- Email: [email protected]
Relationship between elasticity and collagen fiber content in breast disease: A preliminary report
Ultrasonics 57 (2015) 44–49
Contents lists available at ScienceDirect
Ultrasonics journal homepage: www.elsevier.com/locate/ultras
Relationship between elasticity and collagen fiber content in breast disease: A preliminary report Zhi Li Wang a,⇑, Lu Sun b, Ye Li a, Nan Li a a b
Department of Ultrasound, Chinese People’s Liberation Army General Hospital, 28 Fuxing Road, Beijing 100853, China Department of Pathology, Chinese People’s Liberation Army General Hospital, 28 Fuxing Road, Beijing 100853, China
a r t i c l e
i n f o
Article history: Received 19 July 2014 Received in revised form 13 October 2014 Accepted 20 October 2014 Available online 28 October 2014 Keywords: Breast Elasticity Collagen fiber Stroma
a b s t r a c t Objective: To investigate the differences in elasticity and collagen fiber content between malignant and benign breast lesions, and to study the relationship between shear wave elasticity and the content of collagen fiber in extracellular matrix (ECM). Materials and methods: Between May 2012 to May 2013, 106 patients with 116 breast lesions who were referred to our center for ultrasound-guided biopsy of a sonographically apparent breast lesion underwent shear wave elasticity examination. The specimen underwent Van Gieson (VG) dye and Image-Pro Plus 5.1 software was used to quantitatively analyze area of collagen fiber. Results: Malignant lesions exhibited significantly higher maximum elasticity, mean elasticity, and elasticity ratio between lesions and surrounding parenchyma (140.43 ± 70.16 kPa, 63.11 ± 33.68 kPa, 3.49 ± 1.95) than benign lesions (54.64 ± 48.53 kPa, 34.52 ± 25.23 kPa, 2.25 ± 1.48) (t = 5.329, t = 4.382, t = 4.487, P < 0.001). The content of collagen fiber of malignant lesions was significantly higher than that of benign lesions (t = 8.437, p = 0.000). There was a positive correlation between max elasticity and the content of fiber collagen (r = 0.746). Conclusion: The elasticity of breast lesions has a close correlation with the content of collagen fiber, which might have an important impact on tissue stiffness of breast lesions. Ó 2014 Elsevier B.V. All rights reserved.
1. Introduction Elastography is an emerging imaging technique that quantifies the ‘‘stiffness’’ of a breast lesion [1,2]. Studies had shown that the increase of matrix stiffness was one of the characteristics of cancer and stiffness had already been applied in the cancer detection [3,4]. In prior studies, elastography imaging had shown potential for differentiating benign from malignant breast disease and could possibly reduce the overall number of breast biopsies [5,6]. Most of previous studies used elastography scoring (ES) or strain ratio measurement (SR) as the diagnostic parameter, both of which were semi-quantitative parameters. Both of these semi-quantitative parameters were highly dependent on the organ’s compressibility limits under stress and on the skill of the operator to correctly compress the tissue [7]. SWE is a new method of obtaining elastography images, where an acoustic pressure wave induces slow-moving shear waves within the tissue, and the speed of propagation of the shear wave is proportional to the tissue’s elastic stiffness. Shear waves travel ⇑ Corresponding author. Tel.: +86 10 66936848; fax: +86 10 68161218. E-mail address: [email protected] (Z.L. Wang). http://dx.doi.org/10.1016/j.ultras.2014.10.016 0041-624X/Ó 2014 Elsevier B.V. All rights reserved.
relatively faster in stiffer tissue compared to softer tissues. Ultrafast imaging of the propagation of shear waves allows measurement of small changes in velocity that occur when the waves pass through tissues of different stiffness. The velocity information can be mapped to create an image of the stiffness, with the option of measuring SWE features such as the minimum, mean, and maximum elasticity in a region of interest. Deformation of tissue leading to shear waves is created by an acoustic impulse that is generated electronically. With this method, the radiation force produced by the probe is used generate shear waves. This approach is different from conventional quasi-static elastography where the compression is applied externally by the operator. The study by Evans et al. [8] demonstrated that shear wave elastography could give quantitative and reproducible information on solid breast lesions with diagnostic accuracy at least as good as conventional ultrasound with BI-RADS classification. Our previous study demonstrated that when optimal cut-off value of 91.53 kPa was used for max elasticity in the differentiation of breast lesions, the diagnostic sensitivity and specificity were 60.9% and 85.3%, respectively [9]. All these studies showed that shear wave elastography could give quantitative elasticity information that potentially could help in breast lesion characterization.
45
Z.L. Wang et al. / Ultrasonics 57 (2015) 44–49
Although malignant breast tumors exhibited stiff properties compared with benign breast tumors, the reasons for this difference remained to be elucidated. The extracellular matrix (ECM) in primary breast cancers was significantly changed compared with normal breast tissue [10,11]. Collagen fiber is the most abundant structural protein in ECM [12], and increased collagen fiber I had been found to facilitate breast tumor formation, invasion, and metastasis [13,14]. Study has demonstrated that, in cancer progression, there were changes of ECM, such as increasing collagen expression, collagen deposition and structure changes [15]. X-ray mammography, which detected dense fibroglandular tissue in the breast, demonstrated that women with high (50–74% versus <5%) breast density have a 4.64-fold increased risk of developing breast carcinoma [16]. Levental discovered that breast tumorigenesis was accompanied by collagen crosslinking, ECM stiffening, and increased focal adhesions. Induction of collagen crosslinking stiffened the ECM, promoted focal adhesions, enhanced PI3 kinase (PI3K) activity, and induced the invasion of an oncogene-initiated epithelium [17]. Studies also had demonstrated that the proportion of collagen fiber areas was the strongest pathologic determinant of mean stiffness in hepatocellular carcinomas [18]. To our knowledge, there are no investigations comparing elasticity with collagen fiber content in breast lesions. The purpose of this study was to investigate the differences in elasticity and collagen fiber content between malignant and benign breast lesions, and to study the relationship between shear wave elasticity and the content of collagen fiber in ECM. 2. Material and methods 2.1. Patients Between March 2012 to March 2013, a prospective study was conducted at our institute. The study population consisted of consecutive patients referred to our center for ultrasound-guided biopsy of a sonographically apparent breast lesion. Subjects were included following their consent for the index testing. We excluded pregnant and lactating women, those with breast implants, women receiving chemotherapy or radiotherapy for any cancer, skin masses and any which had been biopsied, and patients with a history of ipsilateral breast surgery. Pathologic diagnosis was used as the reference standard. One hundred and six consecutive patients with one hundred and sixteen solid lesions were included in this study. Of these 106 women, 61 women were asymptomatic, 45 women presented
with a palpable mass, and two showed nipple discharge. The age of the patients was 22–82 years (mean age ±standard deviation, 51.8 ± 28.3 years). Of these patients, 96 patients had a single nodule, and 10 patients had two nodules. The maximum diameter of the nodules ranged from 0.7 to 4.6 cm (mean diameter ± standard deviation, 2.1 ± 2.0 cm). Informed consent was obtained from all patients, and the study was approved by our local Ethics Committee. Written informed consent was obtained from every patient at enrollment. 2.2. US examination and SWE examination US examination and SWE examination were performed with AixplorerÒ ultrasound system (SuperSonic Imagine, Aix en Provence, France) with the center frequency of the probe was 12 MHz. The ultrasound images underwent American College of Radiology Breast Imaging Reporting and Data System (ACR BI-RADS) [14] classification by two breast radiologists (Zhi Li Wang and Nan Li) who have twelve years and ten years experience in breast ultrasound, respectively. After the ultrasound examination, the radiologist switched to the elastography model. The probe was applied as light as possible to give no pressure to the lesion. The probe needs to be kept still for 10–20 s during acquisition of the elastography images (due to a slow frame rate) and this was often best done during a breath hold. The elastography views selected were those most clearly displaying abnormal stiffness within the plane but with the absence of movement or pressure artifact, such as the red area under the probe. After the stable image was obtained, the image was recorded and region of interest (ROI) was chosen to calculate the elasticity value. The ROI was chosen as large as possible to cover the whole lesion including calcification potentially present and the edge of the lesion. Because the ROI box was circle, it was hard to cover the whole lesion, especially for the lesions with irregular edge. But we tried our best to cover most part of the lesion, especially for the hardest part of the lesion. Another ROI in peripheral parenchyma was selected to be at a similar depth to that of the breast lesion. For each patient, three ROIs in the lesion and peripheral parenchyma, respectively, were selected and the mean value was regarded as the final value. The ROI in peripheral parenchyma was made as much as possible to be of the same size and depth of the corresponding breast lesion. The maximum value within the ROI, called max elasticity, the mean elasticity, the minimum value within the ROI, called min elasticity and elasticity ratio between lesions and surrounding parenchyma were recorded.
Table 1 Maximum, mean and minimum elasticity and elasticity ratio between lesions and surrounding parenchyma of malignant and benign lesions.
Malignant lesions Benign lesions t test p-value
Max elasticity (kPa)
Mean elasticity (kPa)
Min elasticity (kPa)
Elasticity ratio
140.43 ± 70.16 54.64 ± 48.53 6.329 0.000
63.11 ± 33.68 34.52 ± 25.23 5.382 0.000
25.48 ± 19.85 18.86 ± 15.95 0.567 0.356
3.49 ± 1.95 2.25 ± 1.48 4.487 0.001
Table 2 Maximum, mean and minimum elasticity and elasticity ratio between lesions and surrounding parenchyma among different pathology lesions. Pathological diagnosis
Maximum elasticity (kPa)
Mean elasticity (kPa)
Minimum elasticity (kPa)
Mean elasticity ratio
IDC (n = 51) DCIS (n = 19) ILC (n = 5) Fibroadenoma (n = 19) Fibroadenoses (n = 12) Papillomas (n = 6) Inflammation (n = 4)
142.45 ± 68.37 134.17 ± 61.45 138.89 ± 64.78 53.25 ± 49.63 55.14 ± 50.69 63.62 ± 75.02 44.26 ± 39.47
64.12 ± 33.97 60.62 ± 36.59 61.38 ± 29.43 35.03 ± 29.78 30.84 ± 27.82 40.24 ± 45.47 33.72 ± 33.48
26.31 ± 21.76 25.15 ± 19.11 25.01 ± 20.45 17.58 ± 10.94 23.13 ± 29.31 24.63 ± 21.74 13.58 ± 10.32
3.51 ± 1.89 3.35 ± 1.68 3.26 ± 1.73 2.04 ± 1.22 2.43 ± 1.94 2.57 ± 2.13 2.28 ± 2.14
46
Z.L. Wang et al. / Ultrasonics 57 (2015) 44–49
Fig. 1. Elastography and collagen fiber content of breast carcinoma. (A) A lesion in the upper outer quadrant of the right breast, which was proved to be IDC by pathology; (B) elastography with supersonic shear imaging showed its relatively high stiffness, with maximum elasticity 145.71 kpa; (C) collagen fiber dye (400), the red regions were recognized as collagen fiber areas and collagen fiber was major distributed in mesenchyma; (D) quantitative analysis showed the average area of collagen fiber was 12313.80 ± 4625.25. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Z.L. Wang et al. / Ultrasonics 57 (2015) 44–49
47
Fig. 2. Elastography and collagen fiber content of a benign breast lesion. (A) A lesion in the upper outer quadrant of the left breast, which was proved to be adenosis by pathology; (B) elastography with supersonic shear imaging showed its maximum elasticity was 51.31 kpa; (C) collagen fiber dye (400), the red regions were recognized as collagen fiber areas and collagen fiber was major distributed in basement membrane and mesenchyma; (D) quantitative analysis showed the average area of collagen fiber was 7528.43 ± 4625.25. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
48
Z.L. Wang et al. / Ultrasonics 57 (2015) 44–49
Two independent expert observers performed a blinded review of the static images of all lesions. 2.3. Biopsy procedure 16G core needle biopsy (CNB) are routinely used. All biopsies were performed by two radiologists (Zhi Li Wang and Nan Li) who have ten years and seven years experience in interventional experience, respectively. Two or three core samples were obtained in the area with largest elasticity. The appearance and behavior of the formalin-fixed core samples were examined during the procedure to confirm that the targeted lesion was adequately sampled. 2.4. EVG staining Slides were deparaffinized and hydrated to distilled water; stained in Weigert iron hematoxylin for 10 min, and then rinsed in tap water. After a quick differentiation in 1% hydrochloric acid alcohol, slides were washed in running water for 2 min. After Counterstain in Van Gieson’s solution for 2 min, the slides received quick differentiation by 95% alcohol for a few seconds, then dehydrated in 100% alcohol. After Cleared in xylene for 3 min, slides were coversliped with resinous mounting medium and observed under the microscope. 2.5. Quantification of collagen fiber Image-Pro Plus 5.1 Software was used to quantitatively analyze collagen fiber area. Computer graphic analysis was performed under detailed measurement settings by a board-certified pathologist. 5 fields under high magnification (400) were randomly selected for each slide and the mean value was calculated. 2.6. Statistical analyses All analyses were performed using SPSS11.0, standard version (SPSS Inc, Chicago, IL). All data were described as mean ± SD. The Student’s t-test was used to assess the differences of elasticity and collagen fiber area between benign and malignant lesions. Pearson correlation analysis was used to analyze the correlation between elasticity and collagen fiber area. P < 0.05 was considered statistically significant.
For max elasticity, the optimal cut-off value, yielding the maximal sum of sensitivity and specificity, was 92.32 kPa. With this cut-off value, the sensitivity, specificity, PPV, NPV and accuracy for max elasticity were 84.3%, 92.6%, 78.7%, 77.5%, 84.5%, respectively. 3.3. Comparison of collagen fiber area between malignant and benign lesions The mean areas of collagen fiber of benign and malignant lesions of breast were (7625.14 ± 5321.28) lm2 and (11782.26 ± 8627.54) lm2, respectively. The mean collagen fiber area of malignant lesions was significantly larger than that of the benign lesions (t = 8.437, p = 0.000) (Figs. 1 and 2). The collagen fiber areas of lesions with different pathology types could be seen in Table 3. 3.4. Correlation between elasticity and collagen fiber area Pearson correlation analysis showed a positive correlation between max elasticity of breast lesions and fiber collagen area (r = 0.746) (Fig. 3). 4. Discussion This study demonstrated that max, mean elasticity and elasticity ratio between lesions and surrounding parenchyma of malignant lesions were significantly higher than those in benign lesions, which further demonstrated the important role of elasticity in the differential diagnosis of breast lesions. This study showed that the mean collagen fiber area of malignant breast lesions was significantly larger than that of benign
Table 3 Collagen fiber areas of different pathological types of breast. Pathological diagnosis
Collagen fiber content
IDC (n = 176) DCIS (n = 27) ILC(n = 3) Fibroadenoma (n = 64) Fibroadenosis (n = 32) Papilloma (n = 8) Inflammation (n = 6)
12011.38 ± 81589.34 10623.58 ± 78364.78 10864.75 ± 80585.27 7735.54 ± 5015.56 7536.72 ± 5218.71 8065.56 ± 6188.58 7356.18 ± 5789.47
3. Results 3.1. Pathology Of the 116 masses, 75 (64.7%) were benign and 41 (35.3%) were malignant. Malignant masses included invasive ductal carcinomas (n = 51), ductal carcinoma in situ (DCIS) (n = 19) and invasive lobular carcinoma (ILC) (n = 5). The benign lesions consisted of 19 fibroadenomas, 12 fibroadenoses, 6 papillomas, and 4 inflammations. 3.2. Comparison of elasticity between malignant and benign lesions The max elasticity, mean elasticity, min elasticity and elasticity ratio between lesions and surrounding parenchyma of lesions with different pathology could be seen in Table 1. Malignant lesions exhibited significantly higher max elasticity, mean elasticity, and elasticity ratio between lesions and surrounding parenchyma than benign lesions (t = 5.329, t = 4.382, t = 4.487, P < 0.001). The elasticity parameters of lesions with different pathology types could be seen in Table 2.
Fig. 3. As the Scatter shows, the maximum elasticity and collagen fiber area of breast lesions were positively linear correlated (r = 0.746).
Z.L. Wang et al. / Ultrasonics 57 (2015) 44–49
lesions, which demonstrated that collagen fiber content might play an important role in the progression of breast carcinoma. This study applied Pearson correlation analysis to compare the max elasticity of breast lesion with collagen fiber area, which demonstrated that there was a positive linear correlation between them. It demonstrated that collagen fiber content might have an important impact on tissue stiffness of breast lesions. Our study had several limitations. First, we did not explore associations between stiffness and the other component of ECM, such as elastic fiber, laminin, and fibronectin (FN). However, collagen is the most abundant structural proteins in ECM and its correlation with elasticity could be most important. Second, we did not detect different types of collagen fiber in the breast lesions. It has been demonstrated that type I collagen fiber was the main collagen fiber that prop up the mammary stroma in breast carcinoma [4,5], and further study would be conducted on the impact of different types of collagen fiber on the stiffness of breast lesions. In general, our research demonstrated that the elasticity of breast lesions was closely related to the collagen fiber contents and the collagen fiber content might have an important impact on tissue stiffness of breast lesions. Acknowledgment Financial support from the National Natural Science Foundation (81101059) is gratefully acknowledged. References [1] R.M. Lerner, S.R. Huang, K.J. Parker, ‘‘Sonoelasticity’’ images derived from ultrasound signals in mechanically vibrated tissues, Ultrasound Med. Biol. 16 (3) (1990) 231–239. [2] T.A. Krouskop, T.M. Wheeler, F. Kallel, et al., Elastic moduli of breast and prostate tissues under compression, Ultrason. Imaging 20 (4) (1998) 260–274. [3] A. Modesti, G. D’Orazi, S. Scarpa, et al., Ultrastructural and immunoelectron microscopic study of the desmoplastic stroma in carcinoma of the breast, G. Chir. 10 (5) (1989) 245–249.
49
[4] C. Clavel, P. Birembaut, J.J. Adnet, et al., Breast carcinomas and the extracellular matrix, Ann. Pathol. 8 (2) (1988) 107–113. [5] G.M. Giuseppetti, A. Martegani, B. Di Cioccio, et al., Elastosonography in the diagnosis of the nodular breast lesions: preliminary report, Radiol Med 110 (1– 2) (2005) 69–76. [6] Itoh, E. Ueno, E. Tohno, et al., Breast disease: clinical application of US elastography for diagnosis, Radiology 239 (2006) 341–350. [7] F. Fleury Ede, J.C. Fleury, S. Piato, et al., New elastographic classification of breast lesions during and after compression, Diagn. Interv. Radiol. 15 (2) (2009) 96–103. [8] A. Evans, P. Whelehan, K. Thomson, et al., Quantitative shear wave ultrasound elastography: initial experience in solid breast masses, Breast Cancer Res. 12 (6) (2010) R104. [9] Z.L. Wang, J.L. Li, M. Li, et al., Study of quantitative elastography with supersonic shear imaging in diagnosis of breast tumors, Radiol. Med. 118 (4) (2013) 583–590. [10] A. Bergamaschi, E. Tagliabue, T. Sorlie, et al., Extracellular matrix signature identifies breast cancer subgroups with different clinical outcome, J. Pathol. 214 (3) (2008) 357–367. [11] G. Falzon, S. Pearson, R. Murison, Analysis of collagen fibre shape changes in breast cancer, Phys. Med. Biol. 53 (23) (2008) 6641–6652. [12] L. Kolacna, J. Bakesova, F. Varga, et al., Biochemical and biophysical aspects of collagen nanostructure in the extracellular matrix, Physiol. Res. 56 (Suppl 1) (2007) S51–S60. [13] S. Alowami, S. Troup, S. Al-Haddad, et al., Mammographic density is related to stroma and stromal proteoglycan expression, Breast Cancer Res. 5 (5) (2003) R129–R135. [14] T. Li, L. Sun, N. Miller, et al., The association of measured breast tissue characteristics with mammographic density and other risk factors for breast cancer, Cancer Epidemiol Biomarkers Prev 14 (2) (2005) 343–349. [15] S. Jodele, L. Blavier, J.M. Yoon, et al., Modifying the soil to affect the seed: role of stromal-derived matrix metalloproteinases in cancer progression, Cancer Metastasis Rev. 25 (1) (2006) 35–43. [16] L.J. Martin, N.F. Boyd, Mammographic density. Potential mechanisms of breast cancer risk associated with mammographic density: hypotheses based on epidemiological evidence, Breast Cancer Res. 10 (1) (2008) 201. [17] K.R. Levental, H. Yu, L. Kass, et al., Matrix crosslinking forces tumor progression by enhancing integrin signaling, Cell 139 (5) (2009) 891–906. [18] M. Honjo, F. Moriyasu, K. Sugimoto, et al., Relationship between the liver tissue shear modulus and histopathologic findings analyzed by intraoperative shear wave elastography and digital microscopically assisted morphometry in patients with hepatocellular carcinoma, J. Ultrasound Med. 33 (1) (2014) 61–71.
Contents lists available at ScienceDirect
Ultrasonics journal homepage: www.elsevier.com/locate/ultras
Relationship between elasticity and collagen fiber content in breast disease: A preliminary report Zhi Li Wang a,⇑, Lu Sun b, Ye Li a, Nan Li a a b
Department of Ultrasound, Chinese People’s Liberation Army General Hospital, 28 Fuxing Road, Beijing 100853, China Department of Pathology, Chinese People’s Liberation Army General Hospital, 28 Fuxing Road, Beijing 100853, China
a r t i c l e
i n f o
Article history: Received 19 July 2014 Received in revised form 13 October 2014 Accepted 20 October 2014 Available online 28 October 2014 Keywords: Breast Elasticity Collagen fiber Stroma
a b s t r a c t Objective: To investigate the differences in elasticity and collagen fiber content between malignant and benign breast lesions, and to study the relationship between shear wave elasticity and the content of collagen fiber in extracellular matrix (ECM). Materials and methods: Between May 2012 to May 2013, 106 patients with 116 breast lesions who were referred to our center for ultrasound-guided biopsy of a sonographically apparent breast lesion underwent shear wave elasticity examination. The specimen underwent Van Gieson (VG) dye and Image-Pro Plus 5.1 software was used to quantitatively analyze area of collagen fiber. Results: Malignant lesions exhibited significantly higher maximum elasticity, mean elasticity, and elasticity ratio between lesions and surrounding parenchyma (140.43 ± 70.16 kPa, 63.11 ± 33.68 kPa, 3.49 ± 1.95) than benign lesions (54.64 ± 48.53 kPa, 34.52 ± 25.23 kPa, 2.25 ± 1.48) (t = 5.329, t = 4.382, t = 4.487, P < 0.001). The content of collagen fiber of malignant lesions was significantly higher than that of benign lesions (t = 8.437, p = 0.000). There was a positive correlation between max elasticity and the content of fiber collagen (r = 0.746). Conclusion: The elasticity of breast lesions has a close correlation with the content of collagen fiber, which might have an important impact on tissue stiffness of breast lesions. Ó 2014 Elsevier B.V. All rights reserved.
1. Introduction Elastography is an emerging imaging technique that quantifies the ‘‘stiffness’’ of a breast lesion [1,2]. Studies had shown that the increase of matrix stiffness was one of the characteristics of cancer and stiffness had already been applied in the cancer detection [3,4]. In prior studies, elastography imaging had shown potential for differentiating benign from malignant breast disease and could possibly reduce the overall number of breast biopsies [5,6]. Most of previous studies used elastography scoring (ES) or strain ratio measurement (SR) as the diagnostic parameter, both of which were semi-quantitative parameters. Both of these semi-quantitative parameters were highly dependent on the organ’s compressibility limits under stress and on the skill of the operator to correctly compress the tissue [7]. SWE is a new method of obtaining elastography images, where an acoustic pressure wave induces slow-moving shear waves within the tissue, and the speed of propagation of the shear wave is proportional to the tissue’s elastic stiffness. Shear waves travel ⇑ Corresponding author. Tel.: +86 10 66936848; fax: +86 10 68161218. E-mail address: [email protected] (Z.L. Wang). http://dx.doi.org/10.1016/j.ultras.2014.10.016 0041-624X/Ó 2014 Elsevier B.V. All rights reserved.
relatively faster in stiffer tissue compared to softer tissues. Ultrafast imaging of the propagation of shear waves allows measurement of small changes in velocity that occur when the waves pass through tissues of different stiffness. The velocity information can be mapped to create an image of the stiffness, with the option of measuring SWE features such as the minimum, mean, and maximum elasticity in a region of interest. Deformation of tissue leading to shear waves is created by an acoustic impulse that is generated electronically. With this method, the radiation force produced by the probe is used generate shear waves. This approach is different from conventional quasi-static elastography where the compression is applied externally by the operator. The study by Evans et al. [8] demonstrated that shear wave elastography could give quantitative and reproducible information on solid breast lesions with diagnostic accuracy at least as good as conventional ultrasound with BI-RADS classification. Our previous study demonstrated that when optimal cut-off value of 91.53 kPa was used for max elasticity in the differentiation of breast lesions, the diagnostic sensitivity and specificity were 60.9% and 85.3%, respectively [9]. All these studies showed that shear wave elastography could give quantitative elasticity information that potentially could help in breast lesion characterization.
45
Z.L. Wang et al. / Ultrasonics 57 (2015) 44–49
Although malignant breast tumors exhibited stiff properties compared with benign breast tumors, the reasons for this difference remained to be elucidated. The extracellular matrix (ECM) in primary breast cancers was significantly changed compared with normal breast tissue [10,11]. Collagen fiber is the most abundant structural protein in ECM [12], and increased collagen fiber I had been found to facilitate breast tumor formation, invasion, and metastasis [13,14]. Study has demonstrated that, in cancer progression, there were changes of ECM, such as increasing collagen expression, collagen deposition and structure changes [15]. X-ray mammography, which detected dense fibroglandular tissue in the breast, demonstrated that women with high (50–74% versus <5%) breast density have a 4.64-fold increased risk of developing breast carcinoma [16]. Levental discovered that breast tumorigenesis was accompanied by collagen crosslinking, ECM stiffening, and increased focal adhesions. Induction of collagen crosslinking stiffened the ECM, promoted focal adhesions, enhanced PI3 kinase (PI3K) activity, and induced the invasion of an oncogene-initiated epithelium [17]. Studies also had demonstrated that the proportion of collagen fiber areas was the strongest pathologic determinant of mean stiffness in hepatocellular carcinomas [18]. To our knowledge, there are no investigations comparing elasticity with collagen fiber content in breast lesions. The purpose of this study was to investigate the differences in elasticity and collagen fiber content between malignant and benign breast lesions, and to study the relationship between shear wave elasticity and the content of collagen fiber in ECM. 2. Material and methods 2.1. Patients Between March 2012 to March 2013, a prospective study was conducted at our institute. The study population consisted of consecutive patients referred to our center for ultrasound-guided biopsy of a sonographically apparent breast lesion. Subjects were included following their consent for the index testing. We excluded pregnant and lactating women, those with breast implants, women receiving chemotherapy or radiotherapy for any cancer, skin masses and any which had been biopsied, and patients with a history of ipsilateral breast surgery. Pathologic diagnosis was used as the reference standard. One hundred and six consecutive patients with one hundred and sixteen solid lesions were included in this study. Of these 106 women, 61 women were asymptomatic, 45 women presented
with a palpable mass, and two showed nipple discharge. The age of the patients was 22–82 years (mean age ±standard deviation, 51.8 ± 28.3 years). Of these patients, 96 patients had a single nodule, and 10 patients had two nodules. The maximum diameter of the nodules ranged from 0.7 to 4.6 cm (mean diameter ± standard deviation, 2.1 ± 2.0 cm). Informed consent was obtained from all patients, and the study was approved by our local Ethics Committee. Written informed consent was obtained from every patient at enrollment. 2.2. US examination and SWE examination US examination and SWE examination were performed with AixplorerÒ ultrasound system (SuperSonic Imagine, Aix en Provence, France) with the center frequency of the probe was 12 MHz. The ultrasound images underwent American College of Radiology Breast Imaging Reporting and Data System (ACR BI-RADS) [14] classification by two breast radiologists (Zhi Li Wang and Nan Li) who have twelve years and ten years experience in breast ultrasound, respectively. After the ultrasound examination, the radiologist switched to the elastography model. The probe was applied as light as possible to give no pressure to the lesion. The probe needs to be kept still for 10–20 s during acquisition of the elastography images (due to a slow frame rate) and this was often best done during a breath hold. The elastography views selected were those most clearly displaying abnormal stiffness within the plane but with the absence of movement or pressure artifact, such as the red area under the probe. After the stable image was obtained, the image was recorded and region of interest (ROI) was chosen to calculate the elasticity value. The ROI was chosen as large as possible to cover the whole lesion including calcification potentially present and the edge of the lesion. Because the ROI box was circle, it was hard to cover the whole lesion, especially for the lesions with irregular edge. But we tried our best to cover most part of the lesion, especially for the hardest part of the lesion. Another ROI in peripheral parenchyma was selected to be at a similar depth to that of the breast lesion. For each patient, three ROIs in the lesion and peripheral parenchyma, respectively, were selected and the mean value was regarded as the final value. The ROI in peripheral parenchyma was made as much as possible to be of the same size and depth of the corresponding breast lesion. The maximum value within the ROI, called max elasticity, the mean elasticity, the minimum value within the ROI, called min elasticity and elasticity ratio between lesions and surrounding parenchyma were recorded.
Table 1 Maximum, mean and minimum elasticity and elasticity ratio between lesions and surrounding parenchyma of malignant and benign lesions.
Malignant lesions Benign lesions t test p-value
Max elasticity (kPa)
Mean elasticity (kPa)
Min elasticity (kPa)
Elasticity ratio
140.43 ± 70.16 54.64 ± 48.53 6.329 0.000
63.11 ± 33.68 34.52 ± 25.23 5.382 0.000
25.48 ± 19.85 18.86 ± 15.95 0.567 0.356
3.49 ± 1.95 2.25 ± 1.48 4.487 0.001
Table 2 Maximum, mean and minimum elasticity and elasticity ratio between lesions and surrounding parenchyma among different pathology lesions. Pathological diagnosis
Maximum elasticity (kPa)
Mean elasticity (kPa)
Minimum elasticity (kPa)
Mean elasticity ratio
IDC (n = 51) DCIS (n = 19) ILC (n = 5) Fibroadenoma (n = 19) Fibroadenoses (n = 12) Papillomas (n = 6) Inflammation (n = 4)
142.45 ± 68.37 134.17 ± 61.45 138.89 ± 64.78 53.25 ± 49.63 55.14 ± 50.69 63.62 ± 75.02 44.26 ± 39.47
64.12 ± 33.97 60.62 ± 36.59 61.38 ± 29.43 35.03 ± 29.78 30.84 ± 27.82 40.24 ± 45.47 33.72 ± 33.48
26.31 ± 21.76 25.15 ± 19.11 25.01 ± 20.45 17.58 ± 10.94 23.13 ± 29.31 24.63 ± 21.74 13.58 ± 10.32
3.51 ± 1.89 3.35 ± 1.68 3.26 ± 1.73 2.04 ± 1.22 2.43 ± 1.94 2.57 ± 2.13 2.28 ± 2.14
46
Z.L. Wang et al. / Ultrasonics 57 (2015) 44–49
Fig. 1. Elastography and collagen fiber content of breast carcinoma. (A) A lesion in the upper outer quadrant of the right breast, which was proved to be IDC by pathology; (B) elastography with supersonic shear imaging showed its relatively high stiffness, with maximum elasticity 145.71 kpa; (C) collagen fiber dye (400), the red regions were recognized as collagen fiber areas and collagen fiber was major distributed in mesenchyma; (D) quantitative analysis showed the average area of collagen fiber was 12313.80 ± 4625.25. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Z.L. Wang et al. / Ultrasonics 57 (2015) 44–49
47
Fig. 2. Elastography and collagen fiber content of a benign breast lesion. (A) A lesion in the upper outer quadrant of the left breast, which was proved to be adenosis by pathology; (B) elastography with supersonic shear imaging showed its maximum elasticity was 51.31 kpa; (C) collagen fiber dye (400), the red regions were recognized as collagen fiber areas and collagen fiber was major distributed in basement membrane and mesenchyma; (D) quantitative analysis showed the average area of collagen fiber was 7528.43 ± 4625.25. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
48
Z.L. Wang et al. / Ultrasonics 57 (2015) 44–49
Two independent expert observers performed a blinded review of the static images of all lesions. 2.3. Biopsy procedure 16G core needle biopsy (CNB) are routinely used. All biopsies were performed by two radiologists (Zhi Li Wang and Nan Li) who have ten years and seven years experience in interventional experience, respectively. Two or three core samples were obtained in the area with largest elasticity. The appearance and behavior of the formalin-fixed core samples were examined during the procedure to confirm that the targeted lesion was adequately sampled. 2.4. EVG staining Slides were deparaffinized and hydrated to distilled water; stained in Weigert iron hematoxylin for 10 min, and then rinsed in tap water. After a quick differentiation in 1% hydrochloric acid alcohol, slides were washed in running water for 2 min. After Counterstain in Van Gieson’s solution for 2 min, the slides received quick differentiation by 95% alcohol for a few seconds, then dehydrated in 100% alcohol. After Cleared in xylene for 3 min, slides were coversliped with resinous mounting medium and observed under the microscope. 2.5. Quantification of collagen fiber Image-Pro Plus 5.1 Software was used to quantitatively analyze collagen fiber area. Computer graphic analysis was performed under detailed measurement settings by a board-certified pathologist. 5 fields under high magnification (400) were randomly selected for each slide and the mean value was calculated. 2.6. Statistical analyses All analyses were performed using SPSS11.0, standard version (SPSS Inc, Chicago, IL). All data were described as mean ± SD. The Student’s t-test was used to assess the differences of elasticity and collagen fiber area between benign and malignant lesions. Pearson correlation analysis was used to analyze the correlation between elasticity and collagen fiber area. P < 0.05 was considered statistically significant.
For max elasticity, the optimal cut-off value, yielding the maximal sum of sensitivity and specificity, was 92.32 kPa. With this cut-off value, the sensitivity, specificity, PPV, NPV and accuracy for max elasticity were 84.3%, 92.6%, 78.7%, 77.5%, 84.5%, respectively. 3.3. Comparison of collagen fiber area between malignant and benign lesions The mean areas of collagen fiber of benign and malignant lesions of breast were (7625.14 ± 5321.28) lm2 and (11782.26 ± 8627.54) lm2, respectively. The mean collagen fiber area of malignant lesions was significantly larger than that of the benign lesions (t = 8.437, p = 0.000) (Figs. 1 and 2). The collagen fiber areas of lesions with different pathology types could be seen in Table 3. 3.4. Correlation between elasticity and collagen fiber area Pearson correlation analysis showed a positive correlation between max elasticity of breast lesions and fiber collagen area (r = 0.746) (Fig. 3). 4. Discussion This study demonstrated that max, mean elasticity and elasticity ratio between lesions and surrounding parenchyma of malignant lesions were significantly higher than those in benign lesions, which further demonstrated the important role of elasticity in the differential diagnosis of breast lesions. This study showed that the mean collagen fiber area of malignant breast lesions was significantly larger than that of benign
Table 3 Collagen fiber areas of different pathological types of breast. Pathological diagnosis
Collagen fiber content
IDC (n = 176) DCIS (n = 27) ILC(n = 3) Fibroadenoma (n = 64) Fibroadenosis (n = 32) Papilloma (n = 8) Inflammation (n = 6)
12011.38 ± 81589.34 10623.58 ± 78364.78 10864.75 ± 80585.27 7735.54 ± 5015.56 7536.72 ± 5218.71 8065.56 ± 6188.58 7356.18 ± 5789.47
3. Results 3.1. Pathology Of the 116 masses, 75 (64.7%) were benign and 41 (35.3%) were malignant. Malignant masses included invasive ductal carcinomas (n = 51), ductal carcinoma in situ (DCIS) (n = 19) and invasive lobular carcinoma (ILC) (n = 5). The benign lesions consisted of 19 fibroadenomas, 12 fibroadenoses, 6 papillomas, and 4 inflammations. 3.2. Comparison of elasticity between malignant and benign lesions The max elasticity, mean elasticity, min elasticity and elasticity ratio between lesions and surrounding parenchyma of lesions with different pathology could be seen in Table 1. Malignant lesions exhibited significantly higher max elasticity, mean elasticity, and elasticity ratio between lesions and surrounding parenchyma than benign lesions (t = 5.329, t = 4.382, t = 4.487, P < 0.001). The elasticity parameters of lesions with different pathology types could be seen in Table 2.
Fig. 3. As the Scatter shows, the maximum elasticity and collagen fiber area of breast lesions were positively linear correlated (r = 0.746).
Z.L. Wang et al. / Ultrasonics 57 (2015) 44–49
lesions, which demonstrated that collagen fiber content might play an important role in the progression of breast carcinoma. This study applied Pearson correlation analysis to compare the max elasticity of breast lesion with collagen fiber area, which demonstrated that there was a positive linear correlation between them. It demonstrated that collagen fiber content might have an important impact on tissue stiffness of breast lesions. Our study had several limitations. First, we did not explore associations between stiffness and the other component of ECM, such as elastic fiber, laminin, and fibronectin (FN). However, collagen is the most abundant structural proteins in ECM and its correlation with elasticity could be most important. Second, we did not detect different types of collagen fiber in the breast lesions. It has been demonstrated that type I collagen fiber was the main collagen fiber that prop up the mammary stroma in breast carcinoma [4,5], and further study would be conducted on the impact of different types of collagen fiber on the stiffness of breast lesions. In general, our research demonstrated that the elasticity of breast lesions was closely related to the collagen fiber contents and the collagen fiber content might have an important impact on tissue stiffness of breast lesions. Acknowledgment Financial support from the National Natural Science Foundation (81101059) is gratefully acknowledged. References [1] R.M. Lerner, S.R. Huang, K.J. Parker, ‘‘Sonoelasticity’’ images derived from ultrasound signals in mechanically vibrated tissues, Ultrasound Med. Biol. 16 (3) (1990) 231–239. [2] T.A. Krouskop, T.M. Wheeler, F. Kallel, et al., Elastic moduli of breast and prostate tissues under compression, Ultrason. Imaging 20 (4) (1998) 260–274. [3] A. Modesti, G. D’Orazi, S. Scarpa, et al., Ultrastructural and immunoelectron microscopic study of the desmoplastic stroma in carcinoma of the breast, G. Chir. 10 (5) (1989) 245–249.
49
[4] C. Clavel, P. Birembaut, J.J. Adnet, et al., Breast carcinomas and the extracellular matrix, Ann. Pathol. 8 (2) (1988) 107–113. [5] G.M. Giuseppetti, A. Martegani, B. Di Cioccio, et al., Elastosonography in the diagnosis of the nodular breast lesions: preliminary report, Radiol Med 110 (1– 2) (2005) 69–76. [6] Itoh, E. Ueno, E. Tohno, et al., Breast disease: clinical application of US elastography for diagnosis, Radiology 239 (2006) 341–350. [7] F. Fleury Ede, J.C. Fleury, S. Piato, et al., New elastographic classification of breast lesions during and after compression, Diagn. Interv. Radiol. 15 (2) (2009) 96–103. [8] A. Evans, P. Whelehan, K. Thomson, et al., Quantitative shear wave ultrasound elastography: initial experience in solid breast masses, Breast Cancer Res. 12 (6) (2010) R104. [9] Z.L. Wang, J.L. Li, M. Li, et al., Study of quantitative elastography with supersonic shear imaging in diagnosis of breast tumors, Radiol. Med. 118 (4) (2013) 583–590. [10] A. Bergamaschi, E. Tagliabue, T. Sorlie, et al., Extracellular matrix signature identifies breast cancer subgroups with different clinical outcome, J. Pathol. 214 (3) (2008) 357–367. [11] G. Falzon, S. Pearson, R. Murison, Analysis of collagen fibre shape changes in breast cancer, Phys. Med. Biol. 53 (23) (2008) 6641–6652. [12] L. Kolacna, J. Bakesova, F. Varga, et al., Biochemical and biophysical aspects of collagen nanostructure in the extracellular matrix, Physiol. Res. 56 (Suppl 1) (2007) S51–S60. [13] S. Alowami, S. Troup, S. Al-Haddad, et al., Mammographic density is related to stroma and stromal proteoglycan expression, Breast Cancer Res. 5 (5) (2003) R129–R135. [14] T. Li, L. Sun, N. Miller, et al., The association of measured breast tissue characteristics with mammographic density and other risk factors for breast cancer, Cancer Epidemiol Biomarkers Prev 14 (2) (2005) 343–349. [15] S. Jodele, L. Blavier, J.M. Yoon, et al., Modifying the soil to affect the seed: role of stromal-derived matrix metalloproteinases in cancer progression, Cancer Metastasis Rev. 25 (1) (2006) 35–43. [16] L.J. Martin, N.F. Boyd, Mammographic density. Potential mechanisms of breast cancer risk associated with mammographic density: hypotheses based on epidemiological evidence, Breast Cancer Res. 10 (1) (2008) 201. [17] K.R. Levental, H. Yu, L. Kass, et al., Matrix crosslinking forces tumor progression by enhancing integrin signaling, Cell 139 (5) (2009) 891–906. [18] M. Honjo, F. Moriyasu, K. Sugimoto, et al., Relationship between the liver tissue shear modulus and histopathologic findings analyzed by intraoperative shear wave elastography and digital microscopically assisted morphometry in patients with hepatocellular carcinoma, J. Ultrasound Med. 33 (1) (2014) 61–71.