Understanding mammographic breast density profile in China: A Sino‐Australian comparative study of breast density using real‐world data from cancer screening programs

AIM: This study aims at understanding mammographic density profile in China by comparing the density between women in China and Australia. METHODS: Data of 3250 women aged 45–69 were obtained from the Cancer Screening Program in Urban China and data of 1384 Australian counterparts at same age range were gathered from the Lifepool project. Demographic and reproductive details and mammograms for each cohort were collected. Mammographic density was assessed using AutoDensity, and two metrics, percentage density (PD) and dense area (DA), were applied. T‐tests were used to compare the means of mammographic density between two populations of all, premenopausal, and postmenopausal women. Two‐way ANOVA was conducted to examine interactions of population (Chinese/Australian) and each variable of interest upon mammographic density. RESULTS: Chinese women had 9.61%, 8.20%, and 9.28% higher PD than their Australian counterparts in all, premenopausal, and postmenopausal women, respectively (all p < 0.001). The mean differences in DA between two population were 1.81 cm(2) (p < 0.001), 0.55 cm(2) (p = 0.472), and 1.76 cm(2) (p = 0.003) for all, premenopausal, and postmenopausal women, respectively. There were significant interactions between population and age (F[4, 4624] = 4.12, p = 0.003), BMI (F[2, 4628] = 3.92, p = 0.020), age at first birth (F[1, 4250] = 11.69, p < 0.001), breastfeeding history (F[1, 4479] = 17.79, p < 0.001), and breastfeeding duration (F[1, 3526] = 66.90, p < 0.001) upon PD. Interaction was only found for breastfeeding history (F[1, 4479] = 4.79, p = 0.029) and breastfeeding duration (F[1, 3526] = 17.72, p < 0.001) for DA. CONCLUSIONS: Both PD and DA were found to be higher in Chinese women compared to Australian women. The density difference by menopause status was shown and breastfeeding history affected breast density differently in both populations.