Computed Radiography for Breast Cancer Tatsuya Yamada* and Yukio Muramatsu Department of Radiology, National Cancer Center Hospital, 1-1, Tsukiji 5-chome, Chuo-ku, Tokyo 104
(Jpn. J. Clin. Oncol. 20: 164—168, 1990) Key words: Mammography—Computed radiography—Breast cancer
Introduction
Patients and Methods
Mass surveys for breast cancer have very great importance in reducing the mortality, and it is well known that mammography is both valuable and indispensable in detecting this disease. Examination and diagnostic techniques have already been established.121 The American Cancer Society and the National Cancer Institute (NCI) recommend mammographic screening for detecting breast cancer in its early stages;3'4) but, in fact, mammography has not been fully used because of the possible radiation hazzard.51 On the other hand, mammography which employs the computed radiographic (CR) system enables us to reduce the X-ray dosage.6'7) We have, therefore, evaluated the efficiency of CR in detecting breast cancer compared to that of the conventional non-screen method.
A comparative study between CR and conventional non-screen mammography with reference to detection rate and ability to recognize tumors and their malignant findings such as marginal spicula and tumor calcification from a single projection film, was carried out in 71 patients (mean±SD age, 51 ±11 years), each having a histopathologically verified breast cancer lesion between November, 1985, and July, 1986. The ability to recognize tumor images was classified into the following three groups as evaluated by three radiologists who have been engaged in the diagnosis of breast cancer: (1) CR obviously better than the conventional method, (2) no difference between the two methods, (3) CR inferior to the conventional method. Exposures of- both X-ray films (X-OMAT TL, Kodak, Rochester) and imaging plates (Fuji Film, Tokyo) were carried out in the same machine (Mammoace, Toshiba, Tokyo). Two types of X-ray images were obtained after gradation and frequency processing7' in the CR system (CR 201, Fuji Film and Toshiba). One imaging technique was applied to detect the tumors obtained from our origi-
Received: February 2, 1989 Accepted: January 29,1990 • For reprints and all correspondence 164
Jpn. J. Clin. Oncol. 20(2) 1990
Downloaded from https://academic.oup.com/jjco/article-abstract/20/2/164/818705 by Lancaster University user on 12 January 2019
In order to evaluate the possibility of using computed radiographic mammography in mass surveys of the breast, we have retrospectively examined 71 breast cancer lesions in 71 patients using computed radiographic and conventional non-screen mammographies and have carried out comparative studies on tumor detection rate and calcification. A 95.8% detection rate was obtained for the tumor image (n 71) using computed radiography (CR) and one of 93.0% using non-screen techniques. Three lesions remained undetected by either study. A 100% detection rate was obtained for calcification associated with cancer (n 33) from each method. No significant differences in either detection rate or calcification were seen between the two images. On the other hand, the ability to recognize tumor images (n 66) was as follows; CR superior to non-screen radiography in 53 lesions (80.3%), equal in eight lesions (12.1%) and inferior in five lesions (7.6%). For the calcification images (n 18), CR was superior to non-screen radiography in all 18 lesions. Obviously, CR gives better results than non-screen radiography. Furthermore, an adequate image can be obtained using CR even although the X-ray dosage is only a twentieth of that required for non-screen radiography. It can therefore be applied not only to mass surveys for breast cancer but also to routine clinical diagnoses.
CR MAMMOGRAPHY OF BREAST CANCER
Results From the 71 breast cancer lesions, 68 and 66 were detected as mass shadows on CR and nonscreen mammograms, respectively. Spicula and calcifications within and/or around a mass lesion were recognized in 18 and 33 lesions, respectively, in each study (Table I). No significant differences were seen in tumor detection rates, spicula and calcification images between CR and the non-screen method. Five lesions did not show up when the
non-screen method was employed because of dense breast tissue (Case 1, Fig. la). Two of them were demonstrated, however, using CR (Case 1, Fig. lb). On the other hand, abilities to recognize tumor and calcification images obtained using CR were superior to those obtained using the non-screen method (Table II); tumors were seen more clearly in 53 out of 66 lesions using CR (Case 2, Fig. 2b). Recognition of the calcification image was, moreover, extremely good in 18 lesions (Case 3, Fig. 3b), because it was possible to carry out processing for enhancing the image.
Discussion Breast cancers grow in the mammary gland tissue; therefore, the examinations used for their detection are influenced by their density. The detection of breast cancer can be carried out with ease when the fatty component of the breast tissue is dominant. On the other hand, detecting breast cancer in the dense breast is extremely difficult using conventional mammography. Generally, target-
Fig. 1. Case 1. A dense breast parenchyma in a 47-year-old woman. The tumor is not detected using conventional mammography (a) but is clearly seen using CR mammography (arrowheads; b).
Downloaded from https://academic.oup.com/jjco/article-abstract/20/2/164/818705 by Lancaster University user on 12 January 2019
nal high gamma or "1.0 d" gradation processing curve.8' Another was employed in detecting calcification using both prepared processing gradation with a 45° inclination and frequency processing, from which the area was enhanced with approximately 0.5 cycles/mm. A twentieth (150-200 mR) of the X-ray dosage required for the non-screen method was used in CR in all cases for comparing the quality of images. The cumulative Chi-squared test was used for statistical analysis.
YAMADA ET AL.
In the present comparative study, no significant differences in detection rates of masses, calcifications and spicula could be seen between the two modalities (Table I). The tumor image obtained after processing, however, which had been difficult to display using conventional mammography because of tissue density, was subjectively clearly demonstrated using CR (Table II). In fact, however, CR was equal or inferior to conventional mammogra-
Table I. Comparative Detection Rate Study between Computed Radiography and Non-screen Mammography
Table II. Ability to Recognize Mass Shadows and Calcification. Comparison of Computed Radiography to Non-screen Mammography
Findings Mass shadow Spicula Calcification
No. of lesions Computed with spicula or radiography calcification 71 33 18
Non-screen
68 (95.8) 66 ( 93) 33 (100) 33 (100) 18 (100) 18 (100)
No. of mass shadows (%)
No. with calcification (%)
CR > Non-screen CR = Non-screen CR < Non-screen
53 (80.3) 8(12.1) 5 ( 7.6)
18 (100) 0 0
Total
66 (100)
18 (100)
Category
(a)
Fig. 2. Case 2. A rather dense breast parenchyma in a 43-year-old woman. The spiculated mass is detected by conventional mammography (a) but is more clearly seen on the CR mammogram (b).
im
Jpn. J. Clin. Oncol. 20(2) 1990
Downloaded from https://academic.oup.com/jjco/article-abstract/20/2/164/818705 by Lancaster University user on 12 January 2019
ing spot compression studies for breast masses which have been detected using conventional mammography are used to differentiate cancer from benign disease and to reduce false positive and negative lesions. No spot compression study was performed in CR because the software for dealing with a small part of the breast had not yet been developed. We, therefore, excluded benign diseases from our analysis.
CR MAMMOGRAPHY OF BREAST CANCER
according to the imaging menu. The breast menu was not too appropriate for the diagnosis of breast cancer,8' however, we selected our original "1.0 d" processing curve. X-ray mammography is very useful for detecting breast cancer, but there is a radiation risk when it is used in mass surveys of the breast.5' NCI would restrict mammography for asymptomatic women under the age of 50 years.3' On the other hand, a CR mammogram adequate for diagnosis can be obtained using highly sensitive X-ray imaging plates in place of conventional X-ray films, even although for this method the X-ray dosage required for each exposure is reduced to half (300-400 mR) of that for the film-screen method and a twentieth (150-200 mR) of that for the non-screen
(a)
(b)
Fig. 3. Case 3. Malignant calcifications in the dense breast parenchyma of a 40-year-old woman. The calcifications are detected by both conventional (a) and CR mammography (b), the latter, however, demonstrating them clearly. The calcifications are, moreover, more clearly demonstrated than those shown in b, by means of changing the processing (c).
167
Downloaded from https://academic.oup.com/jjco/article-abstract/20/2/164/818705 by Lancaster University user on 12 January 2019
phy in approximately 20% of patients in spite of the fact that processing had been carried out several times. We speculated that this was the result of the low X-ray dosage. The configuration of calcifications was imaged in detail using the non-screen method, but with CR it was easy to judge the existence of calcifications because their images were enhanced by processing (Table II). This was regarded as one of the reasons for recommending CR for breast screening. X-ray image data are digitalized in CR mammography and, therefore, various changes in image quality can be carried out by the computer.61 In gradation processing, there are several kinds of gradation curves prepared in advance by the CR maker, and a standard density level is set for each
YAMADA ET AL.
4)
5)
6)
7)
8)
9)
Acknowledgments The present study was supported by Grants-in-Aid for Cancer Research from the Ministry of Health and Welfare and the 10 Year Strategy for Cancer Research.
10)
References 1) Shapiro, S., Venet, W., Strax, P., Venet, L. and Roeser, R. Ten- to fourteen-year effect of screening on breast cancer mortality. JNCI 69, 349-355 (1982). 2) Baker, L. H. Breast cancer detection demonstration project: five-year summary report. CA. 32, 194-225 (1982) 3) National Task Force on Breast Cancer Control.
168
11)
Mammography 1982. A statement of the American Cancer Society. CA. 32, 226-230 (1982). US Department of Health and Human Services. Breast Cancer. We're making progress every day. Publication no. (NIH) 82-2409 (1982). Government Printing Office. Cummings, K. M., Funch, D. P. and Mettlin, C. Family physicians' beliefs about breast cancer screening by mammography. J. Fam. Pract. 17, 1029-1034 (1983). Yamada, T. and Muramatsu, Y. Computed Radiography, in "Houshasen Igaku Taikei," pp. 192-199 (1988). Nakayama Syoten, Tokyo, (in Japanese) Sonoda, M., Takano, M., Miyahara, J. and Kato, H. Computed radiography utilizing scanning laser stimulated luminescence. Radiology 148, 833-838 (1983). Muramatsu, Y., Nawano, S., Anan, M., Hanmura, K., Tanaka, T., Matsue, H. and Yamada, T. A study of image processing in CR mammographygradation processing. Rinsho Hoshasen 35, 271-276 (1990). (in Japanese) Takei, H., Miyashita, K., Hirabayashi, T., Ichinose, T., Dobashi, Z., Uehara, H., Tada, R., Fujimatsu, M., Koike, K., Hayashi, M., Nakagawa, M., Kamata, M. and You, T. Experience of common usage of computed radiogrlaphy. Fuji Medical Forum 160, 2-11 (1989). (in Japanese) Hanmura, K., Anan, M., Tanaka, T. and Murakami, T. A trial of imaging calcifications with usage of CR-subtraction. Iryo 41 (Suppl.) 528 (1987). (in Japanese) Asaga, T., Chiyasu, S., Matsuda, S., Matsuura, H., Kato, H., Ishida, M. and Komaki, T. Breast imaging: dual-energy projection radiography with digital radiography. Radiology 164, 869-873 (1987).
Jpn. J. Clin. Oncol. 20(2) 1990
Downloaded from https://academic.oup.com/jjco/article-abstract/20/2/164/818705 by Lancaster University user on 12 January 2019
method.6' In fact, there was no problem in detecting breast cancer in the CR mammography. The application of CR to mass surveys could, therefore, prove valuable. CR is a kind of image processing system and consumes as much time in examinations and management as those of the conventional technique. Whenever there are conventional X-ray systems, CR is available if imaging plates are used in place of X-ray films. Common usage of the CR system and imaging plates could be one way of resolving its economic difficulties." It would be possible to display adequately, in several archives, the levels of calcification found in the mammary gland and/or tumors using CR and new image processing10' (Fig. 3c). The dual-energy projection method which can clearly image tumors in the dense breast has, moreover, been developed.1" We therefore conclude that CR could have a wide application to the diagnosis of breast cancer.
(Jpn. J. Clin. Oncol. 20: 164—168, 1990) Key words: Mammography—Computed radiography—Breast cancer
Introduction
Patients and Methods
Mass surveys for breast cancer have very great importance in reducing the mortality, and it is well known that mammography is both valuable and indispensable in detecting this disease. Examination and diagnostic techniques have already been established.121 The American Cancer Society and the National Cancer Institute (NCI) recommend mammographic screening for detecting breast cancer in its early stages;3'4) but, in fact, mammography has not been fully used because of the possible radiation hazzard.51 On the other hand, mammography which employs the computed radiographic (CR) system enables us to reduce the X-ray dosage.6'7) We have, therefore, evaluated the efficiency of CR in detecting breast cancer compared to that of the conventional non-screen method.
A comparative study between CR and conventional non-screen mammography with reference to detection rate and ability to recognize tumors and their malignant findings such as marginal spicula and tumor calcification from a single projection film, was carried out in 71 patients (mean±SD age, 51 ±11 years), each having a histopathologically verified breast cancer lesion between November, 1985, and July, 1986. The ability to recognize tumor images was classified into the following three groups as evaluated by three radiologists who have been engaged in the diagnosis of breast cancer: (1) CR obviously better than the conventional method, (2) no difference between the two methods, (3) CR inferior to the conventional method. Exposures of- both X-ray films (X-OMAT TL, Kodak, Rochester) and imaging plates (Fuji Film, Tokyo) were carried out in the same machine (Mammoace, Toshiba, Tokyo). Two types of X-ray images were obtained after gradation and frequency processing7' in the CR system (CR 201, Fuji Film and Toshiba). One imaging technique was applied to detect the tumors obtained from our origi-
Received: February 2, 1989 Accepted: January 29,1990 • For reprints and all correspondence 164
Jpn. J. Clin. Oncol. 20(2) 1990
Downloaded from https://academic.oup.com/jjco/article-abstract/20/2/164/818705 by Lancaster University user on 12 January 2019
In order to evaluate the possibility of using computed radiographic mammography in mass surveys of the breast, we have retrospectively examined 71 breast cancer lesions in 71 patients using computed radiographic and conventional non-screen mammographies and have carried out comparative studies on tumor detection rate and calcification. A 95.8% detection rate was obtained for the tumor image (n 71) using computed radiography (CR) and one of 93.0% using non-screen techniques. Three lesions remained undetected by either study. A 100% detection rate was obtained for calcification associated with cancer (n 33) from each method. No significant differences in either detection rate or calcification were seen between the two images. On the other hand, the ability to recognize tumor images (n 66) was as follows; CR superior to non-screen radiography in 53 lesions (80.3%), equal in eight lesions (12.1%) and inferior in five lesions (7.6%). For the calcification images (n 18), CR was superior to non-screen radiography in all 18 lesions. Obviously, CR gives better results than non-screen radiography. Furthermore, an adequate image can be obtained using CR even although the X-ray dosage is only a twentieth of that required for non-screen radiography. It can therefore be applied not only to mass surveys for breast cancer but also to routine clinical diagnoses.
CR MAMMOGRAPHY OF BREAST CANCER
Results From the 71 breast cancer lesions, 68 and 66 were detected as mass shadows on CR and nonscreen mammograms, respectively. Spicula and calcifications within and/or around a mass lesion were recognized in 18 and 33 lesions, respectively, in each study (Table I). No significant differences were seen in tumor detection rates, spicula and calcification images between CR and the non-screen method. Five lesions did not show up when the
non-screen method was employed because of dense breast tissue (Case 1, Fig. la). Two of them were demonstrated, however, using CR (Case 1, Fig. lb). On the other hand, abilities to recognize tumor and calcification images obtained using CR were superior to those obtained using the non-screen method (Table II); tumors were seen more clearly in 53 out of 66 lesions using CR (Case 2, Fig. 2b). Recognition of the calcification image was, moreover, extremely good in 18 lesions (Case 3, Fig. 3b), because it was possible to carry out processing for enhancing the image.
Discussion Breast cancers grow in the mammary gland tissue; therefore, the examinations used for their detection are influenced by their density. The detection of breast cancer can be carried out with ease when the fatty component of the breast tissue is dominant. On the other hand, detecting breast cancer in the dense breast is extremely difficult using conventional mammography. Generally, target-
Fig. 1. Case 1. A dense breast parenchyma in a 47-year-old woman. The tumor is not detected using conventional mammography (a) but is clearly seen using CR mammography (arrowheads; b).
Downloaded from https://academic.oup.com/jjco/article-abstract/20/2/164/818705 by Lancaster University user on 12 January 2019
nal high gamma or "1.0 d" gradation processing curve.8' Another was employed in detecting calcification using both prepared processing gradation with a 45° inclination and frequency processing, from which the area was enhanced with approximately 0.5 cycles/mm. A twentieth (150-200 mR) of the X-ray dosage required for the non-screen method was used in CR in all cases for comparing the quality of images. The cumulative Chi-squared test was used for statistical analysis.
YAMADA ET AL.
In the present comparative study, no significant differences in detection rates of masses, calcifications and spicula could be seen between the two modalities (Table I). The tumor image obtained after processing, however, which had been difficult to display using conventional mammography because of tissue density, was subjectively clearly demonstrated using CR (Table II). In fact, however, CR was equal or inferior to conventional mammogra-
Table I. Comparative Detection Rate Study between Computed Radiography and Non-screen Mammography
Table II. Ability to Recognize Mass Shadows and Calcification. Comparison of Computed Radiography to Non-screen Mammography
Findings Mass shadow Spicula Calcification
No. of lesions Computed with spicula or radiography calcification 71 33 18
Non-screen
68 (95.8) 66 ( 93) 33 (100) 33 (100) 18 (100) 18 (100)
No. of mass shadows (%)
No. with calcification (%)
CR > Non-screen CR = Non-screen CR < Non-screen
53 (80.3) 8(12.1) 5 ( 7.6)
18 (100) 0 0
Total
66 (100)
18 (100)
Category
(a)
Fig. 2. Case 2. A rather dense breast parenchyma in a 43-year-old woman. The spiculated mass is detected by conventional mammography (a) but is more clearly seen on the CR mammogram (b).
im
Jpn. J. Clin. Oncol. 20(2) 1990
Downloaded from https://academic.oup.com/jjco/article-abstract/20/2/164/818705 by Lancaster University user on 12 January 2019
ing spot compression studies for breast masses which have been detected using conventional mammography are used to differentiate cancer from benign disease and to reduce false positive and negative lesions. No spot compression study was performed in CR because the software for dealing with a small part of the breast had not yet been developed. We, therefore, excluded benign diseases from our analysis.
CR MAMMOGRAPHY OF BREAST CANCER
according to the imaging menu. The breast menu was not too appropriate for the diagnosis of breast cancer,8' however, we selected our original "1.0 d" processing curve. X-ray mammography is very useful for detecting breast cancer, but there is a radiation risk when it is used in mass surveys of the breast.5' NCI would restrict mammography for asymptomatic women under the age of 50 years.3' On the other hand, a CR mammogram adequate for diagnosis can be obtained using highly sensitive X-ray imaging plates in place of conventional X-ray films, even although for this method the X-ray dosage required for each exposure is reduced to half (300-400 mR) of that for the film-screen method and a twentieth (150-200 mR) of that for the non-screen
(a)
(b)
Fig. 3. Case 3. Malignant calcifications in the dense breast parenchyma of a 40-year-old woman. The calcifications are detected by both conventional (a) and CR mammography (b), the latter, however, demonstrating them clearly. The calcifications are, moreover, more clearly demonstrated than those shown in b, by means of changing the processing (c).
167
Downloaded from https://academic.oup.com/jjco/article-abstract/20/2/164/818705 by Lancaster University user on 12 January 2019
phy in approximately 20% of patients in spite of the fact that processing had been carried out several times. We speculated that this was the result of the low X-ray dosage. The configuration of calcifications was imaged in detail using the non-screen method, but with CR it was easy to judge the existence of calcifications because their images were enhanced by processing (Table II). This was regarded as one of the reasons for recommending CR for breast screening. X-ray image data are digitalized in CR mammography and, therefore, various changes in image quality can be carried out by the computer.61 In gradation processing, there are several kinds of gradation curves prepared in advance by the CR maker, and a standard density level is set for each
YAMADA ET AL.
4)
5)
6)
7)
8)
9)
Acknowledgments The present study was supported by Grants-in-Aid for Cancer Research from the Ministry of Health and Welfare and the 10 Year Strategy for Cancer Research.
10)
References 1) Shapiro, S., Venet, W., Strax, P., Venet, L. and Roeser, R. Ten- to fourteen-year effect of screening on breast cancer mortality. JNCI 69, 349-355 (1982). 2) Baker, L. H. Breast cancer detection demonstration project: five-year summary report. CA. 32, 194-225 (1982) 3) National Task Force on Breast Cancer Control.
168
11)
Mammography 1982. A statement of the American Cancer Society. CA. 32, 226-230 (1982). US Department of Health and Human Services. Breast Cancer. We're making progress every day. Publication no. (NIH) 82-2409 (1982). Government Printing Office. Cummings, K. M., Funch, D. P. and Mettlin, C. Family physicians' beliefs about breast cancer screening by mammography. J. Fam. Pract. 17, 1029-1034 (1983). Yamada, T. and Muramatsu, Y. Computed Radiography, in "Houshasen Igaku Taikei," pp. 192-199 (1988). Nakayama Syoten, Tokyo, (in Japanese) Sonoda, M., Takano, M., Miyahara, J. and Kato, H. Computed radiography utilizing scanning laser stimulated luminescence. Radiology 148, 833-838 (1983). Muramatsu, Y., Nawano, S., Anan, M., Hanmura, K., Tanaka, T., Matsue, H. and Yamada, T. A study of image processing in CR mammographygradation processing. Rinsho Hoshasen 35, 271-276 (1990). (in Japanese) Takei, H., Miyashita, K., Hirabayashi, T., Ichinose, T., Dobashi, Z., Uehara, H., Tada, R., Fujimatsu, M., Koike, K., Hayashi, M., Nakagawa, M., Kamata, M. and You, T. Experience of common usage of computed radiogrlaphy. Fuji Medical Forum 160, 2-11 (1989). (in Japanese) Hanmura, K., Anan, M., Tanaka, T. and Murakami, T. A trial of imaging calcifications with usage of CR-subtraction. Iryo 41 (Suppl.) 528 (1987). (in Japanese) Asaga, T., Chiyasu, S., Matsuda, S., Matsuura, H., Kato, H., Ishida, M. and Komaki, T. Breast imaging: dual-energy projection radiography with digital radiography. Radiology 164, 869-873 (1987).
Jpn. J. Clin. Oncol. 20(2) 1990
Downloaded from https://academic.oup.com/jjco/article-abstract/20/2/164/818705 by Lancaster University user on 12 January 2019
method.6' In fact, there was no problem in detecting breast cancer in the CR mammography. The application of CR to mass surveys could, therefore, prove valuable. CR is a kind of image processing system and consumes as much time in examinations and management as those of the conventional technique. Whenever there are conventional X-ray systems, CR is available if imaging plates are used in place of X-ray films. Common usage of the CR system and imaging plates could be one way of resolving its economic difficulties." It would be possible to display adequately, in several archives, the levels of calcification found in the mammary gland and/or tumors using CR and new image processing10' (Fig. 3c). The dual-energy projection method which can clearly image tumors in the dense breast has, moreover, been developed.1" We therefore conclude that CR could have a wide application to the diagnosis of breast cancer.
