Full Field Digital Mammography with Breast Cushions in Screening Subjects

To explore the impact of radiolucent breast cushions on breast pain and image quality of full field digital mammography (FFDM) in screening subjects. The American College of Radiology (ACR) phantom and fifty-four women were enrolled in this study. Use of cushions reduced pain (p < 0.001). The imaging quality with cushions was inferior to those without cushions in phantom studies (p ≦ 0.01), on CC views (p = 0.02) and on positioning (p < 0.05). Although use of cushions reduces pain during FFDM, these cushions potentially decreased imaging quality and might hinder breast positioning. More research needs to be done prior to the routine use of these cushions in clinical practice. Correspondence Author to: Yi-Fan Tsai Department of Medical Imaging, Kaohsiung Municipal Hsiao-Kang Hospital, Kaohsiung Medical University, Kaohsiung, Taiwan No. 482, San-Ming Road, Kaohsiung 812, Taiwan J Radiol Sci 2012; 37: 145-152 breast cushions in digital mammography 146 J Radiol Sci December 2012 Vol.37 No.4 screen-film mammography [16, 17]. One preliminary randomized study concluded that there was significantly less pain in the breast that received the cushions as compared to the uncushioned breast in screen-film mammography [18]. However, for 2% of the women, image quality was impaired when breast cushions were used [18]. To the best of our knowledge, no research has been performed to discover the application of these breast cushions in full field digital mammography (FFDM). The purpose of this study was to explore the use of these cushions in a randomized controlled trial with FFDM to assess the cushion’s impact on breast pain, image quality and radiation dose in the screening subjects. MAteRiAlS And MethodS Phantom study This study was approved by our institutional review board. The American College of Radiology (ACR) accreditation phantom for assessing image quality was divided into three groups: cushion 1 (28 kVp, 47 lb) with cushions (MammoPad; BioLucent Inc, Aliso Viejo, CA) on the compression paddle and bucky, cushion 2 (28 kVp, 34 lb) with cushions, and no cushion (28 kVp, 34 lb) as the control group. The compression force of cushion 1 was higher than those of cushion 2 and no cushion to assess whether greater compressive force improves the imaging quality. The ACR accreditation phantom approximates a 4.5cm compressed breast of average glandular/adipose composition with six different size nylon fibers which simulate fibrous structures, five groups of simulated microcalcifications, and five different size tumor-like masses. ACR criteria require a minimum score that includes visibility of the four largest fibrils, the three largest simulated microcaldification groups, and the three largest masses. Phantom images were obtained with W/Rh (target/ filter combination) by an experienced mammographical quality control radiographer (YFT) on a digital mammography (Lorad Selenia, Lorad/Hologic). Following each exposure of the phantom, the radiographer recoded the technical parameters, including compression force (lb), compressed thickness (cm), mAs, kVp and average glandular dose (AGD) (mGy) on the mammography unit (Selina base acquistion workstation). To determine the signal-to-noise ratio (SNR) and the contrast-to-noise ratio (CNR), a region-of-interest (ROI) cursor (about 0.25 cm2) was placed on the masses and next to the masses, and the signal intensity (SI) was recorded by the radiographer. SNR and CNR were calculated using the following equations: SNR= (SI mass –DC offset)/SD mass CNR= (SI background – SI mass)/SD background Where SI mass and SD mass are the mean SI and standard deviation of mass, DC offset is a DC offset added to the detector signal and is equal to 50, and SI background is the mean SI of background next to the mass. Subject population Women were approached by a research assistant after they had changed into a gown and were sitting in the waiting room provided for screening mammography patients only. During the recruiting phase, each potential subject was asked if she would be interested in participating in a brief study to determine the effects of breast cushions on pain and imaging quality. From October 2009 to September 2010, 54 asymptomatic women (50.72 ± 4.30 years, range, 45-61 years) were enrolled in this study. Women with breast implants and those having undergone breast surgery were excluded. After obtaining informed consent, each participant was asked to fill out a brief questionnaire prior to entering the mammography suite. From this, the participant’s age, hormone replacement therapy status, and prior experiences with mammography were recorded. Mammography Each woman had her mammography using a digital mammography (LoRad-Hologic, Selenia, USA). All mammographies were performed in the standard fashion utilizing craniocaudal (CC) and mediolateral oblique (MLO) views with W/Rh (target/ filter combination). One breast in the standard fashion used breast cushions on the compression paddle and bucky (Fig. 1). The opposite breast served as control without the use of cushions. Two parameters were randomized: which breast received the cushions and which breast was imaged first. This design was utilized so that women did not get a double dose of radiation during their screening mammography. The level of compression delivered was determined by experienced mammography radiographers adhering to proper standards. All radiographers had over 10 years experience and were certified in mammography by the Society of Radiological Technologists and Bureau of Health Promotion. Following each breast exposure, radiographers recoded the technical parameters, including breast compression force (lb), compressed breast thickness (cm), mAs, kVp and AGD (mGy) on the mammography unit (Selina base acquistion workstation) and the participant rated her breast pain on a 10-cm horizontal-line visual analogue scale (VAS) with “no pain” on the left and “severe pain” on the far right. In addition, the participant rated her breast pain on an 11-point numeric rating scale (NRS), starting at no pain (score 0) and progressing to severe pain (score 10). A blank section for additional comments was also provided for each participant. The percentage reduction in pain was calculated in the following manner: (VSA value without cushions-VSA value with cushions)/VSA value without cushions imaging analysis and statistics Two experienced radiologists (RSH, TJH) specializing breast cushions in digital mammography 147 J Radiol Sci December 2012 Vol.37 No.4 in breast imaging rated image quality of the mammographies independently on a digital mammography work station (Barco, Coronis/5MP Mammo) with constant window level and window width of 2047 and 4096. They were unaware of the cushion assignment during the image quality rating. They rated the image quality of each view on a 5-point scale (1, bad, no diagnosis possible; 2, poor, diagnostic confidence significantly reduced; 3, moderate, suboptimal exposure or little artifact but sufficient for diagnosis; 4, good, optimal exposure and good for diagnosis; and 5, excellent, good for diagnosis and good present anatomy on mammography) In addition, the imaging quality of positioning was rated as 0–10 as a numeric rating scale by two experienced mammographical quality control radiographers (CJC, IHY) according to the present anatomy on mammography and the adequacy of compression [19, 20] independently on the same mammography work station. This was done to test the influence of these cushions on breast positioning because the placement of the cushions on the compression paddle prevented the radiographers from viewing the breast directly during compression. They were unaware of the cushion assignment during the image quality rating. The fatty and scattered fibroglandular densities of breasts were classed as non-dense breast. The heterogeneously dense and extremely dense breasts were defined as dense breast. All data were expressed as means ± SD. One-way ANOVA (repeated measures) and paired t-tests were applied for exposure parameters (mAs, kVp and AGD) and imaging quality (SNR and CNR) of phantom among the cushion 1, cushion 2 and no cushion groups. Wilcoxon signed-rank test and Spearman rank correlation (rs) statistics were applied for inter-observer agreement of the imaging quality of mammography and the imaging quality of positioning. The imaging quality of mammography and the imaging quality of positioning were reported as means of measurements of two observers. Wilcoxon signed-rank test was used to compare the imaging quality of mammography, the imaging quality of positioning and breast pain (NRS) on cushion assignment. Paired t-test was applied to compare breast pain (VAS) cushion assignment. A level of significance of p < 0.05 was used.