Dataset Information

Deep learning assessment of breast terminal duct lobular unit involution: Towards automated prediction of breast cancer risk.

ABSTRACT: Terminal duct lobular unit (TDLU) involution is the regression of milk-producing structures in the breast. Women with less TDLU involution are more likely to develop breast cancer. A major bottleneck in studying TDLU involution in large cohort studies is the need for labor-intensive manual assessment of TDLUs. We developed a computational pathology solution to automatically capture TDLU involution measures. Whole slide images (WSIs) of benign breast biopsies were obtained from the Nurses' Health Study. A set of 92 WSIs was annotated for acini, TDLUs and adipose tissue to train deep convolutional neural network (CNN) models for detection of acini, and segmentation of TDLUs and adipose tissue. These networks were integrated into a single computational method to capture TDLU involution measures including number of TDLUs per tissue area, median TDLU span and median number of acini per TDLU. We validated our method on 40 additional WSIs by comparing with manually acquired measures. Our CNN models detected acini with an F1 score of 0.73±0.07, and segmented TDLUs and adipose tissue with Dice scores of 0.84±0.13 and 0.87±0.04, respectively. The inter-observer ICC scores for manual assessments on 40 WSIs of number of TDLUs per tissue area, median TDLU span, and median acini count per TDLU were 0.71, 0.81 and 0.73, respectively. Intra-observer reliability was evaluated on 10/40 WSIs with ICC scores of >0.8. Inter-observer ICC scores between automated results and the mean of the two observers were: 0.80 for number of TDLUs per tissue area, 0.57 for median TDLU span, and 0.80 for median acini count per TDLU. TDLU involution measures evaluated by manual and automated assessment were inversely associated with age and menopausal status. We developed a computational pathology method to measure TDLU involution. This technology eliminates the labor-intensiveness and subjectivity of manual TDLU assessment, and can be applied to future breast cancer risk studies.

SUBMITTER: Wetstein SC

PROVIDER: S-EPMC7159218 | biostudies-literature | 2020

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Deep learning assessment of breast terminal duct lobular unit involution: Towards automated prediction of breast cancer risk.

Wetstein Suzanne C SC Onken Allison M AM Luffman Christina C Baker Gabrielle M GM Pyle Michael E ME Kensler Kevin H KH Liu Ying Y Bakker Bart B Vlutters Ruud R van Leeuwen Marinus B MB Collins Laura C LC Schnitt Stuart J SJ Pluim Josien P W JPW Tamimi Rulla M RM Heng Yujing J YJ Veta Mitko M

PloS one 20200415 4

Terminal duct lobular unit (TDLU) involution is the regression of milk-producing structures in the breast. Women with less TDLU involution are more likely to develop breast cancer. A major bottleneck in studying TDLU involution in large cohort studies is the need for labor-intensive manual assessment of TDLUs. We developed a computational pathology solution to automatically capture TDLU involution measures. Whole slide images (WSIs) of benign breast biopsies were obtained from the Nurses' Health ...[more]

PMID: 32294107

Similar Datasets

Project description:BackgroundExposure to certain outdoor air pollutants may be associated with a higher risk of breast cancer, though potential underlying mechanisms are poorly understood. We examined whether outdoor air pollution was associated with involution of terminal duct lobular units (TDLUs), the histologic site where most cancers arise and an intermediate marker of breast cancer risk.MethodsPathologist-enumerated TDLUs were assessed in H&E (hematoxylin and eosin)-stained breast tissue sections from 1904 US women ages 18-75 who donated to the Susan G. Komen Tissue Bank (2009-2012). The 2009 annual fine particulate matter < 2.5 μm in diameter (PM2.5) total mass (μg/m3) at each woman's residential address was estimated from the Environmental Protection Agency's Downscaler Model combining Community Multiscale Air Quality (CMAQ) System modeling with air quality monitoring data. We secondarily considered CMAQ-modeled components of PM2.5 and gaseous pollutants. We used K-means clustering to identify groups of individuals with similar levels of PM2.5 components, selecting groups via cluster stability analysis. Relative rates (RRs) and 95% confidence intervals (95% CIs) for the association between air pollutants and TDLU counts were estimated from a zero-inflated negative binomial regression model adjusted for potential confounders.ResultsPM2.5 total mass was associated with higher TDLU counts among all women (interquartile range (IQR) increase, RR = 1.06; 95% CI: 1.01-1.11). This association was evident among both premenopausal and postmenopausal women (premenopausal RR = 1.05, 95% CI: 1.00-1.11; postmenopausal RR = 1.11, 95% CI: 1.00-1.23). We identified 3 groups corresponding to clusters that varied geographically and roughly represented high, medium, and low levels of PM2.5 components relative to population mean levels. Compared to the cluster with low levels, the clusters with both high (RR = 1.74; 95% CI: 1.08-2.80) and medium (RR = 1.82; 95% CI: 1.13-2.93) levels were associated with higher TDLU counts; although not significantly different, the magnitude of the associations was stronger among postmenopausal women.ConclusionsHigher PM2.5 levels were associated with reduced TDLU involution as measured by TDLU counts. Air pollution exposure may influence the histologic characteristics of normal tissue which could in turn affect breast cancer risk.

Dataset Information

Deep learning assessment of breast terminal duct lobular unit involution: Towards automated prediction of breast cancer risk.

Publications

Deep learning assessment of breast terminal duct lobular unit involution: Towards automated prediction of breast cancer risk.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets