Quantib Prostate

Quantib Prostate version 2.0 is an image post-processing software that enhances the analysis of prostate MRI scans by enabling automatic prostate and sub-region segmentation, multi-parametric image registration, and computation of a biparametric combination image, with user-interaction modules for editing, PI-RADS scoring, and PSA density determination. Its methodology relies on algorithms for automated processing combined with user-interactive tools for refinement, aiming to assist medical professionals in interpreting prostate MR studies and improve diagnostic workflow.

Quantib Prostate version 2.0 is an image post-processing software designed to assist trained medical professionals in the analysis and review of prostate MRI images. It functions as an extension to the Quantib AI Node software platform, providing functionalities such as registered multiparametric-MRI viewing, semi-automatic segmentation of anatomical structures with volume computations, tools for manual editing, and a structured workflow for PI-RADS scoring. The software aims to enhance the interpretation of prostate MR studies by providing additional mathematical and statistical analysis, with results displayable in various formats, including image overlays. Version 2.0 introduces significant modifications, most notably the division of prostate segmentation into prostatic sub-regions (central gland, peripheral zone, urethra) and the initialization of Region of Interest (ROI) location on the PI-RADS sector map. While the core prostate segmentation algorithm remains unchanged, the new sub-region segmentation can be exported, and its outer boundaries adjust with user edits to the main prostate segmentation. The software also expands its output formats to include PDF reports, DICOM RT structure sets, and DICOM segmentation storage. Non-clinical bench testing involved comparing automatic sub-region segmentation to ground truth using Dice overlap and Mean Surface Distance, while clinical performance was qualitatively assessed by radiologists using a Likert scale for sub-region segmentations and ROI initial localizations.

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 the automatic sub-region segmentation algorithm compare to inter-observer variability among expert radiologists in delineating these specific prostatic zones?; What is the quantitative impact of the ROI localization initialization on the PI-RADS sector map on the efficiency and accuracy of radiologists' lesion annotation and scoring?; Are there specific patient demographics or MRI acquisition protocols where the sub-region segmentation accuracy is notably higher or lower?; How does the biparametric combination image quantitatively improve lesion conspicuity or diagnostic confidence compared to viewing individual parametric maps separately?; What is the learning curve for trained medical professionals to proficiently use the new features, particularly the sub-region segmentation editing and ROI localization initialization?