LiverSmart

LiverSmart is a software tool that integrates the functionalities of HepaFat-AI and FerriSmart to provide a consolidated, multi-parametric assessment of liver health from MRI data, measuring both liver fat and iron concentration. It automates the analysis workflow by processing DICOM images through its Data Preparation Module and generating a unified report via its Report Generation Module, utilizing the underlying convolutional neural networks and algorithmic calculations of its predicate devices.

LiverSmart is a software tool designed to streamline the assessment of liver health by integrating data from two pre-existing analysis platforms, HepaFat-AI and FerriSmart. Its primary purpose is to provide a consolidated, multi-parametric report detailing both liver fat content and liver iron concentration from magnetic resonance imaging (MRI) datasets. The methodology involves the user uploading a single zipped folder of DICOM images, acquired according to specific protocols for HepaFat-AI and FerriSmart. LiverSmart then automatically sorts and sends these datasets to the respective analysis tools, and incorporates two new modules: a Data Preparation Module for image validation and error detection, and a Report Generation Module to compile the outputs into a unified report. The scientific basis of LiverSmart relies on the established capabilities of its component devices. HepaFat-AI utilizes MRI to quantify volumetric liver fat fraction (VLFF), proton density fat fraction (PDFF), and determine steatosis grade, while FerriSmart uses MRI to measure liver iron concentration (LIC) by assessing the R2 signal decay rate; both underlying tools employ convolutional neural networks for image analysis and algorithms for quantitative measurements. The key result provided by LiverSmart is a comprehensive report that combines these distinct liver parameters, offering clinicians metrics for diagnosing and monitoring conditions like systemic iron overload, thalassemia, sickle cell disease, fatty liver disease, and for assessing liver transplant donors or patients undergoing weight loss management, with identical quantitative outputs as if the predicate devices were used independently.

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 Data Preparation Module in LiverSmart specifically validate MRI image quality and what are the common anomalies it is designed to detect beyond those handled by HepaFat-AI or FerriSmart individually?; What is the typical processing time for LiverSmart to generate a consolidated report from the moment DICOM images are uploaded, and how does this compare to running HepaFat-AI and FerriSmart analyses separately and manually compiling results?; Are there specific MRI scanner compatibility requirements or limitations for the DICOM images used with LiverSmart that differ from or are more stringent than those for HepaFat-AI and FerriSmart independently?; How does LiverSmart handle discrepancies or conflicting quality flags if one component analysis (e.g., HepaFat-AI) deems an image set acceptable while the other (e.g., FerriSmart) flags an issue from the same uploaded folder?; What are the quantifiable benefits, such as reduction in interpretation time or error rates, observed with the automated Report Generation Module of LiverSmart compared to manual compilation of HepaFat-AI and FerriSmart results?