iQ-solutions

iQ-solutions™ is a software medical device utilizing pre-trained convolutional neural networks to automate the segmentation and volumetric quantification of brain structures from MRI scans. It aims to enhance the efficiency and consistency of neuroimaging analysis by providing quantitative reports on brain volumes and white matter hyperintensities through cross-sectional and longitudinal pipelines, serving as an adjunctive tool for clinicians.

iQ-solutions™ is a standalone software medical device designed to process brain MRI scans for the automatic annotation, visualization, and quantification of segmentable brain structures. Its primary purpose is to accelerate and improve the efficiency of quantifying these structures, a task traditionally performed manually by identifying, annotating, and measuring regions of interest. The software integrates with existing Picture Archiving and Communication Systems (PACS), receiving MRI scans and returning processed results, including quantitative reports and DICOM-formatted annotation masks, to the PACS or other user-specified radiology software systems or databases. The outputs are intended to aid qualified radiologists or physicians and are not standalone diagnostic tools. The core of iQ-solutions™'s analysis capabilities lies in pre-trained convolutional neural networks (CNNs) that segment specific brain structures and generate binary masks. These CNNs are supported by pre-processing components that standardize MRI scans to consistent positions and contrasts, and post-processing components that prepare output annotations and statistical calculations. The device offers two distinct analysis pipelines: a cross-sectional pipeline for structure segmentation and volume analysis from a single time-point MRI, and a longitudinal pipeline for analyzing volume changes using two MRI scans from the same patient taken at different times on the same scanner with consistent protocols. Key outputs include iQ-MS™ reports (detailing whole brain, gray matter, thalami volumes and changes, plus white matter hyperintensities) and iQ-COG™ reports (focusing on cortical gray matter, lobar volumes, hippocampi volumes and changes, asymmetry indices, and white matter hyperintensities), both compared against a healthy cohort based on extensive training data.

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 iQ-solutions™ in segmenting and quantifying brain structures vary across different MRI scanner vendors (Siemens, GE, Philips) and field strengths (1.5T, 3T) beyond the initial training data demographics?; What is the specific methodology and accuracy for the "normalization" of brain volumes when comparing to a healthy cohort, and how is this healthy cohort defined and maintained?; Given that results from cross-sectional and longitudinal pipelines cannot be compared, what are the specific underlying algorithmic or processing differences that necessitate this restriction?; The device reports on "unnormalized volume and volume changes of FLAIR white matter hyperintensities." What is the rationale for not normalizing these volumes, and how does this impact their interpretation, especially in longitudinal studies?; While the device is not for diagnosis, how sensitive are the volumetric change detections in the longitudinal pipeline to subtle pathological changes, and what is the minimum detectable change for key structures like hippocampi or for white matter hyperintensity burden?