Breast Arterial Calcification: A broader proxy for medium and large vessel atherosclerotic calcium

Congress:

ECR 2024

Poster Number:

C-19050

Type:

Scientific Exhibit

Keywords:

Breast, CT, Computer Applications-Detection, diagnosis, Arteriosclerosis

Authors:

C. Parghi, J. Hoffmeister, J. Go, J. Pantleo, N. S. Gonzalez, Z. Zhang, A. Sharma, W. Zhang

DOI:

10.26044/ecr2024/C-19050

DOI-Link:

https://dx.doi.org/10.26044/ecr2024/C-19050

Results

The rate of clinically significant atherosclerotic disease in large and medium vessels was 31.0% (95% CI: 23.7% - 39.4%) when the BAC score was 3 or more greater compared to a rate of 11.0% (95% CI: 9.4% - 12.8%) when the BAC score was less than 3, an increase of 20.0% (95% CI: 11.9% - 28.1%, p < 0.001). Sensitivity was 21.8% (95% CI: 16.4% - 28.2%), Specificity 92.9% (95% CI: 91.4% - 94.2%), and Accuracy 83.8% (95% CI: 81.8% - 85.6%).

Fig 7: Fig 7. BAC AI detection and score examples.

Fig 1: Fig 1. AI BAC Algorithm: Deep convolution neural network models trained to...

Fig 2: Fig 2. U-Net: Vessel detection/segmentation.

Fig 3: Fig 3. ResNet: BAC vessel classification.

Fig 4: Fig 4. Truth Annotation and Scoring: Expert radiologists annotate BAC vessels...

Fig 5: Fig 5. Mammogram and CT abdomen/pelvis performed within 12 months on same...

Fig 6: Fig 6. Model Training: Transfer learning to fine-tune the model.

Fig 7: Fig 7. BAC AI detection and score examples.