RSI-MRI+

RSI-MRI+ is an advanced image processing software that enhances prostate MRI analysis by automatically segmenting the prostate and generating a 'Restricted Signal Map' from diffusion-weighted and T2-weighted images using deep learning. Its purpose is to provide clinicians with improved visualization of restricted diffusion and quantitative metrics to aid in image interpretation, without making a direct diagnosis.

RSI-MRI+ is standalone software designed to assist radiologists, urologists, and other clinicians in the analysis and interpretation of medical images. It functions as an automatic post-acquisition image processing tool for diffusion-weighted (DWI) and anatomical magnetic resonance imaging (MRI) data. Its core capabilities include the automatic fusion of derived DWI data with anatomical T2-weighted MR images, alongside automated prostate segmentation, quantification of prostate volume, and the generation of reports detailing these derived image metrics. The software is intended for use in various clinical settings such as hospitals, clinics, and medical offices, but is not indicated for pediatric populations or for direct diagnosis, treatment, or prevention of diseases. Technologically, RSI-MRI+ utilizes artificial intelligence, specifically a deep learning algorithm, for its automated prostate segmentation on T2-weighted images. A key output is the "Restricted Signal Map," an enhanced DWI map designed to improve the conspicuity of restricted diffusion compared to standard DWI maps. The software can also produce a color fusion series, overlaying the Restricted Signal Map intensity onto anatomical T2-weighted images, and generates an automated segmentation report. All outputs are provided in standard DICOM format, ensuring compatibility with most third-party commercial PACS workstation software. Performance testing has demonstrated the increased conspicuity of its Restricted Signal Map, the accuracy of its automated segmentation when compared to manual segmentations by radiologists, and its ability to normalize diffusion signals across different acquisitions and scanners.

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. What specific deep learning architecture is employed for the automated prostate segmentation, and what was the reported Dice similarity coefficient or other relevant accuracy metric against manual radiologist segmentations in the validation studies?; How does the 'Restricted Signal Map' mathematically derive its improved conspicuity of restricted diffusion compared to standard DWI (e.g., ADC maps), and what quantitative improvements in lesion detection or characterization were observed in performance testing?; Given that RSI-MRI+ supported protocols differ from PI-RADS v2.1 in T2W resolution and GE DWI slice thickness, what is the documented impact of these differences on the software's performance and its utility in a PI-RADS-compliant workflow?; How does the software's diffusion signal normalization algorithm account for variations in MRI acquisition parameters (e.g., different b-values, field strengths, vendor-specific sequences beyond GE) to ensure consistent Restricted Signal Map generation?; What are the computational requirements (e.g., processing time per case, minimum hardware specifications) for RSI-MRI+ to perform a complete analysis, including segmentation and map generation, on typical prostate MRI datasets?