QyScore Software

QyScore Software provides automated volumetric analysis of brain structures and lesions from 3DT1 and T2FLAIR MR images, employing a dynamic probabilistic neuroanatomical atlas and comparison to normative data to assist in neurological assessment. The system quantifies volumes for structures including the hippocampus, grey/white matter, and white matter hyperintensities, delivering detailed reports and visualizations for patients aged 20-90, with the specific limitation that it is not designed for counting individual white matter lesions.

QyScore Software is an advanced medical image processing tool designed for the automatic labeling, visualization, and precise volumetric quantification of segmentable brain structures and lesions using Magnetic Resonance (MR) images. Intended for patients between 20 and 90 years old, the software processes 3DT1 and T2FLAIR MRI series to segment key areas such as grey matter, white matter, hippocampus, amygdala, and white matter hyperintensities; a core feature is its ability to compare these volumetric measurements against reference percentile data derived from a healthy population, offering a quantitative benchmark for clinical assessment, and it integrates with existing DICOM servers and RIS/PACS systems. The underlying methodology of QyScore involves retrieving DICOM MRI data for server-based analysis, including automatic segmentation based on a dynamic probabilistic neuroanatomical atlas that utilizes MR image intensity, alongside processes like bias correction and normalization of volumes against intracranial volume, outputting detailed electronic reports with color overlays accessible via a dedicated graphical user interface. Performance evaluations demonstrate high accuracy, with Dice coefficients exceeding 85% for whole brain regions and 75-85% for subcortical structures from 3DT1 scans, and a mean absolute volume difference of 3.34 mL for lesion segmentation from paired 3DT1 and T2FLAIR sequences, though QyScore is not intended for clinical scenarios requiring the evaluation of the number of white matter hyperintensities.

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 "dynamic probabilistic neuroanatomical atlas" account for anatomical variability due to age (20-90 years) and potential pathological changes, and what are the performance implications if an individual's anatomy significantly deviates from the atlas's training data?; What is the statistical methodology behind the "reference percentile data," including the size, demographics, and MRI acquisition parameters of the normative cohort, and how sensitive are the percentile comparisons to variations in MRI protocols across different institutions?; Considering the mean absolute volume difference of 3.34 mL for lesion segmentation, how does this accuracy translate to longitudinal monitoring of lesion burden in diseases like multiple sclerosis, where small changes can be clinically significant?; What specific features or algorithms within QyScore differentiate its white matter hyperintensity volumetric quantification from the enumeration of lesions, and what is the impact of this distinction on its utility for specific cerebrovascular or neurodegenerative disease assessments?; Beyond volumetric data, does QyScore provide or plan to incorporate any other quantitative imaging biomarkers (e.g., shape analysis, textural features) from the segmented structures that could enhance diagnostic or prognostic capabilities?