Medihub Prostate

MEDIHUB PROSTATE is an image processing software that utilizes an AI-based algorithm for semi-automatic prostate outlining on MR images, requiring user validation to support radiologists and urologists in quantitative analysis and reporting. Its methodology involves multi-parametric MRI processing, image viewing and editing tools, and performance validated by Dice coefficient and Hausdorff distance metrics against expert annotations.

MEDIHUB PROSTATE is an image processing software package designed to assist trained radiologists and urologists in the analysis of prostate MR images. Its core functions include semi-automatic outlining of the prostate, processing, viewing, and editing of MR data. The software supports study review and analysis by providing tools for registered multi-parametric MRI viewing, combining MR sequences into a single image for enhanced visualization, and generating a structured report which includes an edited PI-RADS report and semi-annotated prostate regions. While it uses an AI-based algorithm to initially outline the prostate volume, it requires user interaction to edit, review, and confirm the final segmentation and all other image annotations and measurements, emphasizing that patient management decisions should not solely rely on its analysis. The device's AI functionality is primarily for assessing total prostate volume, not for segmenting lesions. Performance was evaluated using metrics like the Dice coefficient and Hausdorff Distance against ground truth data derived from expert radiologist consensus. Clinical testing on T2 MR images from Siemens 3T Vida and Skyra MRI machines (slice thickness 3.5-5mm, patients aged 55+) showed a mean Dice coefficient of 0.928 and a Hausdorff distance of 2.171 for stand-alone algorithm performance. Reader performance studies indicated improved Dice coefficients for prostate outlining when clinicians used the software's assistance. The system is DICOM compatible and facilitates manual input for PI-RADS scoring, location of seminal vesicles, prostate zones, and DWI/DCE graphs.

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 MEDIHUB PROSTATE's semi-automatic segmentation vary across different MRI acquisition parameters not explicitly validated, such as different field strengths or scanner vendors beyond Siemens?; What is the learning curve for radiologists and urologists to proficiently use MEDIHUB PROSTATE for prostate outlining and PI-RADS reporting, and how does it compare to fully manual methods?; Are there specific prostate pathologies or anatomical variations where the AI-based initial outlining performs less reliably, requiring more significant user correction?; How does the 'biparametric combination image viewer' feature specifically enhance diagnostic interpretation compared to viewing individual sequences separately?; What impact does the use of MEDIHUB PROSTATE have on inter-reader variability in prostate volume measurement and PI-RADS scoring among different users?