A novel two-dimensional quantitative shear wave elastography to make differential diagnosis of breast lesions: Comprehensive evaluation and influencing factors

Abstract

BACKGROUND: Virtual touch imaging quantification (VTIQ), a form of shear wave elastography, may help in the diagnosis of breast lesions.

OBJECTIVE: We aimed to evaluate the diagnostic performance of combined VTIQ and conventional ultrasound (US), and assess the factors influencing VTIQ measurement.

METHODS: From September 2014 to December 2014, 162 patients with breast lesions were examined by US and VTIQ to assess shear wave speed (SWS) and morphological characteristics (lesion shape, orientation, margin and echo pattern). The sensitivity, specificity and accuracy of VTIQ, US and VTIQ+US for the diagnosis of breast lesions was evaluated in comparison to pathological results. Factors influencing deviations in SWS measurements were assessed by logistic regression.

RESULTS: The SWS cut-off between malignant and benign lesions was 3.73 m/s. The sensitivity, specificity and accuracy were: 98.07%, 55.96%, and 69.57% for US; 76.92%, 78.89% and 78.26% for VTIQ; and 98.07%, 84.40% and 88.82% for US+VTIQ. The two factors that influenced the SWS results were the lesion margin (odds ratio [OR], 16.363; 95% confidence interval [CI], 3.220–29.020) and vascularity (OR, 6.712; 95% CI, 1.358–9.072).

CONCLUSIONS: The lesion margin and vascularity could affect the measurement of SWS as well as the experience of examiner. However, VTIQ is still a reliable method that provides valuable information in the differential diagnosis of breast lesions, and may reduce unnecessary biopsies.

Keywords

Elastography breast cancer VTIQ ultrasound

1 Introduction

In recent years, the incidence and prevalence of breast cancer has increased rapidly, with as many as approximately 1 million new patients in each year worldwide, seriously affecting the health and quality of life of women [6]. Early detection of breast cancer is essential to increase life expectancy [20]. Known as the “third eye” of clinicians, ultrasonography (US) possesses major advantages, such as easy operation, no radiation and low cost, and has become the most common method in clinical practice [21]. In 2003, the American College of Radiology (ACR) proposed the breast imaging reporting and data system (BI-RADS), which offers specific criteria for the US diagnosis of breast lesions. The new version of BI-RADS was proposed in 2013 and then worldwide used in breast imaging, which possessed high level of standardization that only a few other domains in imaging field could reach.

However, conventional US diagnosis, mostly based on the morphological features of the lesion, is still unsatisfactory in terms of its specificity and accuracy, because there is always an overlap between the characteristics of benign and malignant lesions. Numerous novel technical approaches, such as ultrasonic elastography, have been developed to solve these problems during the past two decades [17], and have been widely used in the diagnosis of lesions in the breast, liver, thyroid and other organs [12 , 29]. It is known that the stiffness, an inherent feature, has a close relationship with the different pathological types of breast lesion [24]. Based on this principle, elastography, which indirectly reflects the varying stiffness levels in lesions, is considered to be helpful in the diagnosis of benign and malignant lesions.

Virtual Touch Imaging Quantification (VTIQ), a novel two-dimensional (2D) ultrasonic elastography system, has been developed and its diagnostic accuracy for breast lesions has so far been confirmed in preliminary studies [8 , 23]. The measurement of VTIQ is relatively more reliable, because it provides not only quantitative shear wave speed (SWS), but also displays a qualitative map for SWS distribution and a 2D quality mode that can be used to evaluate the accuracy of SWS [14]. However, in clinical practice there are still a proportion of cases in which the measured SWS deviates from the real pathological results. Provided VTIQ, a new generation of elastography, is reliable enough and the investigators have excellent experience with negligible operation error, there are still likely to be certain influencing factors that affect this deviation in the measurement of the SWS. We aimed to determine the main influencing factors and comprehensively analyze the reasons behind such measureddeviations.

2 Materials and method

2.1 Study design and patients

A retrospective study was performed among patients with a breast lesion who had undergone examination by conventional US and VTIQ in our hospital from September 2014 to December 2014. The pathological results of these lesions were confirmed by surgery or biopsy of the lesion. Patients were excluded from this study for the following reasons: (1) simple cystic lesion; (2) previous history of breast surgery or biopsy; (3) unwilling or unsuitable to undergo surgery or biopsy. After exclusion, a total of 161 lesions from 160 patients were enrolled in this study (Fig. 1).

The investigators who performed the examinations had more than 2 years’ experience in breast US. They were neither employees nor consultants for any company or organization that could represent a conflict of interest with the instruments involved in this study.

2.2 Equipment

The US and VTIQ were performed using SIEMENS S3000 ultrasonic diagnostic apparatus equipped with touch tissue imaging quantification (SIEMENS Medical Solutions, Mountain View, CA, USA). Conventional US was first performed, and the obtained morphological characteristics including lesion shape, orientation, margin, echo pattern, posterior acoustic features, calcification and vascularity were recorded.

During the VTIQ examination, the grayscale image was adjusted to ensure the lesion was within the appropriate grayscale range. The region of interest (ROI) was placed in the lesion region, while avoiding obvious calcification. The US probe produces a low-frequency longitudinal impulse, which further generates a transverse shear wave through the tissue. After the shear wave has been encoded into a color image, the apparatus automatically measures and records the SWS in the ROI. The SWS of each breast lesion was measured six to eight times (shown in Fig. 2), and the mean SWS value was calculated.

In addition, the BI-RADS category of the conventional US examination was re-evaluated after combination with the VTIQ results according to a revised rule as follows: (1) when the mean SWS value is greater than the cut-off value, the BI-RADS category is increased by one i.e. category 3 increases to 4A, and category 4A increases to 4B; (2) when the mean SWS value is lower than the cut-off value, the BI-RADS category is decreased by one i.e. category 4B decreases to 4A, and category 4A decreases to 3; (3) as most of the category 5 lesions, which possess obvious malignant sonographic features, can be diagnosed with high accuracy, no adjustment was made to this category. The pathological results from surgery or biopsy were considered to be the gold standard for the evaluation of the sensitivity, specificity and accuracy of US, VTIQ and US+VTIQ using receiver operating characteristic (ROC) curves.

2.3 Statistical analysis

The SPSS16.0 software package was employed to select risk factors from within the data as a whole. Continuous variables were compared by the t test, qualitative data were compared by the X² test or Fisher’s exact test. MedCalc 11 was employed to calculate the cut-off value and perform ROC curve analyses of the different diagnostic methods. The areas under curve (AUCs) were compared using the z test. Differences were statistically significant when p was <0.05. Binary logistic regression was performed from the risk factors selected, with the selection method of the independent variable being “Enter”. The Wald test was used to test the regression parameter estimation value, and p < 0.05 was considered as statistical significance.

3 Result

The ages of the patients ranged from 20 to 84 years. The mean±standard deviation (SD) and median age of patients with malignant breast lesions (58.8±11.2 years and 57 years, respectively) were significantly older than those of patients with benign lesions (40.8±11.1 years and 40 years, respectively; p < 0.01). The mean SWS of the malignant lesions (3.75±0.7 m/s) was significantly higher than that of the benign lesions (2.3±1.1 m/s; p < 0.01). There were also statistically significant differences (p < 0.01) between benign and malignant lesions in terms of their main characteristics, including shape, orientation, margin, echo pattern, posterior acoustic features, calcification and vascularity. Variables from the above characteristics were assigned to perform binary logistic regression, as shown in Table 1.

The best cut-off of SWS for the diagnosis of malignant breast lesions was 3.73 m/s. The sensitivity, specificity and accuracy were calculated to be 98.07, 55.96, and 69.57, respectively, for US, 76.92, 78.89, and 78.26, respectively, for VTIQ, and 98.07, 84.40, and 88.82, respectively, for US+VTIQ, separately as shown in Table 2. The corresponding AUCs were calculated to be 0.784 (95% confidence interval [CI] 0.713–0.845), 0.826 (95% CI 0.759–0.881) and 0.927 (95% CI 0.875–0.962), respectively, indicating that US+VTIQ is more valuable in differentiating benign and malignant lesions. Thirty-six lesions were <10 mm in diameter, 84 lesions were ≥10 to <20 mm and 39 lesions were ≥20 mm, and the diagnostic accuracy were 77.78% (28/36), 89.29% (75/84) and 82.05% (32/39) in each group.

The pathological results of 161 breast lesions were confirmed after biopsy (19.88%, 30/161) or surgical removal (80.12%, 131/161). Among these, 109 lesions were diagnosed as benign, including fibroadenoma with adenosis (n = 45), fibroadenoma (n = 28), adenosis (n = 26), intraductal papilloma (n = 7), epidermoid cyst (n = 1), phyllodes tumor (n = 1) and inflammation (n = 1). The remaining 52 lesions were diagnosed as malignant, including invasive ductal carcinoma (n = 35), invasive lobular carcinoma (n = 9), invasive papillary carcinoma lesions (n = 6), invasive cribriform carcinoma (n = 1) and malignant spindle cell tumor (n = 1).

The false-negative and false-positive results for the US, VTIQ and US+VTIQ methods are showed in Table 3. False-negative results were seen more often with VTIQ than with conventional US, with the corresponding histological types mainly consisting of invasive ductal carcinoma and ductal carcinoma in situ. Meanwhile, false-positive results were more likely to occur using conventional US, with the corresponding histological types mainly consisting of fibroadenoma, adenosis, and fibroadenoma with adenosis as shown in Fig. 3 and Fig. 4.

Furthermore, the BI-RADS category was increased when US+VTIQ compared with conventional US alone in 50 patients: from 3 to 4A in 11 patients, from 4A to 4B in 16 patients, from 4B to 4C in 12 patients, and from 4C to 5 in 22 patients. In contrast, a decrease in the BI-RADS category was seen in 29.81% (48/161) of the lesions: from 4A to 3 in 31 patients, from 4B to 4A in 9 patients, and from 4C to 4B in 8 patients, which could potentially lead to 19.25% (31/161) decrease in the biopsy rate.

The results of logistic regression model analysis (shown in Table 4) meant that only two independent risk factors, lesion margin and vascularity, were selected for inclusion in the regression model, after other characteristics with negligible influence had been excluded. The above model was confirmed to be statistically significant (P < 0.001) by the likelihood ratio test. Further analysis demonstrated that SWS measurement deviations are more likely to occur in lesions with noncircumscribed margins (odds ratio [OR], 16.363; 95% CI, 3.220–29.020), and vascularity within or adjacent to the lesion (OR, 6.712; 95% CI, 1.358–9.072).

4 Discussion

Breast cancer has become a major threat to women’s health as the second most important cause of death in women. Early diagnosis of breast cancer by application of different imaging technologies, such as US, mammography and magnetic resonance imaging (MRI), can greatly improve life expectancy. Elastography, a new sonographic technique, has the unique advantage of being able to acquire information on the tissue stiffness of the tumor [15]. Similar to a clinical palpation examination, elastography is based mainly on the principle that different tissues have different levels of stiffness, and this may help to give useful information on the diagnosis of different pathological types [11].

Previous clinical trials have suggested that elastography may remarkably improve performance for the diagnosis of breast lesions [1 , 27]. In particular, this technique is able to improve the diagnostic specificity without losing sensitivity [2]. Given the importance of elastography, the characteristics of different breast lesions using the technique have now been included in the new edition of the BI-RADS lexicon as associated findings [3].

The main types of elastography include strain elastography (SE) and shear-wave elastography (SWE). The former displays quantitative results for relative tissue displacement under compression, while the latter offers a SWS value using acoustic radiation force excitation [13]. Currently, the latest generation of elastography is VTIQ, which has been developed on the basis of acoustic radiation force impulse. VTIQ offers a value for SWS (m/s) of each pixel in the ROI and codes all the values into a visualized 2D color image [10]. Furthermore, the independence and repeatability of VTIQ has also been proven in previous studies [23, 25]. Taken together, VTIQ is more advanced and efficient compared with other types of elastography.

According to the probability of malignancy, lesions in category 4 have been further divided into 4A (2% – 10%), 4B (10% – 50%) and 4C (50% – 95%) in the new edition of BI-RADS. Lesions in category 4A and above are usually recommended for breast biopsy or surgery in clinical practice. In this study, we noticed that investigators tended to diagnose a certain proportion of category 3 lesions that presented suspicious malignant features, such as uncircumscribed margin or non-parallel orientation, as category 4A in order to avoid missing malignant lesions. The statistical results of our study demonstrate that the false-positive rate of US is as high as 22.36% (36/161). Such lesions are mostly BI-RADS 4A with some indeterminate sonographically features and such over-diagnosis generally results in unnecessary biopsies or surgery. However, a significant decrease in the false-positive rate to 9.94% (16/161) could be achieved after combining US with VTIQ, which would lead to 19.25% (31/161) decrease in the biopsy rate, mainly due to benign quantitative SWS findings. Thus it is reassuring that for lesions in category 4A, due importance should be given to the SWS value, without over-reliance on conventional US features. Our result is consistent with previous reports [9, 10] that have indicated that US+VTIQ can improve diagnostic accuracy and thereby reduce psychological and physical pain for patients.

There are still however a considerable proportion of lesions which were misclassified by VTIQ. The histology of the malignant lesions with measured SWS values lower than the cut-off was mainly invasive ductal carcinoma and ductal carcinoma in situ. This finding is consistent with the previous study of Vinnicombe et al. [26]. Meanwhile, the histology of the benign lesions with SWS values higher than the cut-off were fibroadenoma, adenosis, and fibroadenoma with adenosis.

Besides the conception that VTIQ depends from the experience of the examiner, neither the influencing factors leading to the SWS deviation or the corresponding underlying mechanisms have been fully defined by previous studies. In the present study, we found that the margin and vascularity of lesions were the two risk factors for deviation of the SWS measurement. Moreover, the reason for SWS deviation is likely interference in measurement progress. As is known, the intensity of the ultrasound radiation force generated by VTIQ was limited in order to avoid both mechanical and thermal effects [5, 18], therefore shear waves show high attenuation and small motion amplitude, generally below 20 μm [19]. It is conferred that these drawbacks of shear waves could be further aggravated when facing the uncertain and complex spatial conditions of blood flow and uncircumscribed margins to a lesion, thereby leading to deviation of the SWS measurement; however, the underlying mechanism needs further study. In general, the situation of SWS value to misclassify lesions with blood flow or uncircumscribed margins should be taken into account in the correct classification of lesions.

The present study has some limitations including its retrospective design and small sample size. In particular, some sub-classifications of the sonographic features listed in the BI-RADS US lexicon, such as echo pattern, have not been fully discussed in this paper. More prospective studies with larger sample sizes to investigate the underlying mechanism of SWS measurement deviation will be carried out by us in the near future.

5 Conclusion

VTIQ is a reliable method to provide valuable information on the differential diagnosis of breast lesions, which could reduce unnecessary biopsies. VTIQ can effectively modify the classification of BI-RADS and can improve the accuracy of diagnosis. The risk factors that may affect SWS values are the margin and vascularity of lesions, although the underlying mechanism of this needs further study.

Footnotes

Acknowledgments

This work was supported by the National Natural Science Foundation of China (Grant No. 81471673, 81601499).

References

Barr

R.G.

and Zhang

, Shear-wave elastography of the breast: Value of a quality measure and comparison with strain elastography, Radiology 275 (2015), 45–53.

Berg

W.A.

, Cosgrove

D.O.

, Dore

C.J.

, Schafer

F.K.

, Svensson

W.E.

, Hooley

R.J.

, et al., Shear-wave elastography improves the specificity of breast US: The BE1 multinational study of 939 masses, Radiology 262 (2012), 435–449.

Breast

J.H.

, US: What’s new in BI-RADS 2012?, In: The 44th Annual Congress of Korean Society of Ultrasound in Medicine; Seoul, Korea Seoul: Korean Society of Ultrasound in Medicine, 2013, p. 138.

Chang

J.M.

, Park

I.A.

, Lee

S.H.

, Kim

W.H.

, Bae

M.S.

, Koo

H.R.

, et al., Stiffness of tumours measured by shear-wave elastography correlated with subtypes of breast cancer, Eur Radiol 23 (2013), 2450–2458.

Chen

, Urban

M.W.

, Pislaru

, Kinnick

, Zheng

, Yao

, et al., Shearwave dispersion ultrasound vibrometry (SDUV) for measuring tissue elasticity and viscosity, IEEE T Ultrason Fer 56 (2009), 55–62.

Chen

, Zheng

, Baade

P.D.

, Zhang

, Zeng

, Bray

, et al., Cancer statistics in China, 2015, CA-CANCER J CLIN 66 (2016), 115–132.

Evans

, Whelehan

, Thomson

, Brauer

, Jordan

, Purdie

, et al., Differentiating benign from malignant solid breast masses: Value of shear wave elastography according to lesion stiffness combined with greyscale ultrasound according to BI-RADS classification, Brit J Cancer 107 (2012), 224–229.

Golatta

, Schweitzer-Martin

, Harcos

, Schott

, Gomez

, Stieber

, et al., Evaluation of virtual touch tissue imaging quantification, a new shear wave velocity imaging method, for breast lesion assessment by ultrasound, Biomed Res Int 2014 (2014), 960262.

Gong

, Wang

and Xu

, Application of real-time ultrasound elastography for differential diagnosis of breast tumors, J Ultras Med 32 (2013), 2171–2176.

10.

Ianculescu

, Ciolovan

L.M.

, Dunant

, Vielh

, Mazouni

, Delaloge

, et al., Added value of Virtual Touch IQ shear wave elastography in the ultrasound assessment of breast lesions, Eur J Radiol 83 (2014), 773–777.

11.

Krouskop

T.A.

, Wheeler

T.M.

, Kallel

, Garra

B.S.

and Hall

, Elastic moduli of breast and prostate tissues under compression, Ultrasonic Imaging 20 (1998), 260–274.

12.

Kuehlmann

, Prantl

and Michael Jung

, Imaging of idle breast implants with ultrasound-strain elastography- A first experimental study to establish criteria for accurate imaging of idle implants via ultrasound-strain elastography, Clin Hemorheol Microcirc 61(4) (2016), 645–656.

13.

Lee

S.H.

, Chang

J.M.

, Cho

, Koo

H.R.

, Yi

, Kim

S.J.

, et al., Practice guideline for the performance of breast ultrasound elastography, Ultrasonography 33 (2014), 3–10.

14.

D.D.

, Xu

H.X.

, Guo

L.H.

, Bo

X.W.

, Li

X.L.

, Wu

, et al., Combination of two-dimensional shear wave elastography with ultrasound breast imaging reporting and data system in the diagnosis of breast lesions: A new method to increase the diagnostic performance, Eur Radiol 26 (2015), 3290–3300.

15.

Liu

, Zheng

, Shan

, Lu

, Lin

, Tian

, et al., Elastography by acoustic radiation force impulse technology for differentiation of benign and malignant breast lesions: A meta-analysis, J Med Ultrason (2001) 43 (2016), 47–55.

16.

Liu

, Zhao

L.X.

, Xu

, Yao

M.H.

, Zhang

A.H.

, Xu

H.X.

, et al., Diagnostic value of virtual touch tissue imaging quantification for benign and malignant breast lesions with different sizes, Int J Clin Exp Med 8 (2015), 13118–13126.

17.

Ophir

, Cespedes

, Ponnekanti

, Yazdi

and Li

, Elastography: A quantitative method for imaging the elasticity of biological tissues, Ultrasonic imaging 13 (1991), 111–134.

18.

Palmeri

M.L.

, Wang

M.H.

, Dahl

J.J.

, Frinkley

K.D.

and Nightingale

K.R.

, Quantifying hepatic shear modulus in vivo using acoustic radiation force, Ultrasound Med Bio 34 (2008), 546–558.

19.

Sarvazyan

, Hall

T.J.

, Urban

M.W.

, Fatemi

, Aglyamov

S.R.

and Garra

B.S.

, An overview of elastography - an emerging branch of medical imaging, Curr Med Imaging Re 7 (2011), 255–282.

20.

Siu

A.L.

, Screening for breast cancer: U.S. preventive services task force recommendation statement, Ann Int Med 164 (2016), 279–296.

21.

Tagliafico

A.S.

, Calabrese

, Mariscotti

, Durando

, Tosto

, Monetti

, et al., Adjunct screening with tomosynthesis or ultrasound in women with mammography-negative dense breasts: Interim report of a prospective comparative trial, J Clin Oncol Epub ahead of print <2016 Mar 9> doi: 10.1200/JCO.2015.63.4147

22.

Talwalkar

J.A.

, Elastography for detecting hepatic fibrosis: Options and considerations, Gastroenterology 135 (2008), 299–302.

23.

Tang

, Xu

H.X.

, Bo

X.W.

, Liu

B.J.

, Li

X.L.

, Wu

, et al., A novel two-dimensional quantitative shear wave elastography for differentiating malignant from benign breast lesions, Int J Clin Exp Med 8 (2015), 10920–10928.

24.

Tozaki

, Isobe

and Sakamoto

, Combination of elastography and tissue quantification using the acoustic radiation force impulse (ARFI) technology for differential diagnosis of breast masses, JPN J Radiol 30 (2012), 659–670.

25.

Trottmann

, Marcon

, D’Anastasi

, Karl

, Stief

C.G.

, Reiser

, et al., The role of VTIQ as a new tissue strain analytics measurement technique in testicular lesions, Clin Hemorheol Microcirc 58(1) (2014), 195–209.

26.

Vinnicombe

S.J.

, Whelehan

, Thomson

, McLean

, Purdie

C.A.

, Jordan

L.B.

, et al., What are the characteristics of breast cancers misclassified as benign by quantitative ultrasound shear wave elastography? Eur Radiol 24 (2014), 921–926.

27.

Yao

, Wu

, Zou

, Xu

, Xie

, Wu

, et al., Diagnostic value of virtual touch tissue quantification for breast lesions with different size, Biomed Res Int 2014 (2014), 142504.

28.

Zengel

, Schrötzlmair

, Schwarz

, Paprottka

, Kramer

, Berghaus

, et al., Elastography: A new diagnostic tool for evaluation of obstructive diseases of the salivary glands; primary results, Clin Hemorheol Microcirc 50(1-2) (2012), 91–99.

29.

Zhang

Y.F.

, Xu

J.M.

, Xu

H.X.

, Liu

, Bo

X.W.

, Li

X.L.

, et al., Acoustic radiation force impulse elastography: A useful tool for differential diagnosis of thyroid nodules and recommending fine-needle aspiration: A diagnostic accuracy study, Medicine 94 (2015), e1834.