Transpara Density 1.0.0

Transpara Density 1.0.0 is an artificial intelligence-driven software designed to automatically calculate volumetric breast density, breast volume, and ACR BI-RADS density categories from digital mammography and tomosynthesis images. Its methodology leverages deep learning algorithms operating on processed image data, providing quantitative and categorical outputs to aid healthcare professionals in assessing breast tissue composition, as validated by performance data demonstrating high accuracy, reproducibility, and agreement with radiologist consensus.

Transpara Density 1.0.0 is a software application that utilizes deep learning artificial intelligence algorithms to automatically assess breast tissue composition from digital mammography (DM) and digital breast tomosynthesis (DBT) images. It analyzes processed 'FOR PRESENTATION' mammography images and synthetic 2D DBT images to determine volumetric breast density (VBD) as a percentage, breast volume in cubic centimeters, and an ACR BI-RADS 5th Edition breast density category (A, B, C, or D). The system is designed to aid healthcare professionals by providing these quantitative and categorical density assessments, functioning as a supportive tool rather than a direct diagnostic aid, and requires integration into a DICOM-compatible software environment for clinical operation. The performance of Transpara Density was rigorously evaluated using a large dataset from over 10,000 women, encompassing images from multiple vendors (GE, Siemens, Hologic, Fujifilm). Accuracy assessments showed strong correlations for VBD when compared against a validated physics-based model (Pearson correlation 0.935) and breast MRI studies (Pearson correlation 0.908). The software demonstrated high reproducibility in VBD measurements between different mammographic views (MLO vs. CC, Pearson 0.947), between the left and right breasts of the same patient (Pearson 0.953), and between DM and DBT modalities (VBD Pearson 0.912). In a study comparing its BI-RADS density categorization against a panel majority vote of eight MQSA-qualified radiologists, Transpara Density achieved 70.8% accuracy for the four density categories and 88.9% accuracy for distinguishing dense versus non-dense breasts, with a corresponding sensitivity of 87.3% and specificity of 90.4%.

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 Transpara Density's Volumetric Breast Density (VBD) calculation compare to other commercially available automated breast density assessment tools when using the same imaging datasets?; What is the impact on inter-reader variability among radiologists when Transpara Density's BI-RADS categorization is provided as adjunctive information in clinical practice?; Given that the device utilizes processed 'FOR PRESENTATION' images for mammography and synthetic 2D images for DBT, how sensitive is its performance to variations in image processing parameters across different mammography vendor systems beyond those explicitly tested (GE, Siemens, Hologic, Fujifilm)?; How does Transpara Density's algorithmic approach for density computation using processed images differ fundamentally from methods like the predicate device that utilize raw image data, and what are the implications for accuracy or sensitivity to image acquisition parameters?; What is the potential clinical utility of the 'breast volume in cm3' output, and are there established correlations between this specific measurement provided by Transpara Density and breast cancer risk or other clinical outcomes, independent of VBD and BI-RADS category?