Quantib ND

Quantib ND 1.5 is a medical imaging software that utilizes atlas-based segmentation on MR images to automatically quantify volumes of brain structures and white matter hyperintensities, providing complementary information for neurological assessment. Its methodology incorporates both baseline and longitudinal analysis capabilities, with performance validated against manual segmentations, offering quantitative metrics and comparison to normative data to aid clinicians in quantitative reporting.

Quantib ND 1.5 is a non-invasive medical imaging processing software designed for the automatic labeling, visualization, and volumetric quantification of segmentable brain structures from magnetic resonance (MR) images. Functioning as a post-processing analysis module for Myrian®, it utilizes 3D T1-weighted MR images, with an additional T2-weighted FLAIR MR image for assessing white matter hyperintensities (WMH). The system employs a segmentation methodology based on multiple atlases, each comprising a 3D T1-weighted MR image and a corresponding label map. It offers quantitative data on both absolute and relative volumes of segmented brain regions, including the hippocampus, various lobes (frontal, occipital, parietal, temporal), and the cerebellum, as well as WMH. The software includes Quantib ND Baseline for single time-point analysis and Quantib ND Follow-Up for longitudinal analysis of two time-points, with the latter improving volume change calculation via automated image registration. The output of Quantib ND 1.5 consists of segmentations, visualizations, and volumetric measurements, which can be compared to reference centile data derived from a general population group. While most segmentations are automatic, the WMH segmentation specifically requires review and potential editing by a trained medical professional before validation. Performance validation was conducted by comparing the software's automated segmentations and derived volumes against manual segmentations. For major brain structures like brain tissue and intracranial volume (ICV), Dice indices were high (0.96 and 0.98 respectively). Hippocampus segmentation yielded a Dice index of 0.84. Segmentation of brain lobes generally showed Dice indices ranging from 0.88 (occipital) to 0.98 (cerebellum). WMH segmentation achieved an average Dice overlap of 0.61, with an absolute difference in relative volumes of 0.2 percentage points for non-contrast-enhanced images.

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 atlas-based segmentation algorithm in Quantib ND 1.5 account for anatomical variability across different patient populations, beyond the provided reference centile data?; What specific image registration techniques are employed in the Quantib ND Follow-Up module for longitudinal analysis, and how robust are they to significant interval changes or differing scan acquisition parameters?; Could the performance of White Matter Hyperintensity segmentation (Dice 0.61) be improved, and what are the primary challenges leading to this lower Dice score compared to other brain structures?; What are the demographic and clinical characteristics of the 'group representing the general population' used for the reference centile curves, and how might these factors influence the applicability of these centiles in diverse patient cohorts?; How does the requirement for user review and potential editing of White Matter Hyperintensity segmentations impact overall workflow efficiency and inter-rater reliability in a clinical setting?