Sugar-sweetened beverage consumption and breast composition in a longitudinal study of Chilean girls

Background Frequent sugar-sweetened beverage (SSB) intake has been associated with indirect markers of breast cancer risk, such as weight gain in adolescents and early menarche. How SSB intake relates to breast composition in adolescent girls has not been explored. Methods We evaluated the association between prospective intake of SSB and breast density in a cohort of 374 adolescent girls participating in the Growth and Obesity Cohort Study in Santiago, Chile. Multivariable linear regression models were used to analyze the association between average daily SSB intake quartiles and breast composition (absolute fibroglandular volume [aFGV], percent fibroglandular volume [%FGV], total breast volume [tBV]). Models were adjusted for potential confounding by BMI Z-score, age, daily energy intake (g/day), maternal education, hours of daily television watching after school, dairy intake (g/day), meat intake (g/day), waist circumference, and menarche. To examine the sensitivity of the association to the number of dietary recalls for each girl, analyses were further stratified by girls with one dietary recall and girls with > one dietary recall. Results A total of 881 dietary recalls were available for 374 girls prior to the breast density assessment. More than 60% of the cohort had > one dietary recall available. In multivariable analyses, we found no association between SSB intake quartile and aFGV (Q2 vs Q1 β: − 5.4, 95% CI − 15.1, 4.4; Q3 vs Q1 β: 1.3, 95% CI − 8.6, 11.3; Q4 vs Q1 β: 3.0, 95% CI − 7.1, 13). No associations were noted for %FGV and tBV. Among girls with at least one dietary recall, we found no significant associations between SSB intake quartiles and %FGV, aFGV, or tBV. Conclusion Overall, we observed no evidence that SSB intake was associated with breast density in adolescent Chilean girls.

The Potential Role of the Fat–Glandular Interface (FGI) in Breast Carcinogenesis: Results from an Ultrasound Tomography (UST) Study

This study explored the relationship between the extent of the fat–glandular interface (FGI) and the presence of malignant vs. benign lesions. Two hundred and eight patients were scanned with ultrasound tomography (UST) as part of a Health Insurance Portability and Accountability Act (HIPAA)-compliant study. Segmentation of the sound speed images, employing the k-means clustering method, was used to help define the extent of the FGI for each patient. The metric, α, was defined as the surface area to volume ratio of the segmented fibroglandular volume and its mean value across patients was determined for cancers, fibroadenomas and cysts. ANOVA tests were used to assess significance. The means and standard deviations of α for cancers, fibroadenomas and cysts were found to be 4.0 ± 2.0 cm−1, 3.1 ± 1.7 cm−1 and 2.3 ± 0.9 cm−1, respectively. The differences were statistically significant (p < 0.001). The separation between the groups increased when α was measured on only the image slice where the finding was most prominent, with values for cancers, fibroadenomas and cysts of 5.4 ± 3.6 cm−1, 3.6 ± 2.3 cm−1 and 2.4 ± 1.5 cm−1, respectively. Of the three types of masses studied, cancer was associated with the most extensive FGIs, suggesting a potential role for the FGI in carcinogenesis, a subject for future studies.

Application of convolutional neural networks to breast biopsies to delineate tissue correlates of mammographic breast density

Breast density, a breast cancer risk factor, is a radiologic feature that reflects fibroglandular tissue content relative to breast area or volume. Its histology is incompletely characterized. Here we use deep learning approaches to identify histologic correlates in radiologically-guided biopsies that may underlie breast density and distinguish cancer among women with elevated and low density. We evaluated hematoxylin and eosin (H&E)-stained digitized images from image-guided breast biopsies (n = 852 patients). Breast density was assessed as global and localized fibroglandular volume (%). A convolutional neural network characterized H&E composition. In total 37 features were extracted from the network output, describing tissue quantities and morphological structure. A random forest regression model was trained to identify correlates most predictive of fibroglandular volume (n = 588). Correlations between predicted and radiologically quantified fibroglandular volume were assessed in 264 independent patients. A second random forest classifier was trained to predict diagnosis (invasive vs. benign); performance was assessed using area under receiver-operating characteristics curves (AUC). Using extracted features, regression models predicted global (r = 0.94) and localized (r = 0.93) fibroglandular volume, with fat and non-fatty stromal content representing the strongest correlates, followed by epithelial organization rather than quantity. For predicting cancer among high and low fibroglandular volume, the classifier achieved AUCs of 0.92 and 0.84, respectively, with epithelial organizational features ranking most important. These results suggest non-fatty stroma, fat tissue quantities and epithelial region organization predict fibroglandular volume. The model holds promise for identifying histological correlates of cancer risk in patients with high and low density and warrants further evaluation.