The value of ultrasound elastography combined with multi-parameters in the diagnosis of BI-RADS 4 breast lesions

Abstract

BACKGROUND:

Female breast cancer has surpassed lung cancer as the most commonly diagnosed cancer, with an estimated 2.3 million new cases (11.7%) in the global cancer statistics 2020.

OBJECTIVE:

To evaluate the diagnostic value of ultrasound elastography combined with multi-parameters in differentiating category 4 benign and malignant lesions in the breast imaging reporting and data system (BI-RADS).

METHODS:

This study retrospectively analyzed 206 patients (213 breast lesions) who visited the Department of Breast Surgery and underwent a breast core needle biopsy in the Department of Ultrasound in Peking University First hospital from April to December 2019. The shear wave velocity (SWV) values were collected at the following locations by virtual touch tissue imaging quantification (VTIQ): breast lesion interior, breast lesion margin, surrounding glands, and surrounding fat. Simultaneously, the strain ratio (SR) of breast lesions to glands and the area ratio (AR) of breast lesions were collected under strain elastography and a two-dimensional ultrasound mode.

RESULTS:

Univariate analysis found that the SWV value, measured by ultrasound elastography parameters, and the AR between the elasticity and the two-dimensional ultrasound breast lesions showed statistical differences when differentiating benign and malignant lesions ( $p<$ 0.05). Binary logistic regression analysis found that the SWV values of the lesion interior and the surrounding glands were statistically significant. The joint predictors were calculated and analyzed by Receiver Operating Characteristic (ROC), and it was found that the joint predictors and the SWV values of the lesion interior have great diagnostic value. The cut-off value, sensitivity and specificity of the joint predictor and the SWV value of the lesion interior were $>$ 3.65, 88.35% and 76.36% and $>$ 5.55 m/s, 79.61% and 82.73%, respectively.

CONCLUSIONS:

Ultrasound elastography combined with multi-parameters has good diagnostic value in differentiating BI-RADS 4 breast lesions.

Keywords

Ultrasound elastography strain elastography ARFI VTIQ SWV breast lesion

1. Introduction

Female breast cancer has surpassed lung cancer as the most commonly diagnosed cancer, with an estimated 2.3 million new cases (11.7%) in the global cancer statistics 2020 [1]. Among all malignant tumors in China, the incidence rate of breast cancer ranks fourth, and it ranks first among women in 2020 [2]. From 2003 to 2012, the five-year survival rate for tumors in China increased from 73% to 82%, but there is still a large gap between China and developed countries [3, 4, 5]. Early diagnosis and treatment are of great significance in improving the quality of life (QOL) and five-year survival rate of patients with breast cancer. Conventional ultrasound examination is simple and quick, non-invasive, non-destructive and repeatable. However, the accuracy of this method of diagnosis is insufficient, especially for non-mass breast cancer are easily missed, as it is influenced by the subjective consciousness of the operator [6]. Ultrasound elastography generates images of the elastic properties of tissues according to the difference in hardness between tissues [7]. This technology is non-invasive, does not require contrast medium and is quick, simple and easy to operate. The virtual touch tissue imaging quantification (VTIQ) uses the probe to emit focused acoustic pulses that cause the region of interest (ROI) to generate shear waves transversely. Then, it makes a qualitative and quantitative assessment of tissue hardness. VTIQ can improve the independence and repeatability of the examiner. This decreases the risk of of breast cancer, avoids unnecessary core needle biopsy and surgery. It shows good prospects for application in differentiating benign and malignant breast tumors [8].

Most previous studies have aimed to evaluate the diagnostic value of ultrasound elastography in differentiating benign and malignant breast lesions with respect to BI-RADS 3-5 nodules [9, 10, 11, 12]. while few studies have focused on BI-RADS 4 breast lesions [13, 14, 15]. The study described in this paper applied strain elastography to obtain semi-quantitative parameters (SR and AR) based on two-dimensional ultrasonography. Shear wave elastography was used to obtain a qualitative color-coding graph and a quantitative parameter (SWV value) to assess BI-RADS 4 breast lesions. Then, the diagnostic efficacy of these parameters in distinguishing benign and malignant breast lesions was studied to find a non-invasive, low-cost, simple, quick and easily operable auxiliary diagnostic method that can be introduced in primary hospitals. The hope is to improve the sensitivity and specificity of diagnosis and lessen unnecessary core needle biopsy.

2. Materials and methods

2.1 Patients

This study included patients who visited the Department of Breast Surgery and underwent breast core needle biopsy in the Department of Ultrasound of Peking University First Hospital from April to December 2019. A retrospective study was conducted. This study was conducted with approval from the Ethics Committee of Biomedical research in Peking University first hospital (2020-053). Written informed consent was obtained from all participants.

Inclusion criteria: (1) patients with a breast lesion diagnosed by routine ultrasound examination; (2) patients who received ultrasonography (including conventional ultrasound and ARFI elastography) before core needle biopsy, with complete ultrasound image data recorded; (3) patients with complete clinicopathologic data; (4) patients who received no surgical treatment, chemotherapy or radiotherapy before they underwent breast core needle biopsy.

Exclusion criteria: (1) patients with incomplete ultrasound image data or clinicopathologic data; (2) patients who received surgical treatment, chemotherapy or radiotherapy before they underwent breast core needle biopsy. According to the above inclusion and exclusion criteria, a total of 294 patients (305 breast lesions) were included. Eighty-eight patients (92 breast lesions) were excluded for the following reasons (Fig. 1): (1) 28 patients (29 breast lesions) with BI-RADS 5 lesions, n (breast lesion) $=$ 276; (2) 25 patients (28 breast lesions) with BI-RADS 3 lesions, n (breast lesion) $=$ 248; (3) one patient (one breast lesion) with a BI-RADS 6 lesion that had been confirmed as breast cancer by breast biopsy and who received a repeat biopsy, n (breast lesion) $=$ 247; (4) 29 patients (29 lesions) who received a biopsy of the axillary lymph nodes or axillary lesions only, n (breast lesion) $=$ 218; (5) five patients (five breast lesions) with part of the measurement values missing, n (breast lesion) $=$ 213.

Figure 1.

Patient screening flowchart.

Finally, a total of 206 patients (213 breast lesions) with BI-RADS 4 lesions were included, including 204 female patients (211 breast lesions) and two male patients (2 breast lesions). These patients were aged 16 $\sim$ 93 years, 51.7 ( $\pm$ 14.2) years on average, and had a maximum breast lesion diameter of 0.37–5.0 cm, 2.04 ( $\pm$ 1.28) cm on average. Taking their pathology results as the golden standard, these 213 breast lesions were divided into two groups: the benign group (110/213, 51.6%) and the malignant group (103/213, 48.4%).

2.2 Image acquisition

A Siemens Acuson OXANA3 ultrasound system (Siemens Healthcare, Germany) equipped with a linear 4.0 $\sim$ 9.0 MHz multifrequency probe was used to perform ARFI and US. All examinations were performed by sonographers with more than 10 years’ experience in breast US examination.

The patient was in a supine position and lifted his or her hands to fully expose the bilateral breast glands and armpits. First, the breast tumor was found by conventional two-dimensional ultrasonography and the maximum diameter and vertical diameter of the tumor were measured, with images recorded. Then, in the VTIQ mode, the tumor was placed and observed in the center of the sampling frame as far as possible. The image was recorded, including the tumor margin and part of the gland tissue and adipose tissue around the tumor. At the same time, the patient was instructed to hold his or her breath. The elasticity of the mass and adjacent normal tissue were measured and recorded by using shear wave velocities (SWV). In the VTIQ mode, the hardest part of the tumor was found (avoiding obvious coarse calcification) to measure the SWV value of the tumor’s interior (measuring the SWV value of the part with the maximum hardness inside the tumor) and the SWV value of the tumor margin (measuring the SWV value at the margin with the maximum hardness of the tumor). In addition, the SWV value of the surrounding gland (at the same depth as the tumor) and fat was measured. The images of these values were recorded. The measurement range of the SWV value was 0.5 $\sim$ 10.0 m/s. When VTIQ mode was enabled, the maximum SWV value was 6.5 m/s by default. When the measured value was “XXX” m/s for the part with the greatest local tissue hardness, combined with the images, the SWV value was manually adjusted to the maximum value of 10.0 m/s. Then, the measurement was conducted again, combined with the images, and if the SWV value of the part with the greatest hardness remained “XXX m/s”, then this value was recorded as 10.0 m/s. These values were measured three times and averaged.

Then, in the eSie Touch mode and the two-dimensional mode for dual display was used to measure the SR value of the breast tumor and surrounding normal gland tissue at the same depth. The oval ROI was placed inside the tumor shown by elastography, and the other ROI, of the same size and shape, was placed in the normal gland tissue at the same depth as the lesion ROI. The ROI in the normal gland tissue needed to be separated from the tumor as much as possible, then the instrument automatically calculated their SR values. Next, the tumor area of the two modes was recorded in the strain elastography and two-dimensional dual display mode, and the instrument automatically calculated the AR of the strain elastography in the tumor area in the two-dimensional mode and recorded the image. The above values were measured three times and averaged.

2.3 Statistical analysis

Univariate analysis was conducted first, using the SPSS 25.0 statistical software. The SWV values of the tumor interior, the tumor margin, the surrounding gland and the surrounding fat were tested for normality. If they conformed to the normal distribution, a $t$ -test was conducted in two groups of independent samples, and if not, a non-parametric test was conducted. Next, multivariate analysis was performed. Variables showing a significant difference in univariate analysis were analyzed using binary logistic regression, and regression equations were established to discover the independent influencing factors for the diagnosis of benign and malignant breast lesions and to calculate the joint predictors. Following this, the receiver operating characteristic curve (ROC) of variables with a significant difference in univariate analysis and those with a significant difference after binary logistic regression analysis and joint predictors were generated using MedCalc19.1 software, the area under the curves (AUC) were obtained and the Youden index (sensitivity $+$ specificity $-$ 1) was calculated. The sensitivity, specificity, positive predictive value, negative predictive value, positive likelihood ratio and negative likelihood ratio of the corresponding variables were calculated, and the diagnostic efficiency of various methods was evaluated to discover the variables with the highest diagnostic efficiency and those most easily operable for clinical diagnosis. A figure of $p<$ 0.05 was considered to suggest that the difference was statistically significant.

3. Results

3.1 Benign group and the malignant group

There were 110 cases (110/213, 51.6%) in the benign group and 103 cases (103/213, 48.4%) in the malignant group, among which 107 cases belonged to category 4a (81 cases were benign (81/107, 75.7%) and 26 cases were malignant (26/107, 24.3%)), 60 cases belonged to category 4b (21 cases were benign (21/60, 35.0%) and 39 cases were malignant (39/60, 65.0%)) and 46 cases belonged to category 4c (8 cases were benign (8/46, 17.4%) and 38 cases were malignant (38/46, 82.6%)).

3.2 Pathology results

As pathologically confirmed, there were 110 cases in the benign group, including fibroadenoma (40 cases), intraductal papilloma (5 cases), adenopathy or sclerosing adenosis (9 cases), interstitial fibrous hyperblastosis and collagenosis (35 cases), pseudoangioma-like interstitial proliferation (4 cases), canalicular adenoma (2 cases), breast inflammatory lesion (7 cases), phyllodes tumor (2 cases) and usual ductal hyperplasia (6 cases). There were 103 cases in the malignant group, including invasive ductal carcinoma (71 cases), invasive lobular carcinoma (4 cases), mucinous carcinoma (5 cases), carcinoma in situ (17 cases), malignant epithelial tumor (1 case), lymphoma (1 case), well-differentiated neuroendocrine neoplasm (1 case), metaplastic carcinoma (2 cases) and adenoid cystic carcinoma (1 case) (Table 1).

Table 1
Histopathologic diagnoses of malignant and benign breast lesions in 4a, 4b, 4c

Pathology result	4a ( $N=$ 107)				4b ( $N=$ 60)				4c ( $N=$ 46)
	Benign		Malignant		Benign		Malignant		Benign		Malignant
Total	81	(75.7%)	26	(24.3%)	21	(35.0%)	39	(65.0%)	8	(17.4%)	38	(82.6%)
Fibroadenoma	33	(30.8%)			6	(10.0%)			1	(2.2%)
Intraductal papilloma	5	(4.7%)			0	(0.0%)			0	(0.0%)
Sclerosing adenosis	7	(6.5%)			2	(3.3%)			0	(0.0%)
Interstitial fibrous hyperblastosis	22	(20.6%)			8	(13.3%)			5	(10.9%)
and collagenosis
Pseudoangioma-like interstitial	4	(3.7%)			0	(0.0%)			0	(0.0%)
proliferation
Canalicular adenoma	2	(1.9%)			0	(0.0%)			0	(0.0%)
Breast inflammatory lesions	1	(0.9%)			4	(6.7%)			2	(4.3%)
Phyllodes tumor	1	(0.9%)			1	(1.7%)			0	(0.0%)
Usual ductal hyperplasia	6	(5.6%)			0	(0.0%)			0	(0.0%)
Invasive ductal carcinoma			20	(18.6%)			24	(40.0%)			27	(58.7%)
Invasive lobular carcinoma			0	(0.0%)			2	(3.3%)			2	(4.3%)
Mucinous carcinoma			3	(2.8%)			1	(1.7%)			1	(2.2%)
Carcinoma in situ			3	(2.8%)			10	(16.7%)			4	(8.7%)
Malignant epithelial tumor			0	(0.0%)			1	(1.7%)			0	(0.0%)
Diffuse large b call lymphoma			0	(0.0%)			1	(1.7%)			0	(0.0%)
Well-differentiated neuroendocrine			0	(0.0%)			0	(0.0%)			1	(2.2%)
neoplasm
Metaplastic carcinoma			0	(0.0%)			0	(0.0%)			2	(4.3%)
Adenoid cystic carcinoma			0	(0.0%)			0	(0.0%)			1	(2.2%)

Figure 2.

Female, 24 years old, tumor at 10 o’clock in the outer upper quadrant of the right breast, with a size of about 1.30 $\times$ 1.02 cm, BI-RADS 4b. a. Ordinary two-dimensional ultrasonography showed a hypoechoic nodule at 10 o’clock in the right breast, with oval shape, clear boundaries, and circumscribed margin, non-microcalcification and microcalcification; b. Elastic image in VTIQ mode c. SWV values of breast tumor interior, tumor margin, surrounding gland and surrounding fat measured in VTIQ mode; d. SR values of the breast tumor compared to the surrounding gland of the same depth measured in eSie Touch mode (left: two-dimensional ultrasonography, right: strain elastography); e. AR value of breast tumor in the strain elastography mode compared to that in the two-dimensional ultrasonography mode measured in eSie Touch mode (left: two-dimensional ultrasonography, right: strain elastography); f. Pathology image: breast fibroadenoma.

Figure 3.

Female, 69 years old, tumor at 2 o’clock in the outer upper quadrant of the left breast, with a size of about 2.10 $\times$ 1.82 cm, BI-RADS 4c. a. Ordinary two-dimensional ultrasonography showed a hypoechoic nodule at 2 o’clock in the left breast, with irregular shape, unclear boundaries, indistinct margin, b. Elastic image in VTIQ mode (left: two-dimensional ultrasonography, right: VTIQ mode); c. SWV values of breast tumor interior, tumor margin, surrounding gland and surrounding fat measured in VTIQ mode (left: two-dimensional ultrasonography, right: VTIQ mode); d. SR values of the breast tumor compared to the surrounding gland of the same depth measured in eSie Touch mode (left: two-dimensional ultrasonography, right: strain elastography); e. AR value of breast tumor in the strain elastography mode compared to that in the two-dimensional ultrasonography mode measured in eSie Touch mode (left: two-dimensional ultrasonography, right: strain elastography); f. Pathology image: invasive carcinoma of no special type was observed in the breast tissue by core needle biopsy, Grade I, score 5 (mitotic 2 figures/2 mm ${}^{2}$ ).

3.3 Univariate analysis

Univariate analysis was conducted for the measured SWV values of the tumor interior, the tumor margin, the surrounding gland and the surrounding fat; the SR of the breast tumor to the surrounding normal gland tissue and the AR of the tumor area. It was found that the SWV values of the tumor interior, the tumor margin, the surrounding gland and the surrounding fat and the AR values of the breast lesion were significantly different between the benign group and the malignant group ( $p<$ 0.05). Table 2 shows detailed data. (Figs 2 and 3).

Table 2
eSie Touch and VTIQ measurements for benign and malignant breast lesion

	Benign (Median (upper quartile $\sim$ lower quartile))	Malignant (Median (upper quartile $\sim$ lower quartile))	$P$
Inner SWV (m/s)	3.66 (3.00 $\sim$ 5.07)	6.85 (5.66 $\sim$ 8.27)	$<$ 0.001
Margin SWV (m/s)	3.31 (2.72 $\sim$ 4.30)	4.93 (4.04 $\sim$ 6.07)	$<$ 0.001
Gland SWV (m/s)	2.47 (2.01 $\sim$ 2.92)	2.62 (2.32 $\sim$ 3.16)	0.022
Fat SWV (m/s)	2.27 (1.97 $\sim$ 2.70)	2.58 (2.18 $\sim$ 2.93)	$<$ 0.001
SR	0.38 (0.21 $\sim$ 1.35)	0.36 (0.22 $\sim$ 0.67)	0.509
AR	1.19 (1.04 $\sim$ 1.45)	1.39 (1.21 $\sim$ 1.54)	$<$ 0.001

Note: Nonparametric test results of two groups of independent samples in univariate analysis. eSie Touch mode enables strain elastography, which includes SR and AR in the table. For SR, the strain ratio of breast lesions to surrounding glands is measured, and for AR, the area ratio of elastic mode to two-dimensional mode is measured, Inner SWV is the SWV value of the part with the greatest hardness inside the breast tumor, Margin SWV is the SWV value of the margin with the greatest hardness of the breast tumor, Gland SWV is the SWV value of glands around the breast tumor, and Fat SWV value is the SWV value of fat around the breast.

3.4 Multivariate analysis

Variables with a significant difference in univariate analysis were analyzed using binary logistic regression, and it was found that the SWV values of the tumor interior and the surrounding glands were statistically significant. Table 3 shows the results of binary logistic regression analysis. The regression equation was obtained, and the joint predictor was L $=$ 0.753 $\times$ InnerSWV $+$ 0.264 $\times$ MarginSWV $-$ 0.933 $\times$ GlandSWV $+$ 0.445 $\times$ FatSWV $+$ 0.107 $\times$ AR.

Table 3
Results of multivariate analysis and binary logistic regression analysis

	B	$P$	Exp (B)	EXP (B) 95% confidence interval
				Upper	Lower
Inner SWV (m/s)	0.753	0.000	2.123	1.646	2.739
Margin SWV (m/s)	0.264	0.125	1.302	0.930	1.824
Gland SWV (m/s)	$-$ 0.933	0.007	0.393	0.200	0.772
Fat SWV (m/s)	0.445	0.300	1.560	0.673	3.620
AR	0.107	0.838	1.112	0.401	3.083

Note: Multivariate analysis: Among the five independent variables, only Inner SWV and Gland SWV are independent influencing factors.

3.5 ROC analysis

Table 4 and Fig. 4 show the variables with a statistical difference in univariate analysis, including the AUC about the SWV values of the tumor interior, the tumor margin, the surrounding gland and the surrounding fat and the AR values of the breast lesion; the sensitivity; the specificity; the positive predictive value; the negative predictive value; the positive likelihood ratio and the negative likelihood.

Table 4
ROC curve analysis

	AUC area	95% confidence interval	Youden index	Cutoff	Sensitivity	Specificity	LR ( $+$ )	LR ( $-$ )	PV ( $+$ )	PV ( $-$ )
InnerSWV	0.86	0.806 $\sim$ 0.903	0.623	$>$ 5.55	79.61	82.73	4.61	0.25	81.2	81.2
JointPredictor	0.865	0.812 $\sim$ 0.908	0.647	$>$ 3.65	88.35	76.36	3.74	0.15	77.8	87.5
GlandSWV	0.591	0.522 $\sim$ 0.658	0.207	$>$ 2.19	82.52	38.18	1.33	0.46	55.6	70
MarginSWV	0.789	0.728 $\sim$ 0.842	0.482	$>$ 3.66	85.44	62.73	2.29	0.23	68.2	82.1
FatSWV	0.644	0.576 $\sim$ 0.708	0.282	$>$ 2.32	70.87	57.27	1.66	0.51	60.8	67.7
AR	0.647	0.579 $\sim$ 0.711	0.295	$>$ 1.19	77.67	51.82	1.61	0.43	60.2	71.2

Note: ROC analysis of independent influencing factors and joint predictors after univariate analysis and multivariate analysis demonstrate that Inner SWV and Joint Predictor have a good diagnostic efficiency, while Gland SWV has a poor diagnostic efficiency.

Table 5

ROC curve comparison

Difference ROC	Differencebetween areas	Standard error	95% confidence interval	Z statistic	Significance level
InnerSWV $\sim$ GlandSWV	0.269	0.0386	0.193 $\sim$ 0.344	6.953	$P<$ 0.0001
InnerSWV $\sim$ JointPredictor	0.00574	0.0112	$-$ 0.0162 $\sim$ 0.0277	0.512	$P=$ 0.6090
GlandSWV $\sim$ JointPredictor	0.274	0.0428	0.190 $\sim$ 0.358	6.411	$P<$ 0.0001

Note: ROC of Inner SWV and Joint Predictor are significantly different from those of Gland SWV, but there is no significant difference between Inner SWV and Joint Predictor.

Figure 4.

ROC analysis of the variables with statistical significance in univariate analysis and multivariate analysis and the joint predictors. AUC of InnerSWV, GlandSWV, MarginSWV, FatSWV, AR and Joint Predictor were 0.86, 0.591, 0.789, 0.644, 0.647 and 0.865, respectively.

For variables with statistical significance in the binary logistic regression analysis, an ROC analysis was carried out for the SWV values of the tumor interior and the surrounding glands and the joint predictors. Table 4 and Fig. 4 show the AUC, sensitivity, specificity, positive predictive value, negative predictive value, positive likelihood ratio and negative likelihood ratio. The SWV value of the tumor interior and the joint predictor have a higher diagnostic value for identifying benign and malignant breast lesions, while the SWV value of the surrounding gland has a lower diagnostic value. The Youden index and the cut-off values of the SWV values of the tumor interior and the surrounding gland and the joint predictor were 0.623, $>$ 5.55 m/s, 0.207, $>$ 2.19 m/s, 0.647 and $>$ 3.65. Detailed results are shown in Table 4 and Fig. 4. Comparing the ROC of the three variables, there was no significant difference between the ROC of the SWV value of the tumor interior and that of the joint predictor, while there was a significant difference between the ROC of these two and that of the SWV value of the surrounding gland ( $p<$ 0.05) (Table 5). Therefore, because there is no significant difference between the ROC of the SWV value of the tumor interior and that of the joint predictor, for convenience in clinical work, the SWV value of a tumor’s interior can be used as an important parameter to identify benign and malignant breast lesions. It is considered that a tumor interior SWV value $>$ 5.55 m/s indicates a malignant lesion and $<$ 5.55 m/s indicates a benign lesion.

4. Discussion

In univariate analysis for this group of BI-RADS 4 lesions analyzed retrospectively, there were significant differences between benign and malignant lesions in the SWV values of the tumor interior, the tumor margin, the surrounding gland and the surrounding fat and the AR values of the breast lesion. According to elastography principles, the SWV and the tissue elastic coefficient are closely related to tissue structure. The former two are positively related to the hardness of the tissue. Studies have shown that the elastic coefficient of tissues in the breast is different and is ordered from large to small as invasive ductal carcinoma $>$ non-invasive ductal carcinoma $>$ breast fibrosis $>$ breast tissue $>$ adipose tissue [16, 17]. The parameters of malignant breast lesions were larger than those of benign ones, a result which is similar to those of previous studies [12, 18, 19]. According to a previous study [20] the area of ultrasound elastography is a favorable index for differentiating malignant and benign solid tumors. In this study, the sensitivity, specificity, optimal cut-off value and AUC of the elastic area ratio in the diagnosis of malignant breast tumor were 77.67%, 51.82%, 1.19 and 0.647, respectively. Menezes et al. [11] measured the elastic area ratio of benign breast lesions to malignant breast lesions, and considered that its optimal cut-off value was 1.03, sensitivity was 92% and specificity was 64%. The cut-off value was similar to that determined by this study, but the sensitivity and specificity were greater than those in this study. In addition, another study conducted by Gürüf et al. also confirmed that SWE is capable of differentiating benign and malignant breast lesions (AUC $=$ 0.918) [21].

Gürüf et al. [21] evaluated and compared the diagnostic performances of shear wave elastography (SWE) and strain elastography (SE) in the differentiation of benign and malignant breast lesions. Their study included 87 breast lesions in 84 patients. In their study, the strain ratio (SR) was calculated as the ratio of lesion strain to the adjacent fat strain using SE. Selecting a cutoff SR value of 3.22 led to an 88.1% sensitivity and an 88.4% specificity. Selecting cutoff maximum SWV value of 3.41 m/s led to an 88.1% sensitivity and an 86.7% specificity. They concluded SE and SWE are both feasible imaging modalities in the differentiation of malignant and benign breast lesions with similar diagnostic performances. In this study, it was found that there was no significant difference in SR value between benign and malignant tumors which is not consistent with the previous study [11]. The reason may be the following: (1) BI-RADS 4 breast lesions were selected in this study, and the SR value of category 4 breast lesions compared to surrounding glands in the same layer is generally smaller than that of category 5 breast lesions, there may be no significant difference in statistical analysis in this study. (2) Measurement of the SR is influenced by many factors: it makes high demands on the operator, and the stability of the operation technique and the probe pressure may affect the accuracy of the measurement. In addition, interference from the patient’s respiratory movement, interference from aorta pulsation to adjacent tissues and differing muscular tension also affect the measured SR value. (3) The method selected in this study was to measure the ratio of the breast tumor to the gland tissue at the same depth. Although the ROI was measured by selecting the gland at the same depth, which was as far away from the tumor as possible, some tumors are large, so it was difficult to avoid including any unrecognizable gland tissues which had been infiltrated and hardened by tumors in the ROI when measuring gland tissue at the same level, resulting in a small SR value. Some tumors are relatively superficial, and there are few glandular tissues in the same layer. During the measurement, there may be some adipose tissues in the ROI, which results in a large SR value, and further affects the result.

The results of multivariate analysis showed that the SWV values of the tumor interior and the surrounding glands were statistically significant, creating an independent influencing factor for the differential diagnosis of benign and malignant breast lesions. On this basis, the joint predictor was calculated, and it was analyzed by ROC together with the above two variables. However, it was found that the AUC, sensitivity and specificity of the SWV values of the surrounding glands were 0.591, 82.52% and 38.18%, respectively, which indicates low diagnostic efficiency, high sensitivity and low specificity. Therefore, in clinical application, it is not feasible to use this value alone as an auxiliary index for identifying benign and malignant lesions. The AUC, sensitivity and specificity of the SWV values of the tumor interior and the joint predictors were 0.86, 79.61% and 82.73% and 0.865, 88.35% and 76.36%, respectively; therefore, these two variables show high diagnostic efficiency and can be used as important indexes for identifying benign and malignant lesions.

Comparing the ROC of the SWV value of the tumor interior, the SWV value of the surrounding gland and the joint predictor, it was found that the area under the ROC curve and sensitivity of the joint predictor were not much higher than those of the SWV value of the tumor interior, and there was no significant difference between their ROCs. The measurement of the SWV value of the tumor interior was quite simple and quick in clinical work. Many previous studies [9, 19, 22, 23] have examined the value of traditional ultrasound, contrast-enhanced ultrasonography, strain elastography and shear wave elastography in the differential diagnosis of benign and malignant breast lesions. All have concluded that ultrasound elastography demonstrates good diagnostic value for benign and malignant breast lesions and can be an effective, non-invasive and convenient diagnostic mode alongside traditional ultrasound in clinical work. The most important contribution of this study was to explore a method that is easily operated by sonographers and relatively stable in measurement in clinical work as an effective auxiliary diagnostic method alongside the traditional ultrasound diagnosis mode. When clinicians make recommendations concerning BI-RADS 4 lesions, whether they suggest follow-up examination or core needle biopsy, it is necessary to provide them with a more objective quantitative method rather than a method that relies on subjective judgment. This study suggested that the maximum SWV value of the tumor’s interior allows for effective classification of BI-RADS 4 breast lesions and provides more reasonable clinical suggestions.

This study also has some shortcomings: (1) The overall sample size was not large enough: after being subdivided into categories 4a, 4b and 4c, there were few cases in each category. Therefore, it is necessary to expand the sample size in a future study. (2) Since this was a retrospective study, it was not feasible to control for the operating physicians. Although all physicians had been previously trained to use the instrument, there were inevitably individual differences that may have increased the measurement error. In the subsequent prospective research, it will be necessary to control for the operating physician or conduct strict and standardized operation training for all physicians before the research begins.

5. Conclusion

The VTIQ based on ARFI demonstrates good diagnostic value for the differential diagnosis of benign and malignant breast lesions. It can be a non-invasive and effective auxiliary diagnostic mode alongside traditional ultrasound, and it has a high diagnostic efficiency in distinguishing between benign and malignant BI-RADS 4 breast lesions.

Ethics approval and consent to participate

The study was conducted with approval from the Ethics Committee of Biomedical Research at the Peking University First Hospital (2020-053). The study was conducted in accordance with the Declaration of Helsinki. Written informed consent was obtained from all participants prior to participation.

Funding

No external funding was received to conduct this study.

Footnotes

Conflict of interest

The authors declare that they have no competing interests.

References

Sung

Ferlay

Siegel

Laversanne

Soerjomataram

Jemal

Bray

. Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. CA Cancer J Clin. 2021; 71: 209-249.

Meng

Liu

Niu

Zhou

Zhao

Duan

Zeng

Xiong

Zhou

. Analysis of status and countermeasures of cancer incidence and mortality in China. Sci China Life Sci. 2019; 62: 640-647.

Bai

. Prevalence and disease characteristics of breast cancer in China. Shi Jie Zui Xin Yi Xue Xin Xi Wen Zhai. 2017; 17: 253+6.

. Status of breast cancer in China. Shi Jie Zui Xin Yi Xue Xin Xi Wen Zhai. 2019; 19: 371-2.

Siegel

Miller

Jemal

. Cancer statistics, 2020. CA: A Cancer Journal for Clinicians. 2020; 70: 7-30.

Yang

Hong

. Advances in clinical application of imaging diagnosis of breast tumor. Ren Min Jun Yi. 2017; 60: 1031-3+9.

Ophir

Céspedes

Ponnekanti

Yazdi

. Elastography: A quantitative method for imaging the elasticity of biological tissues. Ultrasound Imaging. 1991; 13: 111-34.

Zhang

Wan

Fang

. Value of conventional ultrasound, ultrasound elasticity imaging, and acoustic radiation force impulse elastography for prediction of malignancy in breast lesions. Clin Hemorheol Microcirc. 2020; 74: 241-253.

Zhao

Wang

Sun

Jiang

Tao

Zhou

. Comparison of strain and acoustic radiation force impulse elastography of breast lesions by qualitative evaluation. Clin Hemorheol Microcirc. 2018; 70: 39-50.

10.

Kim

Park

Kim

Lee

Ryu

. Diagnostic value of elastography using acoustic radiation force impulse imaging and strain ratio for breast tumors. J Breast Cancer. 2014; 17: 76-82.

11.

Menezes

Sardessai

Furtado

Sardessai

. Correlation of Strain Elastography with Conventional Sonography and FNAC/Biopsy. J Clin Diagn Res. 2016; 10: Tc05-10.

12.

Kapetas

Pinker-Domenig

Woitek

Clauser

Bernathova

Spick

Helbich

Baltzer

. Clinical application of Acoustic Radiation Force Impulse Imaging with Virtual Touch IQ in breast ultrasound: diagnostic performance and reproducibility of a new technique. Acta Radiol. 2017; 58: 140-7.

13.

Han

Peng

Huang

Lin

Liu

Cao

Zhou

. Reducing unnecessary biopsy of breast lesions: Preliminary results with combination of strain and shear-wave elastography. Ultrasound Med Biol. 2019; 45: 2317-27.

14.

Jin

Zhou

Chen

Huang

Dai

Chen

. Acoustic radiation force impulse elastography of breast imaging reporting and data system category 4 breast lesions. Clinical Breast Cancer. 2012; 12: 420-7.

15.

Magny

Shikhman

Keppke

16.

Krouskop

Wheeler

Kallel

Garra

Hall

. Elastic moduli of breast and prostate tissues under compression. Ultrasound Imaging. 1998; 20: 260-74.

17.

Samani

Zubovits

Plewes

. Elastic moduli of normal and pathological human breast tissues: An inversion-technique-based investigation of 169 samples. Physics in Medicine and Biology. 2007; 52: 1565-76.

18.

Liu

Huang

Guo

Sun

Fang

. Value of virtual touch tissue imaging quantification for evaluation of ultrasound breast imaging-reporting and data system category 4 lesions. Ultrasound Med Biol. 2016; 42: 2050-7.

19.

Zhang

Wan

Liu

Yao

Xiang

Fang

Jia

20.

Yang

. Application value of ultrasound elastography ratio method and area ratio method in differentiating benign and malignant breast masses. Zhong Guo Quan Ke Yi Xue. 2014; 17: 2904-6.

21.

Gürüf

Öztürk

Bayrak

İK

Polat

. Shear wave versus strain elastography in the differentiation of benign and malignant breast lesions. Turk J Med Sci. 2019; 49: 1509-1517.

22.

Kapetas

Clauser

Woitek

Wengert

Lazar

Pinker

Helbich

Baltzer

PAT

. Quantitative multiparametric breast ultrasound: Application of contrast-enhanced ultrasound and elastography leads to an improved differentiation of benign and malignant lesions. Investigative Radiology. 2019; 54: 257-64.

23.

Wang

Wan

. Preliminary results of acoustic radiation force impulse imaging by combined qualitative and quantitative analyses for evaluation of breast lesions. J Ultrasound Med. 2018; 37: 2405-12.

The value of ultrasound elastography combined with multi-parameters in the diagnosis of BI-RADS 4 breast lesions

Abstract

BACKGROUND:

OBJECTIVE:

METHODS:

RESULTS:

CONCLUSIONS:

Keywords

1. Introduction

2. Materials and methods

2.1 Patients

2.3 Statistical analysis

3. Results

3.1 Benign group and the malignant group

3.2 Pathology results

Table 1 Histopathologic diagnoses of malignant and benign breast lesions in 4a, 4b, 4c

Table 2 eSie Touch and VTIQ measurements for benign and malignant breast lesion

Table 3 Results of multivariate analysis and binary logistic regression analysis

Table 4 ROC curve analysis

5. Conclusion

Ethics approval and consent to participate

Funding

Footnotes

Conflict of interest

References

Table 1
Histopathologic diagnoses of malignant and benign breast lesions in 4a, 4b, 4c

Table 2
eSie Touch and VTIQ measurements for benign and malignant breast lesion

Table 3
Results of multivariate analysis and binary logistic regression analysis

Table 4
ROC curve analysis