ANDI

ANDI is a quantitative neuroimaging software that processes diffusion and T1-weighted MRI data to produce 3D visualizations and measurements of brain white matter tracts, serving as adjunctive information for clinicians. Its methodology leverages advanced image processing techniques, including tractography and a proprietary 3D atlas, to analyze white matter microstructure and output results in a DICOM-compliant PDF report for comparative or longitudinal assessment.

ANDI is a quantitative imaging software designed to process diffusion-weighted MRI (dMRI) sequences, supplemented by T1-weighted images, to generate 3D maps and quantitative measurements of brain white matter tracts. Its purpose is to provide adjunctive information to medical professionals, such as neurologists and radiologists, for visualizing and measuring these tracts, thereby aiding in the assessment of white matter microstructure without offering diagnosis or interpretation of brain health. The system outputs a DICOM encapsulated PDF report containing these analyses, which can be used for comparison against control populations or for longitudinal tracking of a patient's white matter characteristics, with the ultimate clinical interpretation and decision-making resting with the physician. The methodology of ANDI centers on extracting white matter bundles that connect specific brain regions and subsequently performing a detailed microstructure analysis along these pathways. It utilizes local (per-voxel) and global (whole brain) reconstruction algorithms, including modeling, tractography (specifically CSD tractography), and white matter bundling, to map these properties. By leveraging the sensitivity of dMRI to water proton movement and white matter orientation, and T1-weighted images for anatomical context, ANDI quantifies local water diffusion properties. A key component is its proprietary white matter bundles atlas, defined by 3D models, used for segmentation. The software operates via an API for NIfTI image input and DICOM PDF report output, integrating into RIS/PACS systems. Non-clinical performance tests have validated its AI-based brain extraction, processing pipeline robustness, and the content of its generated reports.

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 ANDI's proprietary white matter bundles atlas compare to established atlases like the Glasser atlas in segmenting tracts, particularly in cases with anatomical variations or pathology?; What specific quantitative diffusion metrics (e.g., FA, MD, axial diffusivity, radial diffusivity) are calculated by ANDI, and what is their validated test-retest reliability and sensitivity to subtle microstructural changes?; Given that T2 and FLAIR inputs are not utilized, how does ANDI account for or address potential confounding factors such as white matter lesions or edema that might affect diffusion measurements and tractography accuracy?; What validation studies have been performed on the AI-based brain extraction component, particularly concerning its accuracy across different MRI scanner vendors, field strengths, and patient populations with diverse brain morphologies?; Without specific clinical studies for this device, what is the plan for post-market surveillance or real-world evidence gathering to substantiate the clinical utility of ANDI's outputs in guiding patient management or understanding disease progression?