MuscleView

MuscleView is a machine learning-based software designed for the automatic segmentation of lower extremity muscle and bone structures from MRI scans, providing quantitative metrics for use by trained physicians. Its development involved training and validating an artificial intelligence component on a diverse MRI dataset, demonstrating substantial equivalence to existing devices through robust performance in segmentation accuracy and volume measurement.

MuscleView, a software by Springbok, Inc., automates the segmentation of lower extremity muscle and bone structures from MRI images using a machine learning approach. It generates derived metrics such as muscle volume, bone volume, intramuscular fat percentage, and left/right asymmetry for adult and pediatric patients (18+), intended for use by physicians trained in MRI interpretation as an initial segmentation method that requires subsequent review and editing. The software is explicitly not for direct diagnostic guidance or for use in patients with lower limb tumors. The device achieved substantial equivalence to a comparable legally marketed device, a conclusion supported by comprehensive software verification and validation testing, including a rigorous evaluation of its artificial intelligence (AI) component. The AI was trained and validated on a diverse dataset of MRI scans, consistently demonstrating accurate performance across varied patient demographics, scan sites, and MRI parameters. Performance data, assessed by metrics like the Dice Similarity Coefficient and volume difference, successfully met predefined acceptance criteria, affirming MuscleView's safety and effectiveness comparable to existing devices without introducing new safety or effectiveness concerns.

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. What is the specific performance of MuscleView's segmentation accuracy (e.g., Dice Coefficient) across different lower extremity muscle groups or anatomical variations?; How does the software handle imaging artifacts or variations in MRI scan parameters (e.g., field strength, sequence types) that might affect segmentation quality?; Beyond volume and fat percentage, what other quantitative metrics or insights can MuscleView provide that are clinically relevant for muscle health or disease progression?; Given its non-diagnostic intent, how should clinicians integrate MuscleView's initial segmentation results with their existing diagnostic workflow for optimal patient management?; What is the long-term plan for software updates, dataset expansion, and performance monitoring to ensure MuscleView's continued accuracy and applicability in evolving clinical practice?