Jpn J Radiol DOI 10.1007/s11604-014-0333-x
TECHNICAL NOTE
Feasibility study of a breast density measurement within a direct photon-counting mammography scanner system Youichi Machida • Mitsuhiro Tozaki • Tamiko Yoshida • Ai Saita • Mari Yakabe Kanae Nii
•
Received: 18 February 2014 / Accepted: 9 May 2014 Ó Japan Radiological Society 2014
Abstract Purpose To evaluate the clinical feasibility of breast density measurements by a new application within a direct photon-counting mammography scanner system. Materials and methods A retrospective study of consecutive women who underwent mammography using a direct photon-counting mammography scanner system (MicroDose mammography SI; Philips Digital Mammography Sweden AB) was performed at the authors’ institution between September and December 2013. Quantitative volumetric glandularity measurements were performed automatically for each acquired mammographic image using an application (Breast Density Measurement; Philips Digital Mammography Sweden AB). The quantitative volumetric glandularity of each breast was defined as the average values for the mediolateral oblique (MLO) and craniocaudal (CC) mammogram views. Results Of the 44 women who underwent bilateral mammogram acquisitions, the breast density measurements Y. Machida (&) M. Tozaki T. Yoshida A. Saita Diagnostic Imaging Center, Kameda Kyobashi Clinic, Tokyo Square Garden 4F, 3-1-1 Kyobashi, Chuo-ku, Tokyo 104-0031, Japan e-mail: [email protected] Y. Machida Department of Diagnostic Radiology and Oncology, Tokyo Medical and Dental University, Graduate School of Medicine, Bunkyo-ku, Tokyo 113-8519, Japan M. Tozaki Breast Center, Kameda Medical Center, 929 Higashi-cho, Kamogawa City, Chiba 296-8602, Japan M. Yakabe K. Nii Philips Electronics Japan, Ltd., Minato-ku, Tokyo 108-8507, Japan
were performed successfully in 40 patients (90.9 %). A very good to excellent correlation in the quantitative breast density measurements acquired from the MLO and CC images was obtained in the 40 evaluable patients (R = 0.99). Conclusion The calculated volumetric glandularity using this new application should correspond well with the true volumetric density of each breast. Keywords Mammography Breast density measurement Application BI-RADS
Introduction Breast density has been reported to be a risk factor for breast cancer. Previous studies from Western countries have shown that women with dense tissue in 75 % or more of their breasts have a risk of breast cancer that is four to six times as great as the risk among women with little or no dense tissue [1–9]. Reports from Asian countries, mainly from Japan, have also identified dense breast tissue as a risk factor for breast cancer [10–12]. Quantitative methods of assessing fibroglandular tissue in the breast have been developed. The first generation of these methods involved the review of 2D projection images of compressed breasts [13, 14]. Because 2D methods have several disadvantages, such as a lack of consideration of breast thickness, their accuracy was limited. While more recently proposed methods have attempted to measure the fibroglandular and adipose thickness in each pixel and enable an increased accuracy [15–18], they still bore measurement errors for volumetric breast density. Studies have shown that the accurate measurement of breast density can be achieved using dual-energy imaging [19, 20],
123
Jpn J Radiol
although the additional acquisition of images using highenergy X-rays were indispensable for this measurement, resulting in an increased glandular dose as well as the risk of misregistration artifacts produced by patient movements. A direct photon-counting mammography scanner system (MicroDose mammography SI; Philips Digital Mammography Sweden AB, Solna, Sweden), which is a scanning multi-slit full-field digital mammography system with a photon-counting silicon stripe detector, is capable of sorting photons into low- or high-energy categories, acquiring spectral energy data without additional dose exposures or the risk of misregistration, and enabling the volumetric glandularity of each breast to be calculated automatically by a new application within the system (Breast Density Measurement; Philips Digital Mammography Sweden AB, Solna, Sweden). Although a high accuracy of quantitative breast density measurements by this new application using computer simulations and physical phantoms have been reported [21], the clinical feasibility or validation of this new technology has not yet been reported. The purposes of this study were to evaluate the clinical feasibility of breast density measurements and to assess its correlation with visual assessments performed by a physician.
Materials and methods Patients A retrospective study of 45 consecutive women who underwent mammography using a direct photon-counting mammography scanner system (MicroDose mammography SI; Philips Digital Mammography Sweden AB) was performed at the authors’ institution between September and December 2013. One woman was excluded from the analysis because a comparison between the right and left breasts was unavailable because of a previous unilateral mastectomy for breast cancer. The remaining 44 women who underwent bilateral mammogram acquisitions were included in this study. Correlation of breast density measurements between mediolateral oblique (MLO) and craniocaudal (CC) images Both MLO and CC mammogram views were acquired for each woman. Quantitative volumetric glandularity measurements were performed automatically without any additional procedures for each acquired mammographic image using an application (Breast Density Measurement; Philips Digital Mammography Sweden AB). The breast density data were stored in the Digital Imaging and Communication in Medicine (DICOM) header and as a DICOM
123
Structured Report and were displayed in the mammography diagnosis monitors. The quantitative volumetric glandularity of each patient was compared between the MLO and CC views of the mammograms. Correlation between age and quantitative breast density measurements The quantitative volumetric glandularity of each breast was defined as the average values for the MLO and CC mammogram views. The quantitative volumetric glandularity of each patient was then compared according to age. Correlation between quantitative measurements and visual assessments Qualitative mammographic breast density assessments for each woman were performed by one radiologist (who had 7 years of experience in diagnostic radiology, who was engaged in breast screening and diagnosis, and who had an AS-ranked license in mammography interpretation guaranteed by The Japan Central Committee on Quality Assurance of Breast Cancer Screening, a nonprofit organization) who was blinded to the quantitative volumetric glandularity or any other clinical information. The qualitative visual mammographic breast density assessments were performed according to the Breast Imaging-Reporting and Data System (BI-RADS) reporting criteria for breast density [22] as follows: BI-RADS 1, the breast is almost entirely fat (\25 % glandular); BI-RADS 2, there are scattered fibroglandular densities (*25–50 % glandular); BI-RADS 3, the breast tissue is heterogeneously dense, which could obscure detection of small masses (*51–75 % glandular); BI-RADS 4, the breast tissue is extremely dense. This may lower the sensitivity of the mammography findings ([75 % glandular). The quantitative volumetric glandularity of each breast was defined as the average values of the MLO and CC images. The average of the quantitative volumetric glandularity for each BI-RADS breast density score group was calculated. Statistical analysis The correlation coefficient was calculated for the results of the quantitative volumetric glandularity measurements acquired from MLO and CC images of each breast and for the age and quantitative volumetric glandularity measurements. When interpreting the value of the correlation coefficient, R values from 0 to 0.25 were regarded as indicating the absence of a correlation, R values from 0.25 to 0.50 were regarded as indicating a poor correlation, R values ranging from 0.50 to 0.75 were regarded as
Jpn J Radiol
indicating a moderate to good correlation, and R values from 0.75 to 1.0 were regarded as indicating a very good to excellent correlation between the variables [23]. The statistical analysis was performed using Microsoft Excel 2007 software (Microsoft Corp, Redmond, WA, USA).
Results Patient populations Out of the 44 women who underwent bilateral mammogram acquisitions, comprehensive breast density measurements were unavailable in four patients for the following reasons: in one patient, the thickness of the breast compressed for image acquisition was below 20 mm, which was the lower limit of the breast density measurement software; in three patients, the breast density measurements for the MLO view was unavailable after successful image acquisition. Among these latter three patients, the breast density measurement of the MLO view was unsuccessful because the unilateral pectoralis major muscle was not within the field of view after positioning in two patients, and the breast glandular tissue overlapped with the unilateral pectoralis major muscle in one patient. The breast density measurements were performed successfully in the remaining 40 of the 44 patients (90.9 %). The average glandular dose of these mammographic images ranged from 0.42 to 1.45 mGy (mean 0.74 mGy).
Fig. 1 Correlation of breast density measurements between MLO and CC images. A very good to excellent correlation in the quantitative breast density measurements acquired from the MLO and CC images was obtained in the 40 evaluable patients (R = 0.99)
Correlation of breast density measurements between MLO and CC images A very good to excellent correlation in the quantitative breast density measurements acquired from the MLO and CC images was obtained in the 40 evaluable patients (R = 0.99, Fig. 1). Correlation between age and quantitative breast density measurements The relationship between the quantitative volumetric glandularity of each breast and the age of the 40 patients is shown in Fig. 2. A poor correlation between the two variables was seen (R = –0.34).
Fig. 2 Correlation between age and quantitative breast density measurements. A poor correlation between the two variables was seen (R = –0.34)
(17–54 %, n = 36), and 64.3 % for BI-RADS 4 (37–85 %, n = 32) (Figs. 5, 6), respectively. Distribution of volumetric glandularity for each BI-RADS breast density score is shown in Fig. 7.
Discussion Correlation between quantitative measurements and visual assessments The average quantitative volumetric glandularity of each BI-RADS breast density score group was 9 % for BIRADS 1 (n = 2) (Fig. 3), 21.7 % for BI-RADS 2 (14–38 %, n = 10) (Fig. 4), 34.3 % for BI-RADS 3
Venturini et al. [24] evaluated a breast-screening program for 40- to 49-year-old women living in the town of Segrate, Milan, Italy, using a previous version of the MicroDose mammography system (with the same dose reduction ability, but without the availability of breast density measurements) by Sectra Mamea (Solna, Sweden), which is
123
Jpn J Radiol Fig. 3 A 69-year-old woman with breasts that were assessed to be almost entirely fat (BIRADS 1). Volumetric glandularity was calculated to be 9 % in both breasts. The average glandular dose was 0.60 mGy
Fig. 4 A 40-year-old woman with breasts that were assessed to have scattered fibroglandular densities (BI-RADS 2). Volumetric glandularity was calculated to be 38 % in the right breast and 35 % in the left breast. The average glandular dose was 0.50 mGy. The volumetric glandularity was exceptionally high in this case among the BI-RADS 2 cases
now part of Philips Women’s Healthcare. In their study, they reported that the average glandular doses for CC and MLO projections were 0.73 mGy (range 0.33–1.90 mGy) and 0.76 mGy (range 0.33–1.90 mGy), respectively. In this study, our results regarding the average glandular dose were comparable to those of this previous study, indicating that a similar dose reduction can be achieved among both Asian and European populations. In previous studies, the breast density value was defined as the average of results from MLO and CC images [25, 26], and this new application for breast density measurement also adopts this consensus. However, no previous studies have assessed the clinical feasibility of this new technology or provided any validation of the technology; consequently, the validity of adopting the average value as a representative value for each breast has not been investigated.
123
In this study, 9.1 % (4/44) of the patients were excluded from the breast density measurement analysis even though mammographic images were acquired, and the MLO breast density measurement was unavailable in 75 % (3/4) of these cases. Furthermore, the excellent correlation of the quantitative breast density measurements between the values acquired from the MLO and CC images of each breast (R = 0.99) indicates that breast density information acquired from only the MLO or CC image and that calculated from the average of the MLO and CC images are comparable. Given the reported high accuracy of this system using computer simulations and physical phantoms [21], this excellent correlation also indicates that the calculated value, whether it is from a single MLO or CC image or an average of the two images, seems to correspond well with the true volumetric density of each breast. Even though adopting the calculated value from the CC
Jpn J Radiol Fig. 5 A 28-year-old woman with breasts that were assessed to be extremely dense (BIRADS 4). Volumetric glandularity was calculated to be 59 % in the right breast and 58 % in the left breast. The average glandular dose was 1.10 mGy
Fig. 6 A 64-year-old woman with breasts that were assessed to be extremely dense (BIRADS 4). Volumetric glandularity was calculated to be 37 % in the right breast and 41 % in the left breast. The average glandular dose was 0.57 mGy
image may be promising, MLO images are more widely used for screening mammography. Therefore, adopting the average values acquired from the MLO and CC mammogram views may be an appropriate procedure. Although breast density is known to decrease with age [27], Asian women are more likely to have dense breasts than white women, and this tendency becomes more obvious in older populations [28]. In this study, there was a tendency for breast density to be high, and the tendency
persisted even in older patients, though there was a weak tendency for breast density to decrease with age. These results are in accordance with those of previous studies. The average quantitative volumetric glandularity for each BI-RADS breast density score group was 9 % for BIRADS 1, 21.7 % for BI-RADS 2, 34.3 % for BI-RADS 3, and 64.3 % for BI-RADS 4. These results indicate that, with an exceptional case as shown in Fig. 4, this new method of volumetric breast density measurement is
123
Jpn J Radiol
4.
5.
6.
7.
8.
9. Fig. 7 Box-and-whisker plot of volumetric glandularity was estimated using the BI-RADS breast density score
10.
11.
thought to be comparable with the qualitative visual assessment, which is the current standard for breast density assessments. However, we assessed the correlation between quantitative measurements and qualitative visual assessments for breast density in only 40 patients, and this small population was a serious limitation of the present study. We conducted a preliminary study of this new application, which has not been previously reported. This new application is easy to use without physicians’ evaluation and may have possibilities to select women for whom screening ultrasound will be beneficial. The significance of breast density measurements using a direct photon-counting mammography scanner system for clinical concerns (e.g., its influence on breast cancer risk or the sensitivity of mammography results) needs to be evaluated in a larger population in the future.
12.
13.
14.
15.
16.
17.
18. Acknowledgments
There is no grant support.
Conflict of interest
The authors declare no conflict of interest. 19.
References 1. Boyd NF, Guo H, Martin LJ, Sun L, Stone J, Fishell E, et al. Mammographic density and the risk and detection of breast cancer. N Engl J Med. 2007;356:227–36. 2. Ursin G, Ma H, Wu AH, Bernstein L, Salane M, Parisky YR, et al. Mammographic density and breast cancer in three ethnic groups. Cancer Epidemiol Biomark Prev. 2003;12:332–8. 3. Byrne C, Schairer C, Wolfe J, Parekh N, Salane M, Brinton LA, et al. Mammographic features and breast cancer risk: effects with
123
20.
21.
22.
time, age, and menopause status. J Natl Cancer Inst. 1995;87:1622–9. Boyd NF, Byng JW, Jong RA, Fishell EK, Little LE, Miller AB, et al. Quantitative classification of mammographic densities and breast cancer risk: results from the Canadian National Breast Screening Study. J Natl Cancer Inst. 1995;87:670–5. Wolfe JN, Saftlas AF, Salane M. Mammographic parenchymal patterns and quantitative evaluation of mammographic densities: a case–control study. Am J Roentgenol. 1987;148:1087–92. Byrne C, Schairer C, Brinton LA, Wolfe J, Parekh N, Salane M, et al. Effects of mammographic density and benign breast disease on breast cancer risk (United States). Cancer Causes Control. 2001;12:103–10. Harvey JA, Bovbjerg VE. Quantitative assessment of mammographic breast density: relationship with breast cancer risk. Radiology. 2004;230:29–41. Boyd NF, Rommens JM, Vogt K, Lee V, Hopper JL, Yaffe MJ, et al. Mammographic breast density as an intermediate phenotype for breast cancer. Lancet Oncol. 2005;6:798–808. McCormack VA, dos Santos Silva I. Breast density and parenchymal patterns as markers of breast cancer risk: a meta-analysis. Cancer Epidemiol Biomark Prev. 2006;15:1159–69. Nagata C, Matsubara T, Fujita H, Nagao Y, Shibuya C, Kashiki Y, et al. Mammographic density and the risk of breast cancer in Japanese women. Br J Cancer. 2005;92:2102–6. Kotsuma Y, Tamaki Y, Nishimura T, Tsubai M, Ueda S, Shimazu K, et al. Quantitative assessment of mammographic density and breast cancer risk for Japanese women. Breast. 2008;17:27–35 Epub 2007 Aug 22. Wong CS, Lim GH, Gao F, Jakes RW, Offman J, Chia KS, et al. Mammographic density and its interaction with other breast cancer risk factors in an Asian population. Br J Cancer. 2011;2011(104):871–4. doi:10.1038/sj.bjc.6606085.Epub. Byng JW, Boyd NF, Fishell E, Jong RA, Yaffe MJ. The quantitative-analysis of mammographic densities. Phys Med Biol. 1994;39:1629–38. Sivaramakrishna R, Obuchowski NA, Chilcote WA, Powell KA. Automatic segmentation of mammographic density. Acad Radiol. 2001;8:250–6. Highnam R, Pan X, Warren R, Jeffreys M, Smith GD, Brady M. Breast composition measurements using retrospective standard mammogram form (SMF). Phys Med Biol. 2006;51:2695–713. Highnam R, Jeffreys M, McCormack V, Warren R, Smith GD, Brady M. Comparing measurements of breast density. Phys Med Biol. 2007;52:5881–95. Kaufhold S, Dohrmann R, Ufer K, Meyer FM. Comparison of methods for the quantification of montmorillonite in bentonites. Appl Clay Sci. 2002;22:145–51. Pawluczyk O, Augustine BJ, Yaffe MJ, Rico D, Yang JW, Mawdsley GE, et al. A volumetric method for estimation of breast density on digitized screen-film mammograms. Med Phys. 2003;30:352–64. Ducote JL, Molloi S. Quantification of breast density with dual energy mammography: a simulation study. Med Phys. 2008;35:5411–8. Ducote JL, Molloi S. A quantification of breast density with dual energy mammography: an experimental feasibility study. Med Phys. 2010;37(793–801):24. Ding H, Molloi S. Quantification of breast density with spectral mammography based on a scanned multi-slit photon-counting detector: a feasibility study. Phys Med Biol. 2012;2012(57):4719–38. doi:10.1088/0031-9155/57/15/4719. Epub. D’Orsi CJ, Vassett JW, Berg WA, Feig SA, Jackson VP, Kopans DB, et al. ACR BI-RADS––mammography. 4th ed. In: ACR
Jpn J Radiol breast imaging reporting and data system, breast imaging atlas. Reston: American College of Radiology; 2003. 23. Research questions about relationships among variables. In: Dawson B, Trapp RG, editors. Basic and clinical biostatistics, Chapter 8. 4th ed. New York: McGrawHill; 2004. pp. 190–220. 24. Venturini E, Losio C, Panizza P, Rodighiero MG, Fedele I, Tacchini S, et al. Tailored breast cancer screening program with microdose mammography, US, and MR imaging: short-term results of a pilot study in 40–49-year-old women. Radiology. 2013;2013(268):347–55. doi:10.1148/radiol.13122278.Epub. 25. Heine JJ, Carston MJ, Scott CG, Brandt KR, Wu FF, Pankratz VS, et al. An automated approach for estimation of breast density. Cancer Epidemiol Biomark Prev. 2008;17:3090–7. doi:10.1158/ 1055-9965.EPI-08-0170.
26. Jeffreys M, Warren R, Highnam R, Smith GD. Initial experiences of using an automated volumetric measure of breast density: the standard mammogram form. Br J Radiol. 2006;79:378–82. 27. El-Bastawissi AY, White E, Mandelson MT, Taplin SH. Reproductive and hormonal factors associated with mammographic breast density by age (United States). Cancer Causes Control. 2000;11(10):955–63. 28. El-Bastawissi AY, White E, Mandelson MT, Taplin S. Variation in mammographic breast density by race. Ann Epidemiol. 2001;11(4):257–63.
123
TECHNICAL NOTE
Feasibility study of a breast density measurement within a direct photon-counting mammography scanner system Youichi Machida • Mitsuhiro Tozaki • Tamiko Yoshida • Ai Saita • Mari Yakabe Kanae Nii
•
Received: 18 February 2014 / Accepted: 9 May 2014 Ó Japan Radiological Society 2014
Abstract Purpose To evaluate the clinical feasibility of breast density measurements by a new application within a direct photon-counting mammography scanner system. Materials and methods A retrospective study of consecutive women who underwent mammography using a direct photon-counting mammography scanner system (MicroDose mammography SI; Philips Digital Mammography Sweden AB) was performed at the authors’ institution between September and December 2013. Quantitative volumetric glandularity measurements were performed automatically for each acquired mammographic image using an application (Breast Density Measurement; Philips Digital Mammography Sweden AB). The quantitative volumetric glandularity of each breast was defined as the average values for the mediolateral oblique (MLO) and craniocaudal (CC) mammogram views. Results Of the 44 women who underwent bilateral mammogram acquisitions, the breast density measurements Y. Machida (&) M. Tozaki T. Yoshida A. Saita Diagnostic Imaging Center, Kameda Kyobashi Clinic, Tokyo Square Garden 4F, 3-1-1 Kyobashi, Chuo-ku, Tokyo 104-0031, Japan e-mail: [email protected] Y. Machida Department of Diagnostic Radiology and Oncology, Tokyo Medical and Dental University, Graduate School of Medicine, Bunkyo-ku, Tokyo 113-8519, Japan M. Tozaki Breast Center, Kameda Medical Center, 929 Higashi-cho, Kamogawa City, Chiba 296-8602, Japan M. Yakabe K. Nii Philips Electronics Japan, Ltd., Minato-ku, Tokyo 108-8507, Japan
were performed successfully in 40 patients (90.9 %). A very good to excellent correlation in the quantitative breast density measurements acquired from the MLO and CC images was obtained in the 40 evaluable patients (R = 0.99). Conclusion The calculated volumetric glandularity using this new application should correspond well with the true volumetric density of each breast. Keywords Mammography Breast density measurement Application BI-RADS
Introduction Breast density has been reported to be a risk factor for breast cancer. Previous studies from Western countries have shown that women with dense tissue in 75 % or more of their breasts have a risk of breast cancer that is four to six times as great as the risk among women with little or no dense tissue [1–9]. Reports from Asian countries, mainly from Japan, have also identified dense breast tissue as a risk factor for breast cancer [10–12]. Quantitative methods of assessing fibroglandular tissue in the breast have been developed. The first generation of these methods involved the review of 2D projection images of compressed breasts [13, 14]. Because 2D methods have several disadvantages, such as a lack of consideration of breast thickness, their accuracy was limited. While more recently proposed methods have attempted to measure the fibroglandular and adipose thickness in each pixel and enable an increased accuracy [15–18], they still bore measurement errors for volumetric breast density. Studies have shown that the accurate measurement of breast density can be achieved using dual-energy imaging [19, 20],
123
Jpn J Radiol
although the additional acquisition of images using highenergy X-rays were indispensable for this measurement, resulting in an increased glandular dose as well as the risk of misregistration artifacts produced by patient movements. A direct photon-counting mammography scanner system (MicroDose mammography SI; Philips Digital Mammography Sweden AB, Solna, Sweden), which is a scanning multi-slit full-field digital mammography system with a photon-counting silicon stripe detector, is capable of sorting photons into low- or high-energy categories, acquiring spectral energy data without additional dose exposures or the risk of misregistration, and enabling the volumetric glandularity of each breast to be calculated automatically by a new application within the system (Breast Density Measurement; Philips Digital Mammography Sweden AB, Solna, Sweden). Although a high accuracy of quantitative breast density measurements by this new application using computer simulations and physical phantoms have been reported [21], the clinical feasibility or validation of this new technology has not yet been reported. The purposes of this study were to evaluate the clinical feasibility of breast density measurements and to assess its correlation with visual assessments performed by a physician.
Materials and methods Patients A retrospective study of 45 consecutive women who underwent mammography using a direct photon-counting mammography scanner system (MicroDose mammography SI; Philips Digital Mammography Sweden AB) was performed at the authors’ institution between September and December 2013. One woman was excluded from the analysis because a comparison between the right and left breasts was unavailable because of a previous unilateral mastectomy for breast cancer. The remaining 44 women who underwent bilateral mammogram acquisitions were included in this study. Correlation of breast density measurements between mediolateral oblique (MLO) and craniocaudal (CC) images Both MLO and CC mammogram views were acquired for each woman. Quantitative volumetric glandularity measurements were performed automatically without any additional procedures for each acquired mammographic image using an application (Breast Density Measurement; Philips Digital Mammography Sweden AB). The breast density data were stored in the Digital Imaging and Communication in Medicine (DICOM) header and as a DICOM
123
Structured Report and were displayed in the mammography diagnosis monitors. The quantitative volumetric glandularity of each patient was compared between the MLO and CC views of the mammograms. Correlation between age and quantitative breast density measurements The quantitative volumetric glandularity of each breast was defined as the average values for the MLO and CC mammogram views. The quantitative volumetric glandularity of each patient was then compared according to age. Correlation between quantitative measurements and visual assessments Qualitative mammographic breast density assessments for each woman were performed by one radiologist (who had 7 years of experience in diagnostic radiology, who was engaged in breast screening and diagnosis, and who had an AS-ranked license in mammography interpretation guaranteed by The Japan Central Committee on Quality Assurance of Breast Cancer Screening, a nonprofit organization) who was blinded to the quantitative volumetric glandularity or any other clinical information. The qualitative visual mammographic breast density assessments were performed according to the Breast Imaging-Reporting and Data System (BI-RADS) reporting criteria for breast density [22] as follows: BI-RADS 1, the breast is almost entirely fat (\25 % glandular); BI-RADS 2, there are scattered fibroglandular densities (*25–50 % glandular); BI-RADS 3, the breast tissue is heterogeneously dense, which could obscure detection of small masses (*51–75 % glandular); BI-RADS 4, the breast tissue is extremely dense. This may lower the sensitivity of the mammography findings ([75 % glandular). The quantitative volumetric glandularity of each breast was defined as the average values of the MLO and CC images. The average of the quantitative volumetric glandularity for each BI-RADS breast density score group was calculated. Statistical analysis The correlation coefficient was calculated for the results of the quantitative volumetric glandularity measurements acquired from MLO and CC images of each breast and for the age and quantitative volumetric glandularity measurements. When interpreting the value of the correlation coefficient, R values from 0 to 0.25 were regarded as indicating the absence of a correlation, R values from 0.25 to 0.50 were regarded as indicating a poor correlation, R values ranging from 0.50 to 0.75 were regarded as
Jpn J Radiol
indicating a moderate to good correlation, and R values from 0.75 to 1.0 were regarded as indicating a very good to excellent correlation between the variables [23]. The statistical analysis was performed using Microsoft Excel 2007 software (Microsoft Corp, Redmond, WA, USA).
Results Patient populations Out of the 44 women who underwent bilateral mammogram acquisitions, comprehensive breast density measurements were unavailable in four patients for the following reasons: in one patient, the thickness of the breast compressed for image acquisition was below 20 mm, which was the lower limit of the breast density measurement software; in three patients, the breast density measurements for the MLO view was unavailable after successful image acquisition. Among these latter three patients, the breast density measurement of the MLO view was unsuccessful because the unilateral pectoralis major muscle was not within the field of view after positioning in two patients, and the breast glandular tissue overlapped with the unilateral pectoralis major muscle in one patient. The breast density measurements were performed successfully in the remaining 40 of the 44 patients (90.9 %). The average glandular dose of these mammographic images ranged from 0.42 to 1.45 mGy (mean 0.74 mGy).
Fig. 1 Correlation of breast density measurements between MLO and CC images. A very good to excellent correlation in the quantitative breast density measurements acquired from the MLO and CC images was obtained in the 40 evaluable patients (R = 0.99)
Correlation of breast density measurements between MLO and CC images A very good to excellent correlation in the quantitative breast density measurements acquired from the MLO and CC images was obtained in the 40 evaluable patients (R = 0.99, Fig. 1). Correlation between age and quantitative breast density measurements The relationship between the quantitative volumetric glandularity of each breast and the age of the 40 patients is shown in Fig. 2. A poor correlation between the two variables was seen (R = –0.34).
Fig. 2 Correlation between age and quantitative breast density measurements. A poor correlation between the two variables was seen (R = –0.34)
(17–54 %, n = 36), and 64.3 % for BI-RADS 4 (37–85 %, n = 32) (Figs. 5, 6), respectively. Distribution of volumetric glandularity for each BI-RADS breast density score is shown in Fig. 7.
Discussion Correlation between quantitative measurements and visual assessments The average quantitative volumetric glandularity of each BI-RADS breast density score group was 9 % for BIRADS 1 (n = 2) (Fig. 3), 21.7 % for BI-RADS 2 (14–38 %, n = 10) (Fig. 4), 34.3 % for BI-RADS 3
Venturini et al. [24] evaluated a breast-screening program for 40- to 49-year-old women living in the town of Segrate, Milan, Italy, using a previous version of the MicroDose mammography system (with the same dose reduction ability, but without the availability of breast density measurements) by Sectra Mamea (Solna, Sweden), which is
123
Jpn J Radiol Fig. 3 A 69-year-old woman with breasts that were assessed to be almost entirely fat (BIRADS 1). Volumetric glandularity was calculated to be 9 % in both breasts. The average glandular dose was 0.60 mGy
Fig. 4 A 40-year-old woman with breasts that were assessed to have scattered fibroglandular densities (BI-RADS 2). Volumetric glandularity was calculated to be 38 % in the right breast and 35 % in the left breast. The average glandular dose was 0.50 mGy. The volumetric glandularity was exceptionally high in this case among the BI-RADS 2 cases
now part of Philips Women’s Healthcare. In their study, they reported that the average glandular doses for CC and MLO projections were 0.73 mGy (range 0.33–1.90 mGy) and 0.76 mGy (range 0.33–1.90 mGy), respectively. In this study, our results regarding the average glandular dose were comparable to those of this previous study, indicating that a similar dose reduction can be achieved among both Asian and European populations. In previous studies, the breast density value was defined as the average of results from MLO and CC images [25, 26], and this new application for breast density measurement also adopts this consensus. However, no previous studies have assessed the clinical feasibility of this new technology or provided any validation of the technology; consequently, the validity of adopting the average value as a representative value for each breast has not been investigated.
123
In this study, 9.1 % (4/44) of the patients were excluded from the breast density measurement analysis even though mammographic images were acquired, and the MLO breast density measurement was unavailable in 75 % (3/4) of these cases. Furthermore, the excellent correlation of the quantitative breast density measurements between the values acquired from the MLO and CC images of each breast (R = 0.99) indicates that breast density information acquired from only the MLO or CC image and that calculated from the average of the MLO and CC images are comparable. Given the reported high accuracy of this system using computer simulations and physical phantoms [21], this excellent correlation also indicates that the calculated value, whether it is from a single MLO or CC image or an average of the two images, seems to correspond well with the true volumetric density of each breast. Even though adopting the calculated value from the CC
Jpn J Radiol Fig. 5 A 28-year-old woman with breasts that were assessed to be extremely dense (BIRADS 4). Volumetric glandularity was calculated to be 59 % in the right breast and 58 % in the left breast. The average glandular dose was 1.10 mGy
Fig. 6 A 64-year-old woman with breasts that were assessed to be extremely dense (BIRADS 4). Volumetric glandularity was calculated to be 37 % in the right breast and 41 % in the left breast. The average glandular dose was 0.57 mGy
image may be promising, MLO images are more widely used for screening mammography. Therefore, adopting the average values acquired from the MLO and CC mammogram views may be an appropriate procedure. Although breast density is known to decrease with age [27], Asian women are more likely to have dense breasts than white women, and this tendency becomes more obvious in older populations [28]. In this study, there was a tendency for breast density to be high, and the tendency
persisted even in older patients, though there was a weak tendency for breast density to decrease with age. These results are in accordance with those of previous studies. The average quantitative volumetric glandularity for each BI-RADS breast density score group was 9 % for BIRADS 1, 21.7 % for BI-RADS 2, 34.3 % for BI-RADS 3, and 64.3 % for BI-RADS 4. These results indicate that, with an exceptional case as shown in Fig. 4, this new method of volumetric breast density measurement is
123
Jpn J Radiol
4.
5.
6.
7.
8.
9. Fig. 7 Box-and-whisker plot of volumetric glandularity was estimated using the BI-RADS breast density score
10.
11.
thought to be comparable with the qualitative visual assessment, which is the current standard for breast density assessments. However, we assessed the correlation between quantitative measurements and qualitative visual assessments for breast density in only 40 patients, and this small population was a serious limitation of the present study. We conducted a preliminary study of this new application, which has not been previously reported. This new application is easy to use without physicians’ evaluation and may have possibilities to select women for whom screening ultrasound will be beneficial. The significance of breast density measurements using a direct photon-counting mammography scanner system for clinical concerns (e.g., its influence on breast cancer risk or the sensitivity of mammography results) needs to be evaluated in a larger population in the future.
12.
13.
14.
15.
16.
17.
18. Acknowledgments
There is no grant support.
Conflict of interest
The authors declare no conflict of interest. 19.
References 1. Boyd NF, Guo H, Martin LJ, Sun L, Stone J, Fishell E, et al. Mammographic density and the risk and detection of breast cancer. N Engl J Med. 2007;356:227–36. 2. Ursin G, Ma H, Wu AH, Bernstein L, Salane M, Parisky YR, et al. Mammographic density and breast cancer in three ethnic groups. Cancer Epidemiol Biomark Prev. 2003;12:332–8. 3. Byrne C, Schairer C, Wolfe J, Parekh N, Salane M, Brinton LA, et al. Mammographic features and breast cancer risk: effects with
123
20.
21.
22.
time, age, and menopause status. J Natl Cancer Inst. 1995;87:1622–9. Boyd NF, Byng JW, Jong RA, Fishell EK, Little LE, Miller AB, et al. Quantitative classification of mammographic densities and breast cancer risk: results from the Canadian National Breast Screening Study. J Natl Cancer Inst. 1995;87:670–5. Wolfe JN, Saftlas AF, Salane M. Mammographic parenchymal patterns and quantitative evaluation of mammographic densities: a case–control study. Am J Roentgenol. 1987;148:1087–92. Byrne C, Schairer C, Brinton LA, Wolfe J, Parekh N, Salane M, et al. Effects of mammographic density and benign breast disease on breast cancer risk (United States). Cancer Causes Control. 2001;12:103–10. Harvey JA, Bovbjerg VE. Quantitative assessment of mammographic breast density: relationship with breast cancer risk. Radiology. 2004;230:29–41. Boyd NF, Rommens JM, Vogt K, Lee V, Hopper JL, Yaffe MJ, et al. Mammographic breast density as an intermediate phenotype for breast cancer. Lancet Oncol. 2005;6:798–808. McCormack VA, dos Santos Silva I. Breast density and parenchymal patterns as markers of breast cancer risk: a meta-analysis. Cancer Epidemiol Biomark Prev. 2006;15:1159–69. Nagata C, Matsubara T, Fujita H, Nagao Y, Shibuya C, Kashiki Y, et al. Mammographic density and the risk of breast cancer in Japanese women. Br J Cancer. 2005;92:2102–6. Kotsuma Y, Tamaki Y, Nishimura T, Tsubai M, Ueda S, Shimazu K, et al. Quantitative assessment of mammographic density and breast cancer risk for Japanese women. Breast. 2008;17:27–35 Epub 2007 Aug 22. Wong CS, Lim GH, Gao F, Jakes RW, Offman J, Chia KS, et al. Mammographic density and its interaction with other breast cancer risk factors in an Asian population. Br J Cancer. 2011;2011(104):871–4. doi:10.1038/sj.bjc.6606085.Epub. Byng JW, Boyd NF, Fishell E, Jong RA, Yaffe MJ. The quantitative-analysis of mammographic densities. Phys Med Biol. 1994;39:1629–38. Sivaramakrishna R, Obuchowski NA, Chilcote WA, Powell KA. Automatic segmentation of mammographic density. Acad Radiol. 2001;8:250–6. Highnam R, Pan X, Warren R, Jeffreys M, Smith GD, Brady M. Breast composition measurements using retrospective standard mammogram form (SMF). Phys Med Biol. 2006;51:2695–713. Highnam R, Jeffreys M, McCormack V, Warren R, Smith GD, Brady M. Comparing measurements of breast density. Phys Med Biol. 2007;52:5881–95. Kaufhold S, Dohrmann R, Ufer K, Meyer FM. Comparison of methods for the quantification of montmorillonite in bentonites. Appl Clay Sci. 2002;22:145–51. Pawluczyk O, Augustine BJ, Yaffe MJ, Rico D, Yang JW, Mawdsley GE, et al. A volumetric method for estimation of breast density on digitized screen-film mammograms. Med Phys. 2003;30:352–64. Ducote JL, Molloi S. Quantification of breast density with dual energy mammography: a simulation study. Med Phys. 2008;35:5411–8. Ducote JL, Molloi S. A quantification of breast density with dual energy mammography: an experimental feasibility study. Med Phys. 2010;37(793–801):24. Ding H, Molloi S. Quantification of breast density with spectral mammography based on a scanned multi-slit photon-counting detector: a feasibility study. Phys Med Biol. 2012;2012(57):4719–38. doi:10.1088/0031-9155/57/15/4719. Epub. D’Orsi CJ, Vassett JW, Berg WA, Feig SA, Jackson VP, Kopans DB, et al. ACR BI-RADS––mammography. 4th ed. In: ACR
Jpn J Radiol breast imaging reporting and data system, breast imaging atlas. Reston: American College of Radiology; 2003. 23. Research questions about relationships among variables. In: Dawson B, Trapp RG, editors. Basic and clinical biostatistics, Chapter 8. 4th ed. New York: McGrawHill; 2004. pp. 190–220. 24. Venturini E, Losio C, Panizza P, Rodighiero MG, Fedele I, Tacchini S, et al. Tailored breast cancer screening program with microdose mammography, US, and MR imaging: short-term results of a pilot study in 40–49-year-old women. Radiology. 2013;2013(268):347–55. doi:10.1148/radiol.13122278.Epub. 25. Heine JJ, Carston MJ, Scott CG, Brandt KR, Wu FF, Pankratz VS, et al. An automated approach for estimation of breast density. Cancer Epidemiol Biomark Prev. 2008;17:3090–7. doi:10.1158/ 1055-9965.EPI-08-0170.
26. Jeffreys M, Warren R, Highnam R, Smith GD. Initial experiences of using an automated volumetric measure of breast density: the standard mammogram form. Br J Radiol. 2006;79:378–82. 27. El-Bastawissi AY, White E, Mandelson MT, Taplin SH. Reproductive and hormonal factors associated with mammographic breast density by age (United States). Cancer Causes Control. 2000;11(10):955–63. 28. El-Bastawissi AY, White E, Mandelson MT, Taplin S. Variation in mammographic breast density by race. Ann Epidemiol. 2001;11(4):257–63.
123
