- Email: [email protected]
CTLM as an adjunct to mammography in the diagnosis of patients with dense breast
Clinical Imaging 37 (2013) 289 – 294
CTLM as an adjunct to mammography in the diagnosis of patients with dense breast Jin Qi, Zhaoxiang Ye⁎ Radiology Department, Tianjin Medical University Cancer Institute and Hospital, Key Laboratory of Breast Cancer Prevention and Therapy of the Ministry of Education, Key Laboratory of Cancer Prevention and Therapy, Tianjin 300060, PR China Received 10 April 2012; accepted 1 May 2012
Abstract Purpose: The purpose was to evaluate the utility of computed tomographic laser mammography (CTLM) as an adjunct examination to mammography in women with dense breast tissue. Methods: We retrospectively compared the findings of mammography, CTLM, and adjunct CTLM to mammography with pathology of 155 women scheduled for biopsy or surgery. Results: Positive lesions were observed more significantly in malignant than benign lesions. The sensitivity of mammography vs. mammography+CTLM was 34.4% vs. 81.57% among extremely dense breasts and 68.29% vs. 95.34% among heterogeneously dense breasts. Conclusion: CTLM could distinguish benign lesions from malignant lesions and is not affected by breast density. © 2013 Elsevier Inc. All rights reserved. Keywords: Computed tomographic laser mammography; Laser technique; Tomography; Optical imaging; Mammography; Breast cancer
1. Introduction Early detection and effective treatment of women with a diagnosis of breast cancer are major factors contributing to the decline in the mortality rate. Radiology examination can help in early detection and diagnosis. At the present time, the means of diagnosis of breasts now used include mammography, ultrasound, computed tomography (CT), magnetic resonance imaging (MRI), near and far infrared, and positron emission tomography (PET)-CT [1]. Mammography is the most widely used; however, conventional screenfilm mammography has limited sensitivity for detection of breast cancer especially in breasts with dense tissue. Digital mammography was developed to address some of the limi⁎ Corresponding author. Radiology Department, Tianjin Medical University Cancer Institute and Hospital, Key Laboratory of Breast Cancer Prevention and Therapy of the Ministry of Education, Key Laboratory of Cancer Prevention and Therapy, Tianjin 300060, PR China. Tel.: +86 22 23359337; fax: +86 22 23359337. E-mail address: [email protected] (Z. Ye). 0899-7071/$ – see front matter © 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.clinimag.2012.05.003
tations of screen-film mammography; however, the value of digital mammography is not substantially different from that of screen-film mammography. Radiology physicians are in need of more functional information. Tumor “angiogenesis” is known to be critical for the autonomous growth and spread of breast cancers. Tumor angiogenesis is a complex process that involves both the incorporation of existing host blood vessels into the tumor and the creation of tumor microvessels [2]. If these vessels could be found, the information of function can be used by physicians. The basic principle underlying the new device, computed tomographic laser mammography (CTLM) imaging, is the “angiogenesis.” In the last several years, optical tomography on breast imaging has gained interest all over the world. Light has been investigated since the late 1920s as a diagnostic tool for breast cancer by transillumination; however, it had low spatial resolution and afforded little in spectral quantification of lesions. Hence, it did not attain sufficient sensitivity and specificity to be used clinically. In recent years, optical mammography stepped to new era. Optical introscopy of laboratory animals is extensively developed for
290
J. Qi, Z. Ye / Clinical Imaging 37 (2013) 289–294
studies of DNA and pharmacological preparations. Optical mammography uses either projection or tomographic apparatus. Optical tomographic mammographs are presently available. CTLM (Imaging Diagnosis System Inc., USA) used in our study is now being clinically tested in several countries; the other apparatuses such as Philips OMPS transmission mammography(Optical Mammo Prototype System, Netherlands), SoftScan Optical projection mammography (Advanced Research Technologies Inc., Canada), and ComfortScan mammography (DOBI Medical International Inc., USA) are all being researched [3–5]. The purpose of the study was to evaluate the utility of computed CTLM as an adjunct examination to mammography in women with heterogeneously dense and extremely dense breast tissue [Breast Imaging Reporting and Data System (BI-RADS)3 and 4].
be proper to the size of patient's breasts. In general, 2-mm slice thickness would be set to most patients; when the breast is too large or too small, 1-mm and 4-mm slice thickness should be set. Before scanning, the breast was moved in the middle of scanning ring. The detector system consists of 2 rings with 84 photodetectors each (picture 1). This unit rotates 360° around the breast and takes 17 s. After each rotation, the ring descends, creating a slice at each step. The scanning platform alternates direction to avoid twisting the cables. The whole breast scan would be made up of 10–40 slices. While slice one level is being scanned, our proprietary software is reconstructing the previous slice. The scan is ready to be read immediately after completion. The device produces striking three-dimensional (3D) views, which can be rotated, in real time, along any axis. CTLM also produces sagittal, axial, and coronal views. 2.4. CTLM image interpretation
2. Materials and methods 2.1. Patients From September 2007 to February 2008, 155 women (23–74 years of age; median age: 41years) in Tianjin Medical University Cancer Hospital were scheduled for biopsy or surgery in the short term (within 30 days). All the patients did not undergo current chemotherapy or irradiation of the breast, and were not biopsied within 30 days before the examination. All the patients underwent mammography and CTLM examination before surgery. According to BI-RADS, all women who underwent mammography were classified into heterogeneously dense breast (BI-RADS3) and extremely dense breast ( BI-RADS4). CTLM was performed subsequently. Exclusion criteria. Included (1) ulcer or wounds on the breast (2) various forms of porphyria (3) current biopsy within 30 days (4) chemotherapy or irradiation of the breast. 2.2. Mammography Bilateral 2-view mammography (Senography 2000D, GE Medical Systems), was performed in the craniocaudal and mediolateral projections. Spot compression and additional views were obtained where appropriate. Two mammography independently analyzed the data for this purpose. The physicians classified all patients into two groups: heterogeneously dense breast group (BI-RADS3) and extremely dense breast group (BI-RADS4). 2.3. CTLM CTLM (CTLM 1020; Imaging Diagnostic Systems, USA) was performed in one side with lesion. The patient lay in the prone position on the examination table with one breast pendant into a scanning chamber, where the breast is surrounded by the laser source detector .The scan ring must
CTLM device produces a laser wavelength (808 nm), in the NIR spectrum, which matches the crossover point of absorption of both oxygenated and deoxygenated hemoglobin. The hurdle that has to be overcome is the difference in the way photons transverse the tissue. Light travels in tissue in a random fashion due to scatter. The average path of light can be predicted in spite of the scattering and absorption of light in tissue through. CTLM uses a large number of source and detector positions in each slice to take into account the diffusion approximation of light propagation in tissue and to show the location of the increased vascularity in the breast. The device produces striking 3D views, which can be rotated, in real time, along any axis. It also produces sagittal, axial, and coronal views. An inversion factor is introduced into the reconstruction algorithm so that areas of high absorption (high hemoglobin concentration) are visualized in white. Relatively avascular areas are shown as green or black. The available tools allow us to focus on any area by removing overlying obscuring areas. Image quality can be improved with window and level controls. Images are usually viewed in a maximal intensity projection (MIP) and then in a front to back reconstruction (FTB), also known as a surface rendering mode. The two modes are used to evaluate the vascularization patterns to determine whether the images represent normal or abnormal vascularization. CTLM images were read by a trained radiology physicians, blinded to any imaging and clinical information. To guarantee the accuracy of the data, the difficult cases were review by Eric Mile, MD (USA). Following are the CTLM interpretive data: 1 stands for benign lesions; 2 stands for malignant lesions, 0 stands for the lesions that were not unascertainable. 2.5. Mammography imaging interpretation Following are the mammography interpretive data: 1 stands for benign lesions record; 2 stands for malignant lesions; 0 stands for the lesions which were not unascertainable.
J. Qi, Z. Ye / Clinical Imaging 37 (2013) 289–294
291
Table 1 CTLM+mammography image interpretation
Table 3 Pathologic findings in 155 lesions
Examination
Interpretation
Pathologic findings
Density of breast
Mammography
Benign
Extremely
Heterogeneously
CTLM Mammography+CTLM
1a 1 1
Benign Fibrocystic changes Inflammation Ductal disease Fibroadenomas or adenofibromas Adenosis Cyst Adiponecrosis Normal tissue
37 16 2 4 9 3 1 1 1
39 9 7 9 11 0 0 0 3
Malignant DCIS Invasive ductal cancer Neuroendocrine carcinoma Mucoid adenocarcinoma Apocrine gland carcinoma Interstitial sarcoma Glycogen-rich clear cell carcinoma Mesenchymal sarcoma
37 4 28 0 0 1 2 1 1
42 5 35 1 1 0 0 0 0
a b c
1 2 2
Malignant 1 0 1
2b 1 2
2 2 2
Not ascertain 0c 1 1
2 0 2
0 2 2
0 0 0
1=benign. 2=malignant. 0=not ascertained.
2.6. Mammography+CTLM image interpretation Combine the result of mammography and CTLM, accept it as malignant when any one was malignant, accept it as benign or malignant when they were the same. If one of them is not unascertainable, record depended on another result, as shown in Table 1. 2.7. Statistics
3.2. CTLM
We analyzed the result of mammography, CTLM, and mammography+CTLM to identify true-positive, true-negative, false-positive, and false-negative examination according to pathology .On the basis of these classifications, we measured sensitivity (true-positive/[true-positive+falsenegative]), specificity (true-negative/[false-positive+truenegative]), and positive predictive value (true-positive/ [true-positive+false-positive]). Statistical analysis was performed with the SPSS statistical package .A χ 2 test was used to compare group differences. 3. Results
Radiology physician recorded CTLM impression, as follows: 35 benign lesions, 34 malignant lesions, and 5 ascertained lesions in extremely dense breast, and 33 benign lesions, 43 malignant lesions, and 5 unascertainable lesions in heterogeneously dense breast group, as shown in Table 2. 3.3. Mammography+CTLM There were 32 benign lesions, 40 malignant lesions, and 2 unascertainable lesions in extremely dense breast, and 23 benign lesions, 57 malignant lesions, and 1 unascertainable lesion in heterogeneously dense breast group, as shown in Table 2.
3.1. Mammography 3.4. Pathologic diagnosis Two radiology physicians reviewed the results and classified all patients into two groups: 74 in heterogeneously dense breast group and 81 in the extremely dense breast group. The impression of mammography is as follows: 48 benign lesions, 15 malignant lesions, and 11 unascertainable lesions in extremely dense breast, and 41 benign lesions, 15 malignant lesions, and 6 unascertainable lesions in heterogeneously dense breast group, as shown in Table 2.
In 155 lesions, 150 underwent surgery; 3, biopsy; and 2, needle biopsy. In the 155 breast lesions studied, 76 were malignant and 79 were benign at pathologic analysis. The tumor stages of the 155 breast cancers are listed in Table 3. From our results, CTLM positive predictive value is 72.41% (63/63+24). It supported that CTLM successfully shows the angiogenesis of malignant lesion in breast,
Table 2 Findings in three examinations Breast density
n
Mammography
Rank Extremely (n) Heterogeneously (n)
74 81
1 48 41
2 15 34
CTLM 0 11 6
1 35 33
Mammography+CTLM 2 34 43
0 5 5
1 32 23
2 40 57
0 2 1
292
J. Qi, Z. Ye / Clinical Imaging 37 (2013) 289–294
Table 4 CTLM findings in benign and malignant lesions Pathology Benign Malignant
Table 6 Statistical data
−
+ 24 57
Breast density χ =25.558 P=.000
Mammography
CTLM
Sens
Sens
Mammography +CTLM
2
52 22
+, angiogenesis was shown in CTLM; −, angiogenesis was not shown.
Spec
Spec
Sens
Spec
Extremely 34.40% 90.48% 74.4% 71.00% 81.57% 78.13% Heterogeneously 68.29% 85.00% 85.00% 61.00% 95.34% 55.26% As shown in Table 7: sensitivity of CTLM in heterogeneously dense breast.
However, 57% (24/76) of benign lesions also showed angiogenesis. Among these cases, there are eight fibrocystic changes, six inflammations, four ductal diseases, two adenoses, one fibroadenoma, and one adiponecrosis. Increased absorption was observed significantly more often in malignant than in benign lesions [72.15% (57/79) vs. 31.57% (24/76), χ 2= 25.558, P=.000), as shown in Table 4. Among extremely dense breast, the sensitivity of mammography alone, CTLM alone, and mammography+CTLM was, respectively, 34.40%, 74.40%, and 81.57%. The sensitivity of mammography vs. mammography +CTLM was 34.4% vs. 81.57% (χ 2=13.071, P=.000). The specificity of mammography alone, CTLM alone, and mammography +CTLM was 90.48%, 71.00%, and 78.13%. Among heterogeneously dense breast, the sensitivity of mammography alone, CTLM alone, and mammography + CTLM was 68.29%, 85.00%, and 95.34%. The sensitivity of mammography vs. mammography+CTLM was 68.29% vs. 95.34% (χ 2= 11.131, P=.001). The specificity of mammography alone, CTLM alone, and mammography+CTLM was 85.00%, 61.00%, and 55.26%. The details are shown in Tables 5 and 6. The percentage of dense breast vs. extremely dense breast was 74.40% vs. 85.00% (χ 2=0.446, P=.504); specificity in them was 71.00% vs. 61.00% (χ 2= 0.0001, P= 1.0). 4. Discussion Mammography is the golden standard of breast imaging diagnosis; it has high sensitivity in fatty breast with sharp contrast [6]. Screening mammography and digital mammography are of limited value in dense breasts. In extremely dense and heterogeneously dense breasts, mammography sensitivity is decreased. Mammography sensitivity was 45% to 70% [7]. In our study, in extremely dense breast, the sensitivity of digital mammography was 34.4%, lower than the literature reported. In heterogeneously dense breast, mammographic sensitivity was 68.29%, in accordance with the literature reports. In our study, the specificity declined
from 90.48% to 85%. Relevance research is indicated. The risk of cancer is increased from 2.5 to 5 times with the density ascending. So it is important for the malignant lesions to be detected early in dense breast [8]. Because of the limit of mammography, radiology physicians turn to ultrasonography. In simple cysts, the accuracy of ultrasound is 96% to 100%. However, when ultrasound imaging is used for the differentiation of hyperplasia from solid lesions that mammography could not distinct, it is not high as a result of the overlapping characteristics of solid benign and malignant lesions [9–11]. In the last several years, functional imaging, such as MRI and PET-CT, has gained interest and been paid more attention to. MRI can provide important information in function with high sensitivity in detecting invasive ductal cancer and ductal carcinoma in situ [12]. A new field of research is laser-light-based breast imaging. Initial attempts to use laser light for “transillumination” and detect the angiogenesis of the breast have been developed in the early 20th century. A computed tomographic laser lightbased scanner for the breast, called CTLM, has now been used clinically and has gained some initial experiment. Our study showed that the sensitivity was significantly increased by using CTLM as an adjunct to mammography, from 34.4% to 81.57% (χ 2= 13.071, P=.000), and, in heterogeneously dense breast, from 68.29% to 95.34% (χ 2= 11.131, P=.001). In 1971, Folkman [13] first proposed that tumor cells were capable of stimulating endothelial cell proliferation by means of a “soluble tumor angiogenesis factor". Angiogenesis is defined as the formation of new blood vessels through the sprouting of capillaries from preexisting microvessels [14]. This process serves as the fundamental method by which neovascularization occurs in the human body. In most mature tissues, the rate of endothelial cell turnover is extremely slow, on the order of years. A markedly increased rate of angiogenesis takes place during the normal physiologic processes of embryogenesis, uterine maturation, placental development, corpus luteum formation, and wound healing. In
Table 5 Imaging findings according to pathology Breast density
Extremely Heterogeneously
Mammography
CTLM
Mammography +CTLM
TP
FN
TN
FP
TP
FN
TN
FP
TP
FN
TN
FP
11 28
21 13
38 34
4 6
29 34
10 11
25 22
10 14
31 41
7 2
25 21
11 17
J. Qi, Z. Ye / Clinical Imaging 37 (2013) 289–294 Table 7 CTLM in extremely and heterogeneously dense breasts Breast density
CTLM findings TP
FN
Extremely Heterogeneously
29 34
10 11
χ2=0.446 P=.504
TN
FP
25 22
10 14
χ2=0.0001 P=1.000
each of these instances, the angiogenic process is strictly regulated and is terminated following completion of its intended function. Research has shown, however, that regulated or unregulated angiogenic activity plays a role in a number of different diseases. Examples include diabetic retinopathy, in which associated ocular angiogenesis is the leading cause of blindness in developed countries. A possible role of angiogenesis in the development of atherosclerosis is the focus of an ongoing study .In the most extreme case, angiogenesis has been shown to play an important role in the pathogenesis of tumor growth and metastasis [15–18]. Tumor cells control neovascularization through secretion of angiogenic factors which attract endothelial cells that proliferate and invade the stroma towards the tumor mass; then “angiogenesis” could be found. Tumors are unable to grow larger than circa 1 mm 3 without developing a new, hypoxia-triggered, blood supply [19]. Based on this theory, the CTLM device uses a laser wavelength (808 nm), in the NIR spectrum, which matches the crossover point of absorption of both oxygenated and deoxygenated hemoglobin. This wavelength also minimizes the effects of both fat and water [20]. Laser used by CTLM can so easily penetrate the breast tissue and is not affected by the density. Our study supported it: the CTLM sensitivity of heterogeneously dense breast vs. extremely dense breast was similar (74.40% vs. 85.00%, P=.504), and specificity was also similar(71.00% vs. 61.00%, P= 1.000). The results of the data supported that the results were not varied from heterogeneously dense breast to extremely dense breast. It has been shown that CTLM was not affected by tissue density in showing the angiogenesis. CTLM has the potential to be used as an adjunct tool to mammography in both heterogeneously dense breast and extremely dense breast. We are therefore looking at the absorption pattern of hemoglobin in vessels. In tumor angiogenesis, the vessels are concentrated and are structurally and functionally abnormal. This gives a greater volume of hemoglobin in a confined area that can be visualized by CTLM. In the analysis of cases in this study based on this principle, CT features of malignant angiogenesis are as follows: (a) Abnormal shape: CTLM showed both forms of normal and abnormal deoxyhemoglobin and hemoglobin. It is important for physicians to distinguish normal vein from abnormal “angiogenesis.” On coronal views, abnormal “angiogenesis” showed round and ovoid areas of high absorption (high hemoglobin concentration) visualized in white and distinguished from triangle and cone-shaped areas (picture 2). On 3D-MIP view, the irregular shape of “angiogenesis” showed the following: oblate sphere, dumbbell
293
shape, diverticulum, circle, distinguished from the clear “tunnels.” (b) Isolated areas of high absorption : The normal vessels are shown as “ribbon” running through the breast from chest wall to nipple in MIP-3D view. The branches of vessels shown as star-shaped and branches-shaped could always be found. The oblate areas of high absorption could be found at the nipple area. This high absorption is normal vessel but not “angiogenesis”. The normal vessel at areola is always surrounded by some veins but not isolated. However, the “angiogenesis” of malignant tumor always shows high absorption isolated which is not surrounded by vein. (c) Furthermore, “angiogenesis” of malignant tumor located deep in the breast. The normal vessel of breast is always on the surface of the breast but not deep in the breast. In this study, the high absorption showed in deep tissue is almost “angiogenesis” of malignant tumor except six cases are normal veins. The surface and deep imaging could be distinguished by comparing 3D-MIP mode with 3D-FTB mode. The high absorption on the surface of the breast is always benign lesion such as adiponecrosis and mastitis. These benign lesions could be distinguished from deep malignant absorption. These findings of CTLM help to distinguish benign lesions and malignant lesions and solve the problem of dense breasts. From our results, CTLM positive predictive value was 72.41% (63/63+24). It supported that CTLM successfully shows the angiogenesis of malignant lesion in breast. However, some benign lesion also showed angiogenesis. Although there are some overlap between benign and malignant findings, increased absorption was observed significantly more often in malignant than in benign lesions (72.15% vs. 31.57%, χ 2= 25.558, P=.000). In this hypothesis, a CTLM scanner was used to characterize benign and malignant breast lesions. We can conclude that CTLM could distinguish malignant from benign lesion. Some literature reported that [21,22]. A variety of benign lesions also demonstrated increased vessel as malignant lesions. Determination of confirmatory histologic parameters of breast lesions, such as microvessel counts, would be able to provide additional information about lesion biology. So we concluded that CTLM is able to characterize benign and malignant breast lesions by showing the angiogenesis. Our data indicate that the imaging of CTLM was not affected by tissue density in breast, and provide information about angiogenesis in breast lesions, especially in malignant lesions, when used as an adjunct to mammography in heterogeneously dense breast and extremely dense breast. Sensitivity increased significantly. This study suggests that, in clinical practice, adding CTLM in dense breasts may be useful. References [1] Bao R, Ye Z, Liu P. Make much of breast imaging diagnosis. Chin J Radiol 2007;41. [2] Zhu Q, Cronin EB. Benign versus malignant breast masses: optical differentiation with US-guided optical imaging reconstruction. Radiology 2005;237:57–66.
294
J. Qi, Z. Ye / Clinical Imaging 37 (2013) 289–294
[3] Ntziachristos V, Yodh AG. Concurrent MRI and diffuse optical tomography of breast after indocyanine green enhancement. Proc Natl Acad Sci U S A 2000;97(6):2767–72. [4] Kolb TM, Lichy J. Comparison of the performance of screening mammography, physical examination, and breast US and evaluation of factors that influence them: an analysis of 27,825 patient evaluations. Radiology 2002;225:165–75. [5] Podgaetsky VM, Potapov DA, et al. Development of optical introscopy in medical engineering. Biomed Eng 2007;41(1): 2–6. [6] Carney PA, Kasales CJ, et al. Likelihood of additional work-up among women undergoing routine screening mammography: the impact of age, breast density, and hormone therapy use. Prev Med 2004;39(1): 48–55. [7] Berg WA, Gutierrez L. Diagnostic accuracy of mammography, clinical examination, US, and MR imaging in preoperative assessment of breast cancer. Radiology 2004;233:830–49. [8] Ursin G, Ma H, Wu AH, et al. Mammographic density and breast cancer in three ethnic groups. Cancer Epidemiol Biomarkers Prev 2003;12: 332–8. [9] Majid AS, et al. Missed breast carcinoma pitfalls and pearls. Radiographics 2003;23(4):881. [10] Corsett V, Ferrari A, Ghirardi M. Role of ultrasonography in detecting mammographically occult breast carcinoma in women with dense breasts. Radiology 2006;111:440–8. [11] Zhu Q, Huang M. Ultrasound-guided optical tomographic imaging of malignant and benign breast lesions: initial clinical results of 19 cases. Neoplasia 2003;5(5):379–88.
[12] Orel SG, Schnall MD. MR imaging of the breast for the detection, diagnosis, and staging of breast cancer. Radiology 2001;220:13–30. [13] Folkman J. Tumor angiogenesis: therapeutic implications. N Engl J Med 1971;285:1182–6 2. Gasparini C, Harris A. [14] Battegay F. Angiogenesis: mechanistic in sights, neovascular diseases, and therapeutic prospects. J Mol Med 1995;73:333–46. [15] Berkowitz B, Wilson C. Quantitative mapping of ocular oxygenation using magnetic resonance imaging. Magn Reson Med 1995;33:579–81. [16] Mclaren I, Prentice A, et al. Vascular endothelial growth factor (VEGF) concentrations are elevated in peritoneal fluid of women with endometriosis. Hum Reprod 1996;11:220–3. [17] Arbiser J. Angiogenesis and the skin: a primer. J Am Acad Dermatol 1996;34:486–7. [18] O'Brien E, Garvin M, Dev R, et al. Angiogenesis in human coronary atherosclerotic plaques. Am J Pathol 1994;145:883–94. [19] Van de Wiele C, Oltenfreiter R. Tumour angiogenesis pathways: related clinical issues and implications for nuclear medicine imaging. Eur J Nucl Med 2002;29(5):699–709. [20] Floery D, Helbich TH. Characterization of benign and malignant breast lesions with computed tomography laser mammography (CTLM): initial experience. Invest Radiol 2005;40(6):328–35. [21] Helbich TH, Becherer A, Trattnig S, et al. Differentiation of benign and malignant breast lesions: MR imaging versus Tc-99m sestamibi scintimammography. Radiology 1997;202:421–9. [22] Buadu LD, Murakami J, Murayama S, et al. Breast lesions: correlation of contrast medium enhancement patterns on MR images with histopathologic findings and tumor angiogenesis. Radiology 1996;200: 639–49.
CTLM as an adjunct to mammography in the diagnosis of patients with dense breast Jin Qi, Zhaoxiang Ye⁎ Radiology Department, Tianjin Medical University Cancer Institute and Hospital, Key Laboratory of Breast Cancer Prevention and Therapy of the Ministry of Education, Key Laboratory of Cancer Prevention and Therapy, Tianjin 300060, PR China Received 10 April 2012; accepted 1 May 2012
Abstract Purpose: The purpose was to evaluate the utility of computed tomographic laser mammography (CTLM) as an adjunct examination to mammography in women with dense breast tissue. Methods: We retrospectively compared the findings of mammography, CTLM, and adjunct CTLM to mammography with pathology of 155 women scheduled for biopsy or surgery. Results: Positive lesions were observed more significantly in malignant than benign lesions. The sensitivity of mammography vs. mammography+CTLM was 34.4% vs. 81.57% among extremely dense breasts and 68.29% vs. 95.34% among heterogeneously dense breasts. Conclusion: CTLM could distinguish benign lesions from malignant lesions and is not affected by breast density. © 2013 Elsevier Inc. All rights reserved. Keywords: Computed tomographic laser mammography; Laser technique; Tomography; Optical imaging; Mammography; Breast cancer
1. Introduction Early detection and effective treatment of women with a diagnosis of breast cancer are major factors contributing to the decline in the mortality rate. Radiology examination can help in early detection and diagnosis. At the present time, the means of diagnosis of breasts now used include mammography, ultrasound, computed tomography (CT), magnetic resonance imaging (MRI), near and far infrared, and positron emission tomography (PET)-CT [1]. Mammography is the most widely used; however, conventional screenfilm mammography has limited sensitivity for detection of breast cancer especially in breasts with dense tissue. Digital mammography was developed to address some of the limi⁎ Corresponding author. Radiology Department, Tianjin Medical University Cancer Institute and Hospital, Key Laboratory of Breast Cancer Prevention and Therapy of the Ministry of Education, Key Laboratory of Cancer Prevention and Therapy, Tianjin 300060, PR China. Tel.: +86 22 23359337; fax: +86 22 23359337. E-mail address: [email protected] (Z. Ye). 0899-7071/$ – see front matter © 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.clinimag.2012.05.003
tations of screen-film mammography; however, the value of digital mammography is not substantially different from that of screen-film mammography. Radiology physicians are in need of more functional information. Tumor “angiogenesis” is known to be critical for the autonomous growth and spread of breast cancers. Tumor angiogenesis is a complex process that involves both the incorporation of existing host blood vessels into the tumor and the creation of tumor microvessels [2]. If these vessels could be found, the information of function can be used by physicians. The basic principle underlying the new device, computed tomographic laser mammography (CTLM) imaging, is the “angiogenesis.” In the last several years, optical tomography on breast imaging has gained interest all over the world. Light has been investigated since the late 1920s as a diagnostic tool for breast cancer by transillumination; however, it had low spatial resolution and afforded little in spectral quantification of lesions. Hence, it did not attain sufficient sensitivity and specificity to be used clinically. In recent years, optical mammography stepped to new era. Optical introscopy of laboratory animals is extensively developed for
290
J. Qi, Z. Ye / Clinical Imaging 37 (2013) 289–294
studies of DNA and pharmacological preparations. Optical mammography uses either projection or tomographic apparatus. Optical tomographic mammographs are presently available. CTLM (Imaging Diagnosis System Inc., USA) used in our study is now being clinically tested in several countries; the other apparatuses such as Philips OMPS transmission mammography(Optical Mammo Prototype System, Netherlands), SoftScan Optical projection mammography (Advanced Research Technologies Inc., Canada), and ComfortScan mammography (DOBI Medical International Inc., USA) are all being researched [3–5]. The purpose of the study was to evaluate the utility of computed CTLM as an adjunct examination to mammography in women with heterogeneously dense and extremely dense breast tissue [Breast Imaging Reporting and Data System (BI-RADS)3 and 4].
be proper to the size of patient's breasts. In general, 2-mm slice thickness would be set to most patients; when the breast is too large or too small, 1-mm and 4-mm slice thickness should be set. Before scanning, the breast was moved in the middle of scanning ring. The detector system consists of 2 rings with 84 photodetectors each (picture 1). This unit rotates 360° around the breast and takes 17 s. After each rotation, the ring descends, creating a slice at each step. The scanning platform alternates direction to avoid twisting the cables. The whole breast scan would be made up of 10–40 slices. While slice one level is being scanned, our proprietary software is reconstructing the previous slice. The scan is ready to be read immediately after completion. The device produces striking three-dimensional (3D) views, which can be rotated, in real time, along any axis. CTLM also produces sagittal, axial, and coronal views. 2.4. CTLM image interpretation
2. Materials and methods 2.1. Patients From September 2007 to February 2008, 155 women (23–74 years of age; median age: 41years) in Tianjin Medical University Cancer Hospital were scheduled for biopsy or surgery in the short term (within 30 days). All the patients did not undergo current chemotherapy or irradiation of the breast, and were not biopsied within 30 days before the examination. All the patients underwent mammography and CTLM examination before surgery. According to BI-RADS, all women who underwent mammography were classified into heterogeneously dense breast (BI-RADS3) and extremely dense breast ( BI-RADS4). CTLM was performed subsequently. Exclusion criteria. Included (1) ulcer or wounds on the breast (2) various forms of porphyria (3) current biopsy within 30 days (4) chemotherapy or irradiation of the breast. 2.2. Mammography Bilateral 2-view mammography (Senography 2000D, GE Medical Systems), was performed in the craniocaudal and mediolateral projections. Spot compression and additional views were obtained where appropriate. Two mammography independently analyzed the data for this purpose. The physicians classified all patients into two groups: heterogeneously dense breast group (BI-RADS3) and extremely dense breast group (BI-RADS4). 2.3. CTLM CTLM (CTLM 1020; Imaging Diagnostic Systems, USA) was performed in one side with lesion. The patient lay in the prone position on the examination table with one breast pendant into a scanning chamber, where the breast is surrounded by the laser source detector .The scan ring must
CTLM device produces a laser wavelength (808 nm), in the NIR spectrum, which matches the crossover point of absorption of both oxygenated and deoxygenated hemoglobin. The hurdle that has to be overcome is the difference in the way photons transverse the tissue. Light travels in tissue in a random fashion due to scatter. The average path of light can be predicted in spite of the scattering and absorption of light in tissue through. CTLM uses a large number of source and detector positions in each slice to take into account the diffusion approximation of light propagation in tissue and to show the location of the increased vascularity in the breast. The device produces striking 3D views, which can be rotated, in real time, along any axis. It also produces sagittal, axial, and coronal views. An inversion factor is introduced into the reconstruction algorithm so that areas of high absorption (high hemoglobin concentration) are visualized in white. Relatively avascular areas are shown as green or black. The available tools allow us to focus on any area by removing overlying obscuring areas. Image quality can be improved with window and level controls. Images are usually viewed in a maximal intensity projection (MIP) and then in a front to back reconstruction (FTB), also known as a surface rendering mode. The two modes are used to evaluate the vascularization patterns to determine whether the images represent normal or abnormal vascularization. CTLM images were read by a trained radiology physicians, blinded to any imaging and clinical information. To guarantee the accuracy of the data, the difficult cases were review by Eric Mile, MD (USA). Following are the CTLM interpretive data: 1 stands for benign lesions; 2 stands for malignant lesions, 0 stands for the lesions that were not unascertainable. 2.5. Mammography imaging interpretation Following are the mammography interpretive data: 1 stands for benign lesions record; 2 stands for malignant lesions; 0 stands for the lesions which were not unascertainable.
J. Qi, Z. Ye / Clinical Imaging 37 (2013) 289–294
291
Table 1 CTLM+mammography image interpretation
Table 3 Pathologic findings in 155 lesions
Examination
Interpretation
Pathologic findings
Density of breast
Mammography
Benign
Extremely
Heterogeneously
CTLM Mammography+CTLM
1a 1 1
Benign Fibrocystic changes Inflammation Ductal disease Fibroadenomas or adenofibromas Adenosis Cyst Adiponecrosis Normal tissue
37 16 2 4 9 3 1 1 1
39 9 7 9 11 0 0 0 3
Malignant DCIS Invasive ductal cancer Neuroendocrine carcinoma Mucoid adenocarcinoma Apocrine gland carcinoma Interstitial sarcoma Glycogen-rich clear cell carcinoma Mesenchymal sarcoma
37 4 28 0 0 1 2 1 1
42 5 35 1 1 0 0 0 0
a b c
1 2 2
Malignant 1 0 1
2b 1 2
2 2 2
Not ascertain 0c 1 1
2 0 2
0 2 2
0 0 0
1=benign. 2=malignant. 0=not ascertained.
2.6. Mammography+CTLM image interpretation Combine the result of mammography and CTLM, accept it as malignant when any one was malignant, accept it as benign or malignant when they were the same. If one of them is not unascertainable, record depended on another result, as shown in Table 1. 2.7. Statistics
3.2. CTLM
We analyzed the result of mammography, CTLM, and mammography+CTLM to identify true-positive, true-negative, false-positive, and false-negative examination according to pathology .On the basis of these classifications, we measured sensitivity (true-positive/[true-positive+falsenegative]), specificity (true-negative/[false-positive+truenegative]), and positive predictive value (true-positive/ [true-positive+false-positive]). Statistical analysis was performed with the SPSS statistical package .A χ 2 test was used to compare group differences. 3. Results
Radiology physician recorded CTLM impression, as follows: 35 benign lesions, 34 malignant lesions, and 5 ascertained lesions in extremely dense breast, and 33 benign lesions, 43 malignant lesions, and 5 unascertainable lesions in heterogeneously dense breast group, as shown in Table 2. 3.3. Mammography+CTLM There were 32 benign lesions, 40 malignant lesions, and 2 unascertainable lesions in extremely dense breast, and 23 benign lesions, 57 malignant lesions, and 1 unascertainable lesion in heterogeneously dense breast group, as shown in Table 2.
3.1. Mammography 3.4. Pathologic diagnosis Two radiology physicians reviewed the results and classified all patients into two groups: 74 in heterogeneously dense breast group and 81 in the extremely dense breast group. The impression of mammography is as follows: 48 benign lesions, 15 malignant lesions, and 11 unascertainable lesions in extremely dense breast, and 41 benign lesions, 15 malignant lesions, and 6 unascertainable lesions in heterogeneously dense breast group, as shown in Table 2.
In 155 lesions, 150 underwent surgery; 3, biopsy; and 2, needle biopsy. In the 155 breast lesions studied, 76 were malignant and 79 were benign at pathologic analysis. The tumor stages of the 155 breast cancers are listed in Table 3. From our results, CTLM positive predictive value is 72.41% (63/63+24). It supported that CTLM successfully shows the angiogenesis of malignant lesion in breast,
Table 2 Findings in three examinations Breast density
n
Mammography
Rank Extremely (n) Heterogeneously (n)
74 81
1 48 41
2 15 34
CTLM 0 11 6
1 35 33
Mammography+CTLM 2 34 43
0 5 5
1 32 23
2 40 57
0 2 1
292
J. Qi, Z. Ye / Clinical Imaging 37 (2013) 289–294
Table 4 CTLM findings in benign and malignant lesions Pathology Benign Malignant
Table 6 Statistical data
−
+ 24 57
Breast density χ =25.558 P=.000
Mammography
CTLM
Sens
Sens
Mammography +CTLM
2
52 22
+, angiogenesis was shown in CTLM; −, angiogenesis was not shown.
Spec
Spec
Sens
Spec
Extremely 34.40% 90.48% 74.4% 71.00% 81.57% 78.13% Heterogeneously 68.29% 85.00% 85.00% 61.00% 95.34% 55.26% As shown in Table 7: sensitivity of CTLM in heterogeneously dense breast.
However, 57% (24/76) of benign lesions also showed angiogenesis. Among these cases, there are eight fibrocystic changes, six inflammations, four ductal diseases, two adenoses, one fibroadenoma, and one adiponecrosis. Increased absorption was observed significantly more often in malignant than in benign lesions [72.15% (57/79) vs. 31.57% (24/76), χ 2= 25.558, P=.000), as shown in Table 4. Among extremely dense breast, the sensitivity of mammography alone, CTLM alone, and mammography+CTLM was, respectively, 34.40%, 74.40%, and 81.57%. The sensitivity of mammography vs. mammography +CTLM was 34.4% vs. 81.57% (χ 2=13.071, P=.000). The specificity of mammography alone, CTLM alone, and mammography +CTLM was 90.48%, 71.00%, and 78.13%. Among heterogeneously dense breast, the sensitivity of mammography alone, CTLM alone, and mammography + CTLM was 68.29%, 85.00%, and 95.34%. The sensitivity of mammography vs. mammography+CTLM was 68.29% vs. 95.34% (χ 2= 11.131, P=.001). The specificity of mammography alone, CTLM alone, and mammography+CTLM was 85.00%, 61.00%, and 55.26%. The details are shown in Tables 5 and 6. The percentage of dense breast vs. extremely dense breast was 74.40% vs. 85.00% (χ 2=0.446, P=.504); specificity in them was 71.00% vs. 61.00% (χ 2= 0.0001, P= 1.0). 4. Discussion Mammography is the golden standard of breast imaging diagnosis; it has high sensitivity in fatty breast with sharp contrast [6]. Screening mammography and digital mammography are of limited value in dense breasts. In extremely dense and heterogeneously dense breasts, mammography sensitivity is decreased. Mammography sensitivity was 45% to 70% [7]. In our study, in extremely dense breast, the sensitivity of digital mammography was 34.4%, lower than the literature reported. In heterogeneously dense breast, mammographic sensitivity was 68.29%, in accordance with the literature reports. In our study, the specificity declined
from 90.48% to 85%. Relevance research is indicated. The risk of cancer is increased from 2.5 to 5 times with the density ascending. So it is important for the malignant lesions to be detected early in dense breast [8]. Because of the limit of mammography, radiology physicians turn to ultrasonography. In simple cysts, the accuracy of ultrasound is 96% to 100%. However, when ultrasound imaging is used for the differentiation of hyperplasia from solid lesions that mammography could not distinct, it is not high as a result of the overlapping characteristics of solid benign and malignant lesions [9–11]. In the last several years, functional imaging, such as MRI and PET-CT, has gained interest and been paid more attention to. MRI can provide important information in function with high sensitivity in detecting invasive ductal cancer and ductal carcinoma in situ [12]. A new field of research is laser-light-based breast imaging. Initial attempts to use laser light for “transillumination” and detect the angiogenesis of the breast have been developed in the early 20th century. A computed tomographic laser lightbased scanner for the breast, called CTLM, has now been used clinically and has gained some initial experiment. Our study showed that the sensitivity was significantly increased by using CTLM as an adjunct to mammography, from 34.4% to 81.57% (χ 2= 13.071, P=.000), and, in heterogeneously dense breast, from 68.29% to 95.34% (χ 2= 11.131, P=.001). In 1971, Folkman [13] first proposed that tumor cells were capable of stimulating endothelial cell proliferation by means of a “soluble tumor angiogenesis factor". Angiogenesis is defined as the formation of new blood vessels through the sprouting of capillaries from preexisting microvessels [14]. This process serves as the fundamental method by which neovascularization occurs in the human body. In most mature tissues, the rate of endothelial cell turnover is extremely slow, on the order of years. A markedly increased rate of angiogenesis takes place during the normal physiologic processes of embryogenesis, uterine maturation, placental development, corpus luteum formation, and wound healing. In
Table 5 Imaging findings according to pathology Breast density
Extremely Heterogeneously
Mammography
CTLM
Mammography +CTLM
TP
FN
TN
FP
TP
FN
TN
FP
TP
FN
TN
FP
11 28
21 13
38 34
4 6
29 34
10 11
25 22
10 14
31 41
7 2
25 21
11 17
J. Qi, Z. Ye / Clinical Imaging 37 (2013) 289–294 Table 7 CTLM in extremely and heterogeneously dense breasts Breast density
CTLM findings TP
FN
Extremely Heterogeneously
29 34
10 11
χ2=0.446 P=.504
TN
FP
25 22
10 14
χ2=0.0001 P=1.000
each of these instances, the angiogenic process is strictly regulated and is terminated following completion of its intended function. Research has shown, however, that regulated or unregulated angiogenic activity plays a role in a number of different diseases. Examples include diabetic retinopathy, in which associated ocular angiogenesis is the leading cause of blindness in developed countries. A possible role of angiogenesis in the development of atherosclerosis is the focus of an ongoing study .In the most extreme case, angiogenesis has been shown to play an important role in the pathogenesis of tumor growth and metastasis [15–18]. Tumor cells control neovascularization through secretion of angiogenic factors which attract endothelial cells that proliferate and invade the stroma towards the tumor mass; then “angiogenesis” could be found. Tumors are unable to grow larger than circa 1 mm 3 without developing a new, hypoxia-triggered, blood supply [19]. Based on this theory, the CTLM device uses a laser wavelength (808 nm), in the NIR spectrum, which matches the crossover point of absorption of both oxygenated and deoxygenated hemoglobin. This wavelength also minimizes the effects of both fat and water [20]. Laser used by CTLM can so easily penetrate the breast tissue and is not affected by the density. Our study supported it: the CTLM sensitivity of heterogeneously dense breast vs. extremely dense breast was similar (74.40% vs. 85.00%, P=.504), and specificity was also similar(71.00% vs. 61.00%, P= 1.000). The results of the data supported that the results were not varied from heterogeneously dense breast to extremely dense breast. It has been shown that CTLM was not affected by tissue density in showing the angiogenesis. CTLM has the potential to be used as an adjunct tool to mammography in both heterogeneously dense breast and extremely dense breast. We are therefore looking at the absorption pattern of hemoglobin in vessels. In tumor angiogenesis, the vessels are concentrated and are structurally and functionally abnormal. This gives a greater volume of hemoglobin in a confined area that can be visualized by CTLM. In the analysis of cases in this study based on this principle, CT features of malignant angiogenesis are as follows: (a) Abnormal shape: CTLM showed both forms of normal and abnormal deoxyhemoglobin and hemoglobin. It is important for physicians to distinguish normal vein from abnormal “angiogenesis.” On coronal views, abnormal “angiogenesis” showed round and ovoid areas of high absorption (high hemoglobin concentration) visualized in white and distinguished from triangle and cone-shaped areas (picture 2). On 3D-MIP view, the irregular shape of “angiogenesis” showed the following: oblate sphere, dumbbell
293
shape, diverticulum, circle, distinguished from the clear “tunnels.” (b) Isolated areas of high absorption : The normal vessels are shown as “ribbon” running through the breast from chest wall to nipple in MIP-3D view. The branches of vessels shown as star-shaped and branches-shaped could always be found. The oblate areas of high absorption could be found at the nipple area. This high absorption is normal vessel but not “angiogenesis”. The normal vessel at areola is always surrounded by some veins but not isolated. However, the “angiogenesis” of malignant tumor always shows high absorption isolated which is not surrounded by vein. (c) Furthermore, “angiogenesis” of malignant tumor located deep in the breast. The normal vessel of breast is always on the surface of the breast but not deep in the breast. In this study, the high absorption showed in deep tissue is almost “angiogenesis” of malignant tumor except six cases are normal veins. The surface and deep imaging could be distinguished by comparing 3D-MIP mode with 3D-FTB mode. The high absorption on the surface of the breast is always benign lesion such as adiponecrosis and mastitis. These benign lesions could be distinguished from deep malignant absorption. These findings of CTLM help to distinguish benign lesions and malignant lesions and solve the problem of dense breasts. From our results, CTLM positive predictive value was 72.41% (63/63+24). It supported that CTLM successfully shows the angiogenesis of malignant lesion in breast. However, some benign lesion also showed angiogenesis. Although there are some overlap between benign and malignant findings, increased absorption was observed significantly more often in malignant than in benign lesions (72.15% vs. 31.57%, χ 2= 25.558, P=.000). In this hypothesis, a CTLM scanner was used to characterize benign and malignant breast lesions. We can conclude that CTLM could distinguish malignant from benign lesion. Some literature reported that [21,22]. A variety of benign lesions also demonstrated increased vessel as malignant lesions. Determination of confirmatory histologic parameters of breast lesions, such as microvessel counts, would be able to provide additional information about lesion biology. So we concluded that CTLM is able to characterize benign and malignant breast lesions by showing the angiogenesis. Our data indicate that the imaging of CTLM was not affected by tissue density in breast, and provide information about angiogenesis in breast lesions, especially in malignant lesions, when used as an adjunct to mammography in heterogeneously dense breast and extremely dense breast. Sensitivity increased significantly. This study suggests that, in clinical practice, adding CTLM in dense breasts may be useful. References [1] Bao R, Ye Z, Liu P. Make much of breast imaging diagnosis. Chin J Radiol 2007;41. [2] Zhu Q, Cronin EB. Benign versus malignant breast masses: optical differentiation with US-guided optical imaging reconstruction. Radiology 2005;237:57–66.
294
J. Qi, Z. Ye / Clinical Imaging 37 (2013) 289–294
[3] Ntziachristos V, Yodh AG. Concurrent MRI and diffuse optical tomography of breast after indocyanine green enhancement. Proc Natl Acad Sci U S A 2000;97(6):2767–72. [4] Kolb TM, Lichy J. Comparison of the performance of screening mammography, physical examination, and breast US and evaluation of factors that influence them: an analysis of 27,825 patient evaluations. Radiology 2002;225:165–75. [5] Podgaetsky VM, Potapov DA, et al. Development of optical introscopy in medical engineering. Biomed Eng 2007;41(1): 2–6. [6] Carney PA, Kasales CJ, et al. Likelihood of additional work-up among women undergoing routine screening mammography: the impact of age, breast density, and hormone therapy use. Prev Med 2004;39(1): 48–55. [7] Berg WA, Gutierrez L. Diagnostic accuracy of mammography, clinical examination, US, and MR imaging in preoperative assessment of breast cancer. Radiology 2004;233:830–49. [8] Ursin G, Ma H, Wu AH, et al. Mammographic density and breast cancer in three ethnic groups. Cancer Epidemiol Biomarkers Prev 2003;12: 332–8. [9] Majid AS, et al. Missed breast carcinoma pitfalls and pearls. Radiographics 2003;23(4):881. [10] Corsett V, Ferrari A, Ghirardi M. Role of ultrasonography in detecting mammographically occult breast carcinoma in women with dense breasts. Radiology 2006;111:440–8. [11] Zhu Q, Huang M. Ultrasound-guided optical tomographic imaging of malignant and benign breast lesions: initial clinical results of 19 cases. Neoplasia 2003;5(5):379–88.
[12] Orel SG, Schnall MD. MR imaging of the breast for the detection, diagnosis, and staging of breast cancer. Radiology 2001;220:13–30. [13] Folkman J. Tumor angiogenesis: therapeutic implications. N Engl J Med 1971;285:1182–6 2. Gasparini C, Harris A. [14] Battegay F. Angiogenesis: mechanistic in sights, neovascular diseases, and therapeutic prospects. J Mol Med 1995;73:333–46. [15] Berkowitz B, Wilson C. Quantitative mapping of ocular oxygenation using magnetic resonance imaging. Magn Reson Med 1995;33:579–81. [16] Mclaren I, Prentice A, et al. Vascular endothelial growth factor (VEGF) concentrations are elevated in peritoneal fluid of women with endometriosis. Hum Reprod 1996;11:220–3. [17] Arbiser J. Angiogenesis and the skin: a primer. J Am Acad Dermatol 1996;34:486–7. [18] O'Brien E, Garvin M, Dev R, et al. Angiogenesis in human coronary atherosclerotic plaques. Am J Pathol 1994;145:883–94. [19] Van de Wiele C, Oltenfreiter R. Tumour angiogenesis pathways: related clinical issues and implications for nuclear medicine imaging. Eur J Nucl Med 2002;29(5):699–709. [20] Floery D, Helbich TH. Characterization of benign and malignant breast lesions with computed tomography laser mammography (CTLM): initial experience. Invest Radiol 2005;40(6):328–35. [21] Helbich TH, Becherer A, Trattnig S, et al. Differentiation of benign and malignant breast lesions: MR imaging versus Tc-99m sestamibi scintimammography. Radiology 1997;202:421–9. [22] Buadu LD, Murakami J, Murayama S, et al. Breast lesions: correlation of contrast medium enhancement patterns on MR images with histopathologic findings and tumor angiogenesis. Radiology 1996;200: 639–49.