Comparison of the SR max,SR ave,and color map of strain-elastography in differentiating malignant from benign breast lesions

Abstract

Background

The maximum value of the strain ratio (SR) is a newly developed measure in strain-elastography.

Purpose

To prospectively compare the diagnostic performance of three different measures of strain-elastography, the maximum value of the SR (SR_max), the average value of the SR (SR_ave), and the color map, for differentiating benign and malignant breast lesions.

Material and Methods

We obtained the SR_max and SR_ave of 314 lesions from 290 patients with the tissue to nodule SR and color map using a five-degree scoring system. The diagnostic performances of the SR_max, SR_ave, and color map were compared after obtaining the area under the receiver operating characteristic (ROC) curves (AUCs) of each parameter.

Results

The AUC of the SR_max (0.7674) was larger than the AUCs of the SR_ave (0.7138) and color map (0.6324), with statistical significance (P = 0.0383 for SR_max vs. SR_ave, P = 0.0000 for SR_max vs. color map). The AUC of the SR_ave was larger than that of the color map; however, there was no significant difference. The optimal cut-off point of the SR_max that balanced the sensitivity (91.12%) and specificity (50.81%) was 5.16.

Conclusion

The SR_max is a more reliable diagnostic tool than the SR_ave and color map for differentiating benign and malignant breast lesions.

Keywords

Breast ultrasound strain-elastography strain ratio solid mass

Introduction

Breast elastography has developed rapidly and has been used to help differentiating benign and malignant solid masses based on their tissue stiffness (1,2). Malignant lesions of the breast tend to be stiffer than benign lesions; therefore, elastography imaging promises to be a highly specific method for distinguishing benign and malignant breast lesions and has the potential to reduce the number of benign breast biopsies performed (3 –5). Strain-elastography is displayed as a color overlay on the B-mode ultrasound (US) with red to green to blue used to represent the degree of stiffness. The color map with a 4- or 5-point scoring system compares the strain, i.e. the elastographic color of the lesion, with the scale of the relative tissue displacement using manual compression (1,6). However, using the color map of strain-elastography has several limitations in that the acquired information is operator dependent and non-quantitative (6,7). Garra et al. reported quantitative measurements by comparing the stiffness of the target lesion with the reference tissue within the breast; this value is known as the strain ratio (SR) (2). The SR offers semi-quantitative information about the stiffness of breast lesions, which is obtained by the reference tissue strain divided by the lesion strain. A stiff lesion with a low strain tends to produce a high SR. Therefore, a higher SR is correlated with a greater likelihood of malignancy (4). Previous studies (4,8,9) have suggested the use of the average value of the SR (SR_ave); however, recent technological advances in elastography have permitted the evaluation of the maximum value of the SR (SR_max) as well. To the best of our knowledge, this study is the first to assess and compare the SR_max with the SR_ave.

Thus, the purpose of this study was to compare the diagnostic performance of the SR_max, SR_ave, and color map in the differentiation of benign and malignant breast lesions in the same breasts using histologic analysis as the reference standard.

Material and Methods

Patient selection and data collection

This prospective study was approved by the institutional review board. All of the patients provided written informed consent.

From January 2014 to March 2015, we performed elastography in 304 patients before biopsy or surgery. We excluded nine patients who had undergone chemotherapy or radiotherapy for previous cancer and three patients who had undergone neoadjuvant chemotherapy before US examination. Two cases of lesions that were larger than the probe were excluded. All of these factors can affect the SR of elastography (10). Therefore 290 patients with 314 breast lesions were prospectively enrolled (mean age = 45.5 ± 10.1 years; age range = 17–81 years). Twenty-two patients had two lesions in the unilateral or bilateral breasts. Biopsy or surgery results in the pathology reports were considered as reference standards.

Image acquisition, US, and elastography

Two radiologists who had six and eight years of experience in breast imaging, respectively, examined whole breast US in the transverse and longitudinal planes for all 290 patients. In addition, radial or oblique scanning was also performed of some of the lesions. US examinations were performed using 5- to 12-MHz linear transducers with EPIQ (Philips Healthcare, Bothwell, WA, USA).

The US features of each lesion were evaluated according to the standardized Breast Imaging Recording and Data System (BI-RADS) Classification recommended by the American College of Radiology (ACR) (11). All patients with BI-RADS scores of 4 or 5 for the lesions underwent biopsies, and in some cases scores of 2 or 3 led to biopsies for reasons such as the patient's or clinician's request. Before US-guided biopsy, the radiologist who performed the US examination also performed elastography on those lesions to be scheduled for biopsy.

Elastographic images were obtained as real-time images using EPIQ by a freehand manual compression technique. For standardization of a freehand manual compression technique by two radiologists, we had training period with 13 cases before the beginning the study. As in Itoh et al. (1), we applied the probe with only light pressure to the lesion vertically, and the pectoralis muscle was displayed parallel to the probe oriented perpendicular to the chest wall (2,6,7). The top of the color box on the elastographic images included the subcutaneous fat and the bottom of the regions of interest (ROI) included a pectoralis muscle. The lateral borders of the ROI were set > 5 mm from the lesion's boundary.

For qualitative assessment, each lesion was assigned an elasticity score according to the 5-point scoring system proposed by Itoh et al. (1). A score of 1 indicated even strain for the entire hypoechoic lesion (i.e. the entire lesion was evenly shaded in green). A score of 2 indicated strain in most of the hypoechoic lesion, with some areas exhibiting no strain (i.e. the hypoechoic lesion had a mosaic pattern of green and blue). A score of 3 indicated strain at the periphery of the hypoechoic lesion, while sparing of the center of the lesion (i.e. the peripheral part of the lesion was green and the central part was blue). A score of 4 indicated that there was no strain in entire hypoechoic lesion (i.e. the entire lesion was blue, but its surrounding area was not included). A score of 5 indicated that there was no strain in the entire hypoechoic lesion or in the surrounding area (i.e. both the entire hypoechoic lesion and its surrounding area were blue) (Fig. 1). Scores of 1, 2, and 3 indicated a benign nodule, whereas scores of 4 and 5 indicated a malignant nodule. On elastography, color-coded red indicated soft tissue, whereas color-coded blue indicated hard tissue. The risk of malignancy increases from 1 (benign lesion) to 5 (malignant lesion).

Fig. 1.

A 5-point elasticity score by Ito et al (1). A score of 1 represents an entirely green lesion with the same elasticity throughout the lesion; a score of 2 represents a lesion of which the greater part can be deformed although it may also contain non-deformable areas (green and blue mosaic); a score of 3 corresponds to a high elasticity level in the periphery of the lesion (green) while the center of the lesion is blue; a score of 4 indicates no deformability throughout the lesion (the entire lesion is blue although the adjacent tissue is not affected); and a score of 5 indicates that there is no deformation throughout the hypoechoic lesion or the adjacent tissue (the lesion and adjacent tissue are blue). The risk of malignancy increases from 1 (benign lesion) to 5 (malignant lesion).

In the semi-quantitative evaluation of the elastographic images, the SR_max and SR_ave were considered. The SR is designed to compare the strain in two manually selected ROIs on the elastography. The SR is automatically calculated by the elastography software and yields a fraction of the average strain in the reference area divided by the average strain in the lesion (4). For the automatically calculated elastographic analyses, a short video composed of several frames was stored in a program called QLAB-EQ for quantification. The SR_ave is the average value of several frames in the video. Recently developed US provides the SR_max, which is the maximum value of the SR over, several frames in the video. The technique of SR involves computing a ratio to compare the strain in two selected ROIs, one ROI within the mass and the other in the adjacent breast tissue at the same depth. We selected the fat-to-lesion SR of each lesion, as recommended by Zhou et al (12). In the elastographic image, the radiologist selected two round-shaped ROIs with an average diameter of 2 mm; one ROI was positioned at the hardest area, which presented a blue color, within the mass and the softest area presented a red color for the fat outside the mass, at the same depth. The SR_max and SR_ave were automatically calculated by software. Each lesion was assessed at least twice and the average value was recorded (13).

After the US and elastography examination, the radiologists performed US-guided biopsies using 14-gauge core needle devices (Stericut, TSK, Laboratory, Tochigi, Japan) for all of the lesions. There were no lesions for which the diagnosis changed from benign to malignant on the basis of surgical pathology results. For the patients who underwent surgery, the surgical pathology results were regarded the reference standard rather than the core biopsy pathology results. Thus, lesions upgraded on the basis of surgical pathology results had no influence on our conclusion.

Statistical analysis

We compared the mean of the SR_max and SR_ave between the benign and malignant breast lesions using Student's t-test. Receiver operating curves (ROCs) were obtained and we evaluated the area under the ROC curve (AUC) to compare the diagnostic performance of the SR_max, SR_ave, and color map. From the ROC curves, we calculated the optimal cut-off values of the SR_max and SR_ave for balancing the sensitivity and specificity. In addition, the sensitivities, specificities, positive predictive values (PPV), and negative predictive values (NPV) of the SR_max, SR_ave, and color map were obtained. These analyses used STATA software (version 10.0; Stata Corp, College Station, TX, USA). For all of the statistical analyses a P value < 0.05 was considered statistically significant.

Results

Of 314 lesions, 246 (78.34%) were benign and 68 (21.66%) were malignant (Table 1). The average diameter ( ± SD) of the benign lesions was 19.7 ± 9.8 mm (range = 50–54 mm; median = 18.0 mm) and the average diameter of the malignant lesions was 22.2 ±16.1 mm (range = 55–55 mm; median = 16.9 mm). The subtypes of the benign and malignant lesions are listed in Table 1. Two cases of ductal carcinoma in situ were upgraded to invasive ductal carcinoma on the basis of the surgical pathology results. On conventional US, five lesions (1.9%) were categorized as BI-RADS category 2 (typical benign lesion), 71 (22.6%) as category 3 (probably benign lesion), 164 (52.2%) as category 4A (low suspicion for malignancy), 25 (8.0%) as category 4b (intermediate suspicion for malignancy), 17 (5.4%) as category 4C (moderate concern for malignancy), and 32 (10.2%) as category 5 (highly suggestive of malignancy) (Table 2).

Table 1.

Final pathologic diagnosis in 314 breast lesions.

Pathologic diagnosis	Lesions (n)
Malignant lesions	68
Invasive ductal carcinoma	50
Ductal carcinoma in situ	10
Invasive lobular carcinoma	2
Mucinous carcinoma	2
Other malignant lesions: medullary carcinoma, metaplastic carcinoma, invasive tubular carcinoma, invasive papillary carcinoma	1, each
Benign lesions	246
Fibroadenoma	116
Fibrocystic change	27
Stromal sclerosis	25
Intraductal papilloma	20
Adenosis	13
Fibroadenomatous mastopathy	9
Stromal fibrosis	9
Ductal ectasia	7
Florid ductal hyperplasia	6
Atypical ductal and lobular hyperplasia	4
Benign pyllodes tumor	3
Columnar cell change	2
Other benign lesions: lobular carcinoma in situ, chronic granulomatous mastopathy, radial sclerosing lesion, fat necrosis, tubular adenoma	1, each

Table 2.

BI-RADS categorization on conventional US.

BI-RADS category	Lesions (n (%))
2, typical benign lesion	5 (1.9)
3, probably benign lesion	71 (22.6)
4A, low suspicious for malignancy	164 (52.2)
4B, intermediate suspicion for malignancy	25 (8.0)
4C, moderate concern for malignancy	17 (5.4)
5, highly suggestive of malignancy	32 (10.2)

On the color map, three (1.0%) lesions were scored as 1, 25 lesions (8.0%) as 2, 51 (16.2%) as 3, 180 (57.3%) as 4, and 55 (17.8%) as 5 (Table 3).

Table 3.

Elasticity score by color map.

Elasticity score	Lesions (n (%))
1	3 (1.0)
2	25 (8.0)
3	51 (16.2)
4	180 (57.3)
5	55 (17.8)

The mean SR_ave ( ± SD) was 3.09 ± 0.66 (range =0.76–7.50; median = 2.87) in the benign group and 4.15 ± 0.59 (range = 1.48–7.80; median = 4.16) in the malignant group. The mean SR_max was 8.21 ± 1.23 (range = 0.34–16.07; median = 8.04) in the benign group and 9.53 ± 2.62 (range = 1.99–20.01; median = 8.61) in the malignant group (Table 4). Both the mean and maximum SR distributions were higher in the malignant lesions than in the benign lesions and the difference was statistical significant (P < 0.05) (Fig. 2).

Table 4.

Mean of SR_max and SR_ave.

	SR_max	SR_ave
Benign lesions	8.21 ± 1.23	3.09 ± 0.66
Malignant lesions	9.53 ± 2.62	4.15 ± 0.59

Fig. 2.

Box plot distribution of the SR_ave and SR_max in benign and malignant breast lesions.

The AUC was 0.6324 (95% confidence interval [CI] = 0.5945–0.6704) for the color map, 0.7138 (95% CI = 0.6456–0.7820) for the SR_ave, and 0.7674 (95% CI = 0.7046–0.8302) for the SR_max (Fig. 3). The AUC of the SR_max was larger than the AUCs of the SR_ave and color map, and the differences were statistically significant (P = 0.0383 for SR_max vs. SR_ave, P = 0.0000 for SR_max vs. color map). The AUC of the SR_ave was larger than the AUC of the color map; however, the differences were not statistically significant (P = 0.194 for SR_ave vs. color map) (Table 5). These results showed that the SR_max is a more useful measure than the SR_ave and color map for the differentiation of the benign and malignant disease.

Fig. 3.

The ROC of the SR_max, SR_ave, and color map. The AUC is 0.7674 for the SR_max (solid dot line), 0.7138 for the SR_ave (empty dot line), and 0.6324 for the color map (dashed line).

Table 5.

ROC comparison for SR_max, SR_ave, and color map.

ROC compare	P value
SR_max vs. SR_ave	0.0383
SR_max vs. color map	0.0000
SR_ave vs. color map	0.194

SR, strain ratio.

The sensitivity, specificity, PPV, and NPV were 76.2%, 73.2%, 40.5%, and 88.7%, respectively, for the SR_max, 73.6%, 70.5%, 38.3%, and 88.1%, respectively, for the SR_ave, and 85.6%, 51.5%, 37.4%, and 86.3%, respectively, for the color map (Table 6).

Table 6.

Statistical analysis of SR_max, SR_ave, and color map.

Value	Performance measure
Value	Sensitivity (%)	Specificity (%)	PPV (%)	NPV (%)
SR_max	76.2	73.2	40.5	88.7
SR_ave	73.6	70.5	38.3	88.1
Color map	85.6	51.5	37.4	86.3

PPV, positive predictive value; NPV, negative predictive value.

The optimal cut-off value of the SR_max was 5.16, which created a balance with the 88.1% sensitivity and 70.2% specificity; that of SR_ave was 3.65, which balanced with the 68.7% sensitivity and 70.7% specificity (Fig. 4).

Fig. 4.

(a–c) A 52-year-old woman with invasive ductal carcinoma. The US showed a microlobulated hypoechoic mass with echogenic rim (BI-RADS category 5). On elastography, the color map score was 4, the SR_max was 8.55, and the SR_ave was 5.51.

Discussion

Our study first evaluated the diagnostic performance of the SR_max and demonstrated that the SR_max was more useful than the SR_ave and color map for differentiating benign and malignant breast lesions.

Previous studies have discussed the SR in elastography, which uses the average value of the SR_ave. Recently developed US provide the SR_max; in the elastographic image, these two values for the targeted lesion are displayed at the same time. When radiologists perform elastography and measure the SR, sometimes it is difficult for them to differentiate benign and malignant breast lesions. Therefore, we investigated the difference between these two values in the diagnostic performance and cut-off values compared with the color map.

In the current study, both the SR_max and SR_ave were higher in malignant lesions than in benign lesions; this difference was significant (P < 0.0000). Previous studies (4,9) that discussed the SR_ave revealed it was higher in malignant lesions than in benign lesions and, thus, the calculation of the SR_ave allows for the differentiation of significant differences of between benign and malignant breast lesions and also contributes to the standardization of strain-elastography. In accordance with these results, the SR_max can be similarly used to differentiate benign from malignant breast lesions.

Several investigators have reported the usefulness of elastography based on the color map and SR_ave, either alone or together. Thomas et al. (4) reported a sensitivity of 81% and a specificity of 89% for elastography, and a sensitivity of 90% and a specificity of 89% for the SR_ave. These authors reported that the SR_ave contributes to improving the sensitivity; however, the specificity of the SR_ave was not higher than that of the color map. Zhi et al. (9) found that the SR_ave can provide a new, more reliable diagnostic tool in comparison with the five-point scoring system of the color map. In their study, the SR_ave showed better diagnostic performance than those of the color map or conventional US. However, there were significant statistical differences between the two studies.

In the present study, the AUC of the SR_ave was larger than that of the color map; however, this difference was not statistically significant. The AUC of the SR_max was larger than those of the SR_ave and color map, and these differences were statistically significant. This result revealed that the SR_max is a more reliable value than the SR_ave and color map for differentiating benign and malignant breast lesions. Therefore, we recommend the SR_max with 5.16 as the cut-off value for differentiating benign and malignant breast lesions.

The current study revealed that the cut-off value of the SR_ave was 3.65 and the sensitivity and specificity were 68.7% and 70.7%, respectively. Several investigators suggested variable degrees of cut-off values of the SR_ave for the differentiation of benign and malignant lesions. Thomas et al. suggested 2.45 as the cut-off value of the SR_ave and the sensitivity and specificity were 90% and 89%, respectively. Todd et al. showed a sensitivity of 79% and a specificity of 76%, with a cut-off value of 4.5; Zhi et al. showed a sensitivity of 92.4% and a specificity of 91.1%, with a cut-off value of 3.05 (4,5,8). The reasons for the differences in the cut-off values were that there were differences in the reference tissue component for the ROI and in the position and size of the ROI. For the calculation of the SR, the reference tissue placement of the ROI was in fat tissue in some of the studies (4,5,14) and in glandular tissue in other studies (8,9). In the present study, fat tissue was selected over glandular tissue as the reference, because glandular tissue shows various moduli depending on the compression level, whereas fat tissue shows a constant modulus over various compression loadings (15). Additionally, several studies show better diagnostic performance in the fat-to-lesion SR than in the gland-to-lesion SR (12,14,15). With regard to the position of the ROI, some researchers place the reference ROI at the superficial tissue of the targeted lesion (4,5). However, Zui et al. (8,9) and Cho et al. (14) used the same depth level as that of the targeted lesion as the ROI position that was used for the reference. In our study, the ROI for the reference was placed at the same depth. The superficial layer under the transducer compressed more than the deep layer, and thus the strain value of an object in the superficial layer would be higher than that in the deep layer (8). The ROI sizes were variable among the previous studies; however, most of the investigators drew the first ROI using a rectangular or oval figure that was manually constructed and placed in such a way that it was bounded by the inner margin of the target lesion (9,16,17). The size of the second ROI for the fat strain was the same or similar to that of the first ROI. However, in lesions that were >3 cm in size, assessment of the lesion with respect to the adjacent fat tissue was challenging because the fat tissue might not be within the field of view. Furthermore, in some studies, investigators used different sized ROIs for the targeted lesion and reference tissue (4,18). The amount of included tissue is obviously different between the ROI for the lesion and that for the reference tissue. For this reason, and to maximize the comparison, we strictly selected a round-shaped ROI with an approximate average width of 2 mm, and we placed the first ROI at the hardest area that presented a blue color within the targeted lesion and the second ROI in the fat tissue that presented red color just outside of the targeted lesion at the same level (19,20). However, the sensitivity and specificity of our study were lower than in previous studies. This difference might be because the location and reference tissue ROIs were different in size. There is a need for a consistent standard when performing an assessment of the ROI of the SR.

Despite the differences, many research studies, including the current study, have in common that there was a significant improvement in the specificity but not in the sensitivity when using the SR_ave because conventional breast US was already highly sensitive (2,6,21,22).

This study has some limitations. First, the inter-observer variability among US practitioners was not considered. Two radiologists were involved and they had six and eight years of experience in breast imaging, respectively, and both two radiologists had three years of experience in elastography. However, some differences between these radiologists may exist. In performing elastography, it is inevitable that different operators will exert a different degree of manual compression on the breast mass. Second, we only compared the diagnostic performance of the SR_max, SR_ave, and color map. A comparison of the SR_max, SR_ave, and B-mode US or combined SR and B-mode US would be necessary to assess the usefulness of the SR in clinical practice. Third, we used only strain-elastography for this study. However, shear wave elastography or Acoustic radiation force impulse (ARFI) techniques using velocity and Young's modulus are more quantitative method than strain-elastography. Thus, future studies to compare both methods for quantitative assessment of elastography may be useful. Lastly, the sizes of lesions were not measured on elastographic images. Some previous studies have shown a difference in sizes of the lesions on US and elastographic images (23,24). It might affect the result of current study.

In conclusion, the SR_max is a more reliable diagnostic tool in comparison to the SR_ave and color map of elastography in differentiating benign and malignant breast lesions. When balancing diagnostic accuracy, we recommend 5.16 of the SR_max as the optimal cut-off value.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported by Hallym University Research Fund 2017 (HURF-2017-14).

References

Itoh

Ueno

Tohno

et al.

Breast disease: clinical application of US elastography for diagnosis. Radiology 2006; 239: 341–350.

Garra

Cespedes

Ophir

et al.

Elastography of breast lesions: initial clinical results. Radiology 1997; 202: 79–86.

Raza

Odulate

Ong

et al.

Using real-time tissue elastography for breast lesion evaluation our initial experience. J Ultrasound Med 2010; 29: 551–563.

Thomas

Degenhardt

Farrokh

et al.

Significant differentiation of focal breast lesions: calculation of strain ratio in breast sonoelastography. Acad Radiol 2010; 17: 558–563.

Kumm

Szabunio

. Elastography for the characterization of breast lesions: Initial clinical experience. Cancer Control 2010; 17: 156–161.

Burnside

Hall

Sommer

et al.

Differentiating benign from malignant solid breast masses with US strain imaging. Radiology 2007; 245: 401–410.

Chang

Moon

Cho

et al.

Breast mass evaluation: Factors influencing the quality of US elastography. Radiology 2011; 259: 59–64.

Zhi

Xiao

Yang

et al.

Semi-quantitating stiffness of breast solid lesions in ultrasonic elastography. Acad Radiol 2008; 15: 1347–1353.

Zhi

Xiao

Yang

et al.

Ultrasonic elastography in breast cancer diagnosis: strain ratio vs 5-point scale. Acad Radiol 2010; 17: 1227–1233.

10.

Chang

Moon

Cho

et al.

Clinical application of shear wave elastography (SWE) in the diagnosis of benign and malignant breast diseases. Breast Cancer Res Treat 2011; 129: 89–97.

11.

American College of Radiology. ACR BI-RADS® Atlas, Breast Imaging Reporting and Data System, Reston, VA: ACR, 2013.

12.

Zhou

Zhan

et al.

Elastography ultrasound for breast lesions: fat-to-lesion strain ratio vs gland-to-lesion strain ratio. Eur Radiol 2014; 24: 3171–3177.

13.

Alhabshi

Rahmat

Abdul Halim

et al.

Semi-quantitative and qualitative assessment of breast ultrasound elastography in differentiating between malignant and benign lesions. Ultrasound Med Biol 2013; 39: 568–578.

14.

Cho

Moon

Kim

et al.

Sonoelastographic strain index for differentiation of benign and malignant nonpalpable breast masses. J Ultrasound Med 2010; 29: 1–7.

15.

Krouskop

Wheeler

Kallel

et al.

Elastic moduli of breast and prostate tissues under compression. Ultrason Imaging 1998; 20: 260–274.

16.

Jiang

Zhang

Chen

et al.

Technical evaluation of Virtual Touch™ tissue quantification and elastography in benign and malignant breast tumors. Exp Ther Med 2014; 8: 1059–1064.

17.

Carlsen

Ewertsen

Săftoiu

et al.

Accuracy of visual scoring and semi-quantification of ultrasound strain elastography–a phantom study. PLoS One 2014; 9: e88699.

18.

Gheonea

Stoica

Bondari

. Differential diagnosis of breast lesions using ultrasound elastography. Indian J Radiol Imaging 2011; 21: 301–305.

19.

Ricci

Maggini

Mancuso

et al.

Clinical application of breast elastography: state of the art. Eur J Radiol 2014; 83: 429–437.

20.

Graziano

Bitencourt

AGV

Cohen

et al.

Elastographic evaluation of indeterminate breast masses on ultrasound. Rev Bras Ginecol Obstet 2017; 39: 72–79.

21.

Leong

Sim

Lee

et al.

A prospective study to compare the diagnostic performance of breast elastography versus conventional breast ultrasound. Clin Radiol 2010; 65: 887–894.

22.

Regner

Hesley

Hangiandreou

et al.

Breast lesions: Evaluation with US strain imaging–clinical experience of multiple observers. Radiology 2006; 238: 425–437.

23.

Meier-Meitinger

Häberle

Fasching

et al.

Assessment of breast cancer tumour size using six different methods. Eur Radiol 2011; 21: 1180–1187.

24.

Zippel

Shalmon

Rundstein

et al.

Freehand elastography for determination of breast cancer size. J Ultrasound Med 2014; 33: 1441–1446.