Visage Imaging, Inc.

Visage Breast Density

1

01/29/2021

K201411

MIMPS

Images must be 2D projection images, either original 2D projections or synthetic reconstructions from DBT volumes (e.g., 'C-Views' from Hologic scanners). The modality must be 'MG' (Mammography). The manufacturer must match the devices for which the CNN has been trained, in this case Holologic. The presentation intent type must be 'for presentation'. If defined, the pixel intensity relationship must be 'LOG'. The image laterality must be 'R' (Right) or 'L' (Left). The view position must be 'CC' (Cranio-caudal) or 'MLO' (Medio-lateral oblique). The patient orientation must have the row direction encoded as 'P' (Positive) or 'A' (Anterior). The content of 'breast implant present' is recorded, regardless of its value. Images should be turned into 8-bit grey level images via application of the first recommended DICOM window.

Reconstructed slices of DBT volumes via their SOP class UID tag (0008, 00016) are excluded. Images from devices for which the CNN has not been trained are excluded. Images with view modifiers 'Spot Compression' or 'View Magnification' are excluded. If the content of 'breast implant present' is 'YES', the image is only accepted if the view modifier 'Implant Displaced' is set.

Not available

Visage Breast Density is a software application intended for use with compatible full field digital mammography and digital breast tomosynthesis systems. Visage Breast Density assesses breast density from a mammography study and provides an ACR BI-RADS Atlas 5th Edition breast density category to aid radiologists in the assessment of breast tissue composition.

Radiologist

Adult

Chest

MAM

ACR BI-RADS Atlas 5th Edition breast density category to aid in the assessment of breast tissue composition(adjustable by radiologist), confidence score.

33,000

Unspecified

Hologic\Lorad

Selenia Dimensions

Reader Study

N/A

N/A

N/A

N/A

N/A

N/A

N/A

N/A

N/A

N/A

3

1,200

2

Unspecified

Unspecified

Hologic\Lorad

Selenia Dimensions

Site A: 84

Not computed - Confusion matrices used for validation of Breast Density

Not computed - Confusion matrices used for validation of Breast Density

For model performance rubric and additional details - please use this link - https://pubmed.ncbi.nlm.nih.gov/37421715/ Post-deployment performance study of Visage Breast density showed very high radiologist agreement, indicating strong support for use as an adjunct or standalone density assessment tool. For a total of 291 days and 93,560 mammograms, the agreement of the radiologists with the AI was 99.35%. Of the 0.65% cases in which there was disagreement, the difference was only by one category. There were no cases of disagreement of more than one category. In case of radiologist disagreement, 59% cases were upgrade and 41% were downgraded. The change in density assignment for upgraded cases was: A to B (62%), B to C (30%) , C to D (8%) and for downgrading: B to A (24%), C to B (48%), D to C (28%). The most commonly changed categories were upgrading A to B and downgrading C to B. Reference: Philpotts L, Szabla P, Lin M. Artificial Intelligence Platform for Determining Breast Density in Digital Breast Tomosynthesis Mammography: How Does It Perform?, American Roentgen Ray Society (ARRS) 2023, presentation. https://www.auntminnie.com/clinical-news/womens-imaging/breast/article/15633258/ai-radiologists-match-when-it-comes-to-classifying-breast-density

Hologic\Lorad

Unspecified

Unspecified

Unspecified

Unspecified

Unspecified

Unspecified

Unspecified

Unspecified