Author: Taylor A Beal, Kari J. Brewer Savannah, Kristy K. Brock, Alejandro Contreras, Natalie W Fowlkes, Megan Kalambo, Gregory P Reece, Erin P Snoddy, Tien T Tang ๐จโ๐ฌ
Affiliation: The University of Texas MD Anderson Cancer Center, Baylor College of Medicine ๐
Purpose: Current anatomical and surgical research does not adequately detail the breast fascial systemโs ligaments and connective tissues. Most available information stems from cadaver dissections, which alter the three-dimensional integrity of the breast. We introduce a cryo-fluorescence tomography (CFT) imaging and image processing workflow that characterizes specific features of the breast fascial system while preserving spatial information.
Methods: Seven sagittal excised clinical breast tissue slices were flash-frozen, embedded in an optimal cutting temperature (OCT) medium, and imaged using a CFT imager. The CFT imaging leverages fluorescence imaging capabilities to capture the natural autofluorescence signals of collagen fibers in ligaments. To process the images, we utilized a Frangi vessel enhancement filter to segment the ligaments before skeletonizing them (creating a simplified representation of the ligaments). To validate the skeletonization model, we compared segmented slices with images annotated by a radiologist. Finally, branchpoints/nodes and ligament length were extracted from the segmented images, and Spearmanโs Rank correlation analysis was performed with clinical information (age, BMI, and bra cup size).
Results: False positive and false negative rates were 0.6% and 34.3%, respectively, with a sensitivity of 65.7% and specificity of 99.4%. The Spearmanโs Rank correlation analysis showed a strong positive correlation between bra cup size and the number of ligaments in the superior region of the breast (ฯ = 0.77, p = 0.05) and a strong negative correlation between bra cup size and the number of nodes in the superior region (ฯ = -0.66, p = 0.11). This suggests that while women with larger breasts might have more ligament attachments at the top of the breast, the actual branching network is more simplified.
Conclusion: This work demonstrates the potential of characterizing the fascial tissues of the entire breast. Future work will focus on expanding the sample size to provide a more comprehensive regression analysis.