatomic segmentation, virtual needle placement, and quantitative analysis of ablation zones. It employs AI algorithms for liver and vessel segmentation and a structured workflow to enhance the precision and evaluation of liver ablation treatments, though final clinical accuracy relies on user judgment.

VisAble.IO is a user-controlled software application designed for processing Computed Tomography (CT) images specifically for liver ablation procedures. It facilitates the import of images from CT scanners and PACS systems, allowing physicians to utilize these images for both planning and confirmation stages of liver ablation. Its core purpose is to assist in identifying ablation targets, planning virtual needle placements, and subsequently confirming the achieved ablation zones post-procedure, though it is explicitly not intended for diagnostic purposes or predicting treatment outcomes. The software operates through a five-step workflow: data import, anatomic structures segmentation (including liver, vessels, and target), instrument placement planning, ablation zone segmentation, and treatment confirmation involving image registration and quantitative analysis. Key functionalities include manual and AI-driven automated segmentation tools, virtual instrument overlay, image fusion, and the computation of achieved margins and missed volumes to evaluate target coverage. VisAble.IO supports specific Medtronic and NeuWave Medical needle models for planning and utilizes AI algorithms for key segmentation tasks; validation studies demonstrated a mean DICE score of 0.98 for liver segmentation (goal 0.92), 0.80 for ablation target segmentation (goal 0.70), 0.86 for ablation zone segmentation (goal 0.70), 0.72 for liver vessel segmentation (goal 0.70), and a mean corresponding distance of 4.11 mm for pre-post ablation image registration (goal 6.06 mm).

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 performance of the AI-driven liver and vessel segmentation algorithms (e.g., DICE scores) compare to manual segmentation by experienced radiologists in terms of time efficiency and accuracy for complex cases?; What are the specific limitations of the software when dealing with CT images of varying quality or from different CT scanner manufacturers, beyond general image resolution dependencies?; Can the system's quantitative analysis of 'achieved margins' and 'missed volumes' be correlated with long-term patient outcomes or recurrence rates in liver ablation?; How does the 'automated local rigid registration' handle significant anatomical changes or deformations between pre- and post-ablation scans, and what is the typical extent of user modification required?; Given the software supports specific needle models for planning, how adaptable is it to new or different ablation needle technologies, and what is the process for incorporating them?