The effects of menstrual cycle, site and individual variation on breast elasticity and thickness

Abstract

Ultrasound elasticity imaging using quasi-static compression has been widely evaluated in breast tumor diagnosis. Although it is promising in differentiating between benign and malignant breast tumors, factors such as semi-quantization without scaling the force applied and physiological changes of breast elasticity during the menstrual cycle may result in misdiagnosis. It has been reported that, due to female sex hormones, female breast morphology, proliferative and secretive activities, density, volume, electrical property and blood flow exhibit cyclic changes during the menstrual cycle. However, there is a lack of in vivo information about the menstrual cycle dependence of breast tissue elasticity changes. One recent study based on magnetic resonance elastography (MRE) unveiled the variations of breast elasticity during the menstrual cycle. However, the theory of MRE and the position of subjects in the examination are different compared with the measurements performed using ultrasound elastography. Additionally, the potential differences within the breast and among individuals have not been well considered in the previous studies about breast elasticity measurement in vivo. Using a 2-D ultrasound indentation system, the aims were to quantify the elasticity and thickness of breast tissues in vivo and to investigate the changes as a function of the menstrual cycle, different parts of a breast, and individual subjects. A specialized 2-D ultrasound probe with a force sensor attached was developed to record the force applied onto the tissue and the corresponding ultrasound images during manual loading and unloading processes (named 'indentation'). The initial tissue thickness and the deformation caused by indentation were extracted from the ultrasound images. Tissue elasticity was calculated from the force-deformation data. The good repeatability of the elasticity and thickness measurement in vivo (n=163, ICC>0.950, variability <4.9% for the elasticity, and <0.7% for the thickness) guaranteed the feasibility of repeated tests during menstrual cycles. 21 subjects were recruited for breast elasticity and thickness measurements at an interval of two to three days for one complete menstrual cycle. The measurements were conducted on the four quadrants of both breasts for each subject. Concerning each quadrant, five repeated tests were performed and the values were averaged. The menstrual cycle length of each subject was normalized to 28 days using the ovulation day predicted from the pelvis ultrasound scanning. Accordingly, the number of tests for each subject was interpolated to 28 data points by linear interpolation. The results indicated that the breast elasticity and thickness varied as a function of menstrual cycle days with a trough and a peak in the mid-cycle (Days 11-13) and before menstruation (Days 26-28), respectively. Breast elasticity and thickness in the proliferative phase were significantly smaller (p<0.05) than that in the menstrual phase and the secrectoy phase in the uterine event. The thickness was significantly smaller (p=0.013) in the follicular phase than the luteal phase in the ovary event, but not for the elasticity (p=0.069). All the individual quadrants showed similar trends in either elasticity or thickness during the menstrual cycle. There were significant differences among the four quadrants, between the upper and lower parts, and between the inner and outer parts of the breast for both the tissue thickness and elasticity (p<0.05). Breast thickness and elasticity were different among individual subjects, which might be caused by the potential differences of their physiological body indices such as BMI, breast size, breast pain, and handedness. It was found that a significant linear correlation existed between breast elasticity and BMI using Pearson's linear regression (n=26, r=0.546, p=0.004), and between breast thickness and BMI (n=26, r=0.871, p=0.000). A significant positive linear correlation (n=26, r=0.53, p=0.005) was also found between breast elasticity and thickness. Bra cup size, as an indicator of breast size, affected breast elasticity and thickness which showed significant difference among the bra cup sizes A (n=10), B (n=8) and C (n=8) (p<0.005). The breast was stiffer in the group with breast pain (n=15) than the group without breast pain (n=11), but the significant difference failed to be demonstrated in this study (p=0.165). In the strong right handed group, a significant difference was found in elasticity between the bilateral breasts (n=20, p=0.007). In conclusion, the within- and between-subject differences of breast elasticity and thickness were found. It is believed that the findings of site dependence and menstrual cycle dependence of breast tissue elasticity will provide a useful reference and guidance for the clinical application of ultrasound elasticity imaging. The ultrasound elastography is recommended to be scheduled before ovulation, about Days 11-13 in a cycle length of 28 days, when the best elasticity contrast between lesion and background emerges. The correlations between breast elasticity/thickness and the physiological body indices may potentially help increase the understanding of between-subject difference in breast elasticity and thickness in vivo.