Application of contrast-enhanced ultrasound in the diagnosis of small breast lesions

Abstract

BACKGROUND:

Breast cancer is the most common cancer in women worldwide. The purpose of the study was to observe the features of contrast-enhanced ultrasound (CEUS) and the combination with Breast Imaging-Reporting and Data System (BI-RADS) of conventional ultrasound for assessing small breast lesions.

OBJECTIVES:

The study was to explore the small breast lesions’ features of contrast-enhanced ultrasound (CEUS) and the combination with Breast Imaging-Reporting and Data System (BI-RADS) of conventional ultrasound, in order to improve the diagnostic accuracy of early breast cancer.

METHODS:

105 lesions were subject to conventional US (ultrasound) and CEUS before operations or biopsies. Among 105 breast lesions, six patient diagnoses were established by thick core-needle biopsy, while the rest were all confirmed by surgery and pathology.

RESULTS:

Significant differences were found between benign and malignant lesions in qualitative and quantitative indexes (peak) of CEUS (P < 0.05). The qualitative features of malignant small breast lesions were as follows: (1) enhanced intensity within the lesion was not uniform (61/61,100%); (2) the speed of wash-in was earlier than the surrounding tissue (58/61, 95.1%); (3) lesion interior and the surrounding tissues had contrast vessel performance (61/61,100%). Peak of malignant lesions (35.77±11.45) was higher than that of benign lesions (31.96±10.76) (P < 0.05). The diagnostic performance of BI-RADS-US plus qualitative indexes (method one) in terms of area under receiver operating characteristic curve (AUROC) were the highest (i.e., AUROC = 0.817), in comparison with other combined diagnostic methods. The associated sensitivity, specificity and accuracy were 78.69%, 84.09% and 80.95%, respectively. With method one, however, was similar with US-BI-RADS in specificity, 11 malignant breast lesions were regarded as a higher classification of BI-RADS and classified into malignant group, which were identified as benign on US-BI-RADS originally.

CONCLUSIONS:

CEUS was useful to differentiate benign from malignant small breast lesions, and the combination of CEUS and BI-RADS-US can improve the early diagnosis of breast cancers.

Keywords

Small breast lesions contrast-enhanced ultrasound peak intensity breast imaging-reporting and data system

1 Introduction

Breast cancer is the most common cancer in women worldwide [1]. In the United States, it is estimated that 231840 of women will be diagnosed with breast cancer in 2015 [2]. Early breast cancer detection, diagnosis, and treatment have been shown to be the key to improving prognosis of patients and reducing mortality [3]. However, diagnosis of early breast cancer by ultrasound (US) is difficult. Clinically, there is a high rate of misdiagnosis of small size, non-palpable, and atypical ultrasonographic early stage breast cancer. The small breast lesions, the maximum diameters of which are less than 10 mm, are often in the early stage of diseases [4]. Thus, technological improvements to increase the rate of accurate diagnosis of early breast cancer is indispensable to improve early detection and patient outcomes.

Nowadays, numerous novel technical approaches come into use, one of which is ultrasound elastography (UE). It has been developed to measure the tissue stiffness so as to accurately distinguish benign and malignant breast lesions [5, 6]. And with the combination of UE and BI-RADS had a better diagnostic performance in the diagnosis of breast lesions [7, 8]. Contrast-enhanced ultrasound (CEUS) is also one of these technical approaches, and it has been widely used in the diagnosis of abnormalities in abdominal organs. In the diagnosis of liver tumors, its accuracy can be comparable with CT (Computed Tomography) and MRI (Magnetic Resonance Imaging) [9]. However, although many studies have been carried out, there is still no unified diagnostic criteria for examination of the breast. Some studies show that benign and malignant breast tumors can be differentiated by CEUS using SonoVue [10], as the morphology and distribution of the blood vessels are different in benign and malignant tumors. In malignant lesions, vessels are tortuous, irregular, and mainly focus on the peripheral region of the tumor. Conversely, the vessels of benign lesions are natural, and gently curves and courses along the margin of the tumor [11 –13]. Some studies show that the combined use of conventional US and CEUS has a better diagnostic performance than either method alone, and displays good agreement with MRI in the differentiation capability for benign and malignant breast lesions [14]. The study of Xiang LH et al. [15] show that the combined diagnosis of SWE and CEUS was a significant difference with SWE alone and CEUS alone, so it could help to differentiate between malignant and benign breast lesions. Others demonstrate that CEUS quantitative analysis alone offers an objective and reproducible assessment of lesion vascularization, with good correlation with the results of MRI [16]. Ji CL et al. [17] found that quantitative analysis of CEUS may be a useful and objective method in predicting pathological prognostic factors in breast IDCs. To date there have been no systematic and complete studies done on CEUS of breast cancer lesions of ≤10 mm in diameter.

The purpose of this study was to observe CEUS features alone and in combination with Breast Imaging Reporting and Data System contained Ultrasonography (BI-RADS-US) classification for the assessment of small breast lesions.

2 Materials and methods

2.1 Patients

This retrospective study was approved by the Ethical Committee of the Rui Jin Hospital and informed consent was obtained from all the patients. From December 2013 to October 2016, there were 1748 patients with BI-RADS category 3–5 lesions in the hospital breast disease diagnosis and treatment center who underwent surgery or core-needle biopsy. Among which there were 324 patients with lesions smaller than 10 mm in maximal diameter according to the final pathological assessment. From these 324 patients, 105 patients were enrolled in the study and examined by conventional US combined with CEUS before operation or core-needle biopsy. The enrollment criteria for the patients were as follows: (i) BI-RADS category3–5 lesions with size smaller than 10 mm in maximal diameter according to the final pathological assessment; (ii) lesions were examined by both conventional US and CEUS before operation or core-needle biopsy. (iii) Patients with complete medical information and no prior treatment performed on the lesions. From the population, 1424 lesions with size larger than 10 mm in maximal diameter, 216 lesions not examined by CEUS, and three lesions which had previously been treated by neoadjuvant chemotherapy were excluded. For patients with multiple lesions (nine patients had two lesions and four patients had three or more lesions), only the most suspicious and biggest lesion on US was included. The final diagnoses for all lesions were derived from histopathological results after core biopsy or surgery. The flowchart for the selection of patients is presented in Fig. 1.

Fig.1

Flowchart of the selection of breast lesions. CEUS = Contrast-Enhanced Ultrasonography.

2.2 Ultrasound examination and evaluation

2.2.1 Ultrasound examination

Conventional US and CEUS scanning were performed using the same ultrasound machine Mylab 90 (Esaote, Genoa, Italy). Conventional US and color Doppler US were performed by LA 532 transducer with a frequency of 13-4 MHz, while CEUS was evaluated by LA 522 transducer with a frequency of 9-3 MHz. The contrast agent was SonoVue (BR1, Bracco SpA, Milan, Italy), a sulfur hexafluoride-filled microbubble contrast agent. To avoid inter-observer variability, all lesions were examined with the patient in a supine position by one of the investigators with more than 5 years’ experience in breast US. All ultrasound scanning was typically performed less than one week before surgery.

First, conventional US and color Doppler US scanning were carried out to observe general features of breast tumors and to select the best tumor imaging in the maximum plane, respectively, from which both the tumors and the normal adjacent breast tissue could be observed. Subsequently, CEUS was performed with the following scanner settings: the selected plane included the lesion and its surrounding normal tissue if possible; range, 70 dB; the image depth was 3 or 4 cm, the single focus was at the bottom of the image, and the probe was stabilized manually and no pressure was exerted. The contrast agent SonoVue was injected at the same dose and fashion with CEUS (2.4 mL of SonoVue as a bolus through an antecubital vein, followed by a flush of 5 mL of 0.9% saline). During the CEUS scanning, the transducer was kept in a stable position without movement and the patient was asked to quiet breathing. The selected plane remained unchanged during the examination and real time imaging was recorded for up to 3 min. During the examination, close attention would be paid to the discomfort of patients, and the patients could leave if they had no abnormal reaction after rest for 15 min. All static and dynamic images were stored in the ultrasound systems, and then single frames in JPEG format and movie files in digital imaging communications in medicine (DICOM) format were stored in the hard disk of the ultrasonography machine for further analysis.

2.2.2 CEUS imaging analysis

Two investigators who had more than 5 y of experience in breast US and more than 3 y in breast CEUS were asked to analyze the images, blind and without access to patient clinical information. Disagreement was solved by consensus, and if different assessment were assigned, a consensus reached after discussion. Both investigators had been trained to review the images before the study.

Qualitative and quantitative analysis of the image was done.

Qualitative analysis includes: (1) enhanced uniformity with in the lesion (uniform or not), (2) speed of wash-in compared to the surrounding tissue (faster than the surrounding tissue or slower than the surrounding tissue), (3) lesion interior and the surrounding tissues contrast vessel performance (radiation-like peripheral blood vessels, penetrating blood vessels, ring-like peripheral high enhancement, no blood vessels).

Quantitative analysis was done by QontraXt contrast-enhanced ultrasound imaging quantitative analysis software. First, uniformly-spaced software choose 100 frames automatically. After tracing the observation area and tumor boundary, the three-dimensional perfusion parameters and perfusion curve would be generated automatically. The recording parameters include: (i) the peak [intensity/100]; (ii) the time to peak (TTP) [ms]; (iii) the Sharpness [l/s]; (iv) the area under the curve (AUC) [l/s].

2.3 Statistical analysis

Statistical analysis was conducted using the SPSS statistical package, version 20.0 for Windows (SPSS Institute, Cary, NC, USA). We comparatively analyzed the qualitative description of the CEUS between benign and malignant lesions by Chi-square test, and we got the P value of each index. The parameters of intensity-time curve were analyzed by t test. Any P values <0.05 were considered statistically significant. Quantitative data were expressed as mean±standard deviation (SD) and range, if normal distribution was achieved. The best cut-off value for each significant independent variable was obtained when the Youden index was maximal (sensitivity+specificity-1). Receiver operating characteristic (ROC) curves were plotted to assess the diagnostic performance of quantitative indexes, BI-RADS-US+ qualitative indexes, BI-RADS-US+ quantitative indexes and BI-RADS-US+ qualitative indexes + quantitative indexes.

2.4 Pathological analysis

All cases were specifically evaluated by one pathologist with 10-y experience in breast diseases. The maximal histologic section of each mass was selected for comparison with US images. For doubtful lesions, consensus was obtained between the radiologist and pathologist by comparing CEUS images and histologic localization and the enhancing and non-enhancing areas. The ultrasonic description of the lesion and the location of the pathological description were carefully checked.

We used three methods to adjust the BI-RADS-US: (1) Method One (BI-RADS-US+ qualitative indexes): Lesions with all the qualitative signs mentioned above were regarded as a higher classification of BI-RADS (eg.3 to 4a, 4a to 4b, 4b to 4c, 4c to 5). Others were retained in their original categories. (2) Method Two (BI-RADS-US+ quantitative indexes)+ Lesions with all the quantitative signs higher than their best cut-off values mentioned above were regarded as a higher classification of BI-RADS. Others were retained in their original categories. (3) Method Three (BI-RADS-US+ qualitative indexes + quantitative indexes): Lesions with all the qualitative and quantitative indexes mentioned above were regarded as a higher classification of BI-RADS. Others were retained in their original categories. The sensitivity, specificity, accuracy, positive predictive value and negative predictive value were calculated to compare the diagnostic performances for BI-RADS-US, qualitative indexes, quantitative indexes, BI-RADS-US+ qualitative indexes, BI-RADS-US+ quantitative indexes and BI-RADS-US+ qualitative indexes + quantitative indexes.

3 Results

3.1 Pathologic diagnoses

A total of 105 breast lesions in 105 women comprised the final cohort. Mean patient age was 53.3 y±13.4 y (range, 21–82 y). Of the 105 breast lesions, six patient diagnoses were established by thick core-needle biopsy (2 ductal carcinoma in situ (DCIS), 2 invasive ductal carcinoma (IDC), 1 intraductal papilloma and 1 adenopathy), while the rest were all confirmed by surgery and pathology. The final pathologic diagnoses confirmed that 61 (58.1%) were malignant (Fig. 2) and 44 (41.9%) were benign (Fig. 3). The 44 benign lesions included fibroadenoma (n = 11, 25.0%), adenosis (n = 13, 29.5%), intraductal papilloma (n = 13, 29.5%), inflammation (n = 3, 6.8%), interstitial fibrosis of breast (n = 2, 4.5%), lacteal cyst (n = 1, 2.3%), and mammary duct ectasia (n = 1, 2.3%). The 61 malignant lesions included IDC (n = 33, 54.1%), DCIS (n = 10, 16.4%), mixed carcinoma (n = 12, 19.7%) (mixed carcinoma refers to the small breast lesions include both ductal carcinoma in situ and invasive ductal carcinoma. Different lesions had different proportions of these two kinds of cancer), adenocarcinoma (n = 3, 4.9%), invasive papillary carcinoma (n = 2, 3.3%), and solid papillary carcinoma (n = 1, 1.6%). According to the pathological results, the mean maximum diameter of the lesions was 7.5 mm±2.8 mm (range, 2–10 mm).

Fig.2

An 82-y-old woman with ductal carcinoma in situ. (a) A solid, hypoechoic, irregular shaped, indistinct margin lesion (arrows) was shown on US, which was classified as BI-RADS 4c. (b) 20 s after contrast agent injection Contrast enhanced ultrasound showed a heterogeneously and hyper-enhanced lesion (arrows) with penetrating blood vessels (red with white arrows). (c) The three-dimensional perfusion parameters and perfusion curve analysis showed the peak intensity was 54.3 (intensity/100), time to peak 44408[ms], sharpness 0.151[1/s], the area under the curve 8.2[1/s], respectively (d) Histopathological analysis showed as ductal carcinoma in situ (0.8 cm in diameter) revealed many and large neovascularization in and around the lesion (Hematoxylin-eosin stain,×100).

Fig.3

A 35-y-old woman with fibroadenoma. (a) A solid, hypoechoic, regular shaped, well-defined margin lesion (arrows) was shown on US, which was classified as BI-RADS 4a. (b) 30 s after contrast agent injection Contrast enhanced images showed a homogeneous and equal-enhanced lesion (arrows). (c) The three-dimensional perfusion parameters and perfusion curve analysis showed the peak intensity was 24.7 (intensity/100), time to peak 81295[ms], sharpness 0.098[1/s], the area under the curve 2.4[1/s] respectively. (d) Histopathological analysis showed as fibroadenoma (1×0.6×0.4 cm) revealed few neovascularization inside the lesion (Hematoxylin-eosin stain,×100).

3.2 Qualitative and quantitative analysis of CEUS images

There were statistically significant differences between benign and malignant small breast lesions in many qualitative indexes of CEUS (P < 0.05). The qualitative features of malignant small breast lesions were as follows: (1) enhanced intensity within the lesion was not uniform (61/61, 100%); (2) the speed of wash-in was earlier than the surrounding tissue (58/61, 95.1%); (3) lesion interior and the surrounding tissues had contrast vessel performance (radiation-like peripheral blood vessels, penetrating blood vessels, ring-like peripheral high enhancement) (61/61, 100%).

In quantitative analysis, it was found that there was a significant difference between malignant group and benign group in quantitative indexes Peak (intensity/100) (P < 0.05). Peak of malignant lesions (35.77±11.45) was higher than that of benign lesions (31.96±10.76) (P < 0.05). According to the ROC curve, the best cut-off values of Peak was 25.65 (Youden index = 0.619). The sensitivity and specificity of peak intensity are 90.91% and 70.97%, respectively, at this cut-off value.

3.3 BI-RADS regulate category on CEUS

B-mode US BI-RADS assigned categories of the breast lesions before modification are shown in Table 1. The malignancy rates of small breast lesions before adjustment were as follows: 50.0% (1 of 2) for category 3, 20.0% (7 of 35) for category 4a, 68.2% (15 of 22) for category 4b, 80.6% (25 of 31) for category 4c and 86.7% (13 of 15) for category 5. We modified the category of BI-RADS according to the qualitative and quantitative features of CEUS. There was 1 malignant lesion of BI-RADS 3, 3 malignant lesions of BI-RADS 4a, and 3 malignant lesions of BI-RADS 4b that were modified to the higher malignant potential category (Table 1).

Table 1
B-mode US BI-RADS before and after modification

Before adjustment After adjustment

Overall Malignant Benign Overall Malignant Benign

BI-RADS 3 2 1 (1.6%) 1 (2.2%) 1 0 (0%) 1 (2.2%)

BI-RADS 4a 35 7 (11.4%) 28 (63.6%) 32 4 (6.8%) 28 (63.6%)

BI-RADS 4b 22 15 (24.6%) 7 (15.9%) 19 12 (19.7%) 7 (15.9%)

BI-RADS 4c 31 25 (40.9%) 6 (13.6%) 25 20 (32.8%) 5 (11.4%)

BI-RADS 5 15 13 (21.3%) 2 (4.5%) 28 25 (40.9%) 3 (6.8%)

	Before adjustment	After adjustment
BI-RADS 3	2	1 (1.6%)	1 (2.2%)	1	0 (0%)	1 (2.2%)
BI-RADS 4a	35	7 (11.4%)	28 (63.6%)	32	4 (6.8%)	28 (63.6%)
BI-RADS 4b	22	15 (24.6%)	7 (15.9%)	19	12 (19.7%)	7 (15.9%)
BI-RADS 4c	31	25 (40.9%)	6 (13.6%)	25	20 (32.8%)	5 (11.4%)
BI-RADS 5	15	13 (21.3%)	2 (4.5%)	28	25 (40.9%)	3 (6.8%)

3.4 Diagnostic performances of the combined diagnosis

The sensitivity, specificity, accuracy, positive predictive value and negative predictive value for various methods are shown in Table 2. Since the lesions of BI-RADS category 4c were regarded as malignant, conventional US achieved 60.66% (37 of 61) sensitivity, 84.09% (37 of 44) specificity, and 70.48% (74 of 105) diagnostic accuracy. The sensitivity and accuracy of the various methods combining BI-RADS-US with qualitative indexes, quantitative indexes (peak), and qualitative and quantitative (peak) indexes together were each higher than that of BI-RADS-US alone. Interestingly, the specificity was similar to that of BI-RADS-US in all cases. With regard to sensitivity, the combination of BI-RADS-US and qualitative indexes achieved the best sensitivity of 83.61%, which was higher than other combined diagnostic methods.

Table 2
The diagnostic performance between malignant and benign lesions for US BI-RADS, qualitative indexes, quantitative indexes (Peak) and combinations

Methods sensitivity (%) specificity (%) accuracy (%) PPV (%) NPV (%) AUROC P value

BI-RDAS-US 60.66 (37 of 61) 84.09 (37 of 44) 70.48 (74 of 105) 84.09 (37 of 44) 60.66 (37 of 61) 0.726 0.000

Method One 78.69 (48 of 61) 84.09 (37 of 44) 80.95 (85 of 105) 87.27 (48 of 55) 74.00 (37 of 50) 0.817 0.000

Method Two 83.61 (51 of 61) 77.27 (34 of 44) 80.95 (85 of 105) 83.61 (51 of 61) 77.27 (34 of 44) 0.808 0.000

Method Three 72.13 (44 of 61) 84.09 (37 of 44) 77.14 (81 of 105) 86.27 (44 of 51) 68.52 (37 of 54) 0.783 0.000

Methods	sensitivity (%)	specificity (%)	accuracy (%)	PPV (%)	NPV (%)	AUROC
BI-RDAS-US	60.66 (37 of 61)	84.09 (37 of 44)	70.48 (74 of 105)	84.09 (37 of 44)	60.66 (37 of 61)	0.726
Method One	78.69 (48 of 61)	84.09 (37 of 44)	80.95 (85 of 105)	87.27 (48 of 55)	74.00 (37 of 50)	0.817
Method Two	83.61 (51 of 61)	77.27 (34 of 44)	80.95 (85 of 105)	83.61 (51 of 61)	77.27 (34 of 44)	0.808
Method Three	72.13 (44 of 61)	84.09 (37 of 44)	77.14 (81 of 105)	86.27 (44 of 51)	68.52 (37 of 54)	0.783

AUROC = area under the ROC curve; PPV = positive predictive value; NPV = negative predictive value; Method One = BI-RADS-US+ qualitative indexes; Method Two = BI-RADS-US+ quantitative indexes Peak; Method Three = BI-RADS-US+ qualitative indexes + quantitative indexes Peak.

With the combination of BI-RADS-US, qualitative indexes and Peak, five benign lesions were misdiagnosed as malignant. They were BI-RADS category 4c and 5 lesions on conventional US. Ten malignant lesions misdiagnosed as benign were 5 IDC, 2 DCIS, 1 mixed carcinoma and 2 others.

4 Discussion

Small breast cancer lesions have distinct ultrasonographic features from those typically observed in larger tumors, thus, it is difficult to diagnose these breast diseases. If these lesions are malignant, they are still in the early stages of invasion. Early detection positively impacts on the treatment, prognosis, and survival rate of breast cancer patients. There are numerous differences between treatments for tumors of varying size described in the National Comprehensive Cancer Network (NCCN) guidelines [3]. Therefore, it is necessary to find a method to improve diagnostic efficiency and to reduce unnecessary biopsies or medical examinations, which increases costs for patients and causes pain [18]. The small lesions of small breast lumps are small in the early stage of the disease, and the lack of typical ultrasonographic signs is also one of the causes of missed diagnosis.

From the perspective of clinical management for small breast lesions, a high sensitivity is necessary. Several studies have found that categorizing by conventional US-BI-RADS provides a high diagnostic specificity and accuracy [19]. In the present study, US-BI-RADS had high specificity (84.09%), but low sensitivity (60.66%) for small breast cancers, meaning that many malignancies would be missed. BI-RADS assessment is usually completed after routine work-up by investigators. This leads to poor inter-observer consistency in classification. This poor reproducibility reflects the highly subjective nature of this classification [20]. The small lesions with small size and atypical ultrasonographic in the early stage of breast cancer is also one of the causes of missed diagnosis [21].

Many studies have demonstrated that some qualitative characteristics are useful in differential diagnosis of breast lesions [22 –24], but there are few studies about small breast lesions. Our study showed that there are statistically significant differences in a total of 3 single qualitative CEUS features between benign and malignant small breast lesions. 100% of small malignant lesions were enhanced heterogeneously, 95.1% enhanced faster than the surrounding tissue, and 100% of small malignant lesions had the lesion interior and the surrounding tissues contrast vessel performance. This was similar with the work by Wang Y et al. [25], who reported that there are significant differences in enhancement degree, enhancement order, internal homogeneity, enhancement margin, surrounding vessels and enlargement of diameters between benign and malignant lesions. The reason for this difference might be the richer blood supply and tumor heterogeneity in malignant lesions. The CEUS contrast agent is purely intravascular, and endothelial permeability does not affect the distribution of contrast agent [26]. On the other hand, malignant breast lesions have increased angiogenic activity in order to support faster growth [27]. Furthermore, pathologic study revealed that malignant tumor vessels were caliber irregular, with more arterio-venous shunts, sinusoids, incomplete vascular walls and irregular course of intra-tumoral vessels [28, 29]. As a result, blood perfusion will be increased, allowing contrast agent to flow more quickly through the tumor microvasculature in CEUS time-course imaging. Although other studies are focused on breast tumors which were not limited in size, we found that small breast tumors have the typical characteristics of CEUS, and it provides a diagnosis basis for early breast cancer.

As for quantitative analysis, the CEUS time-intensity curve reflects the speed and quantity of contrast agent within the lesion consecutively. It has been proven that the peak was significantly higher [22] in malignant breast lesions. In our study, compared to benign lesions, the peak intensity of malignant small lesions was significantly higher (P < 0.05). The density of capillaries in the tumor may contribute this phenomenon. It has been reported that the faster and higher enhancement in contrast signal in malignant tumors may be associated with increased tumor microvascular density [23]. The sensitivity and specificity of the peak were 90.91% and 70.97%, respectively, meaning that many unnecessary biopsies were carried out. Therefore, obtaining a balance between sensitivity and specificity is important and relevant in clinical practice.

In general, the combined diagnostic techniques all improved the diagnostic performance for BI-RADS category 3–5 small breast lesions, as shown in the present study. The combined use of the CEUS techniques might be another option to improve the sensitivity. In a study of 91 breast lesions, the combination of qualitative with quantitative analysis achieved 83.0% sensitivity and 88.6% specificity [22]. In the present study, method one (BI-RADS-US with qualitative indexes) yielded sensitivity and specificity of 78.69% and 84.09%, method two (BI-RADS-US with quantitative (peak) indexes) yielded sensitivity and specificity of 83.61% and 77.27%, and method three (BI-RADS-US with both qualitative and quantitative indexes) yielded sensitivity and specificity of 72.13% and 84.09%, respectively. The sensitivity of the combination methods we examined were higher than that of BI-RADS-US used alone. This was similar with the work by Zhang JX et al. [21] who reported that qualitative features could improve diagnostic sensitivity, reduce the negative likelihood ratio, and improve the Negative Predictive Value (NPV). However, this was not sufficient to improve diagnostic specificity or the Positive Predictive Value (PPV). In terms of AUROC, method one achieved the highest overall diagnostic performance, compared with methods two and three. By using the combined method one, 11 lesions were changed from category 4b to category 4c and 7 lesions from category 4a to category 4b, according to the qualitative indexes of enhancement determined using CEUS. All of these lesions were malignant, and they all meet the three qualitative indicators at the same time. This led to a reduction of missed diagnosis, and only 15.91% of malignancies would be missed if this approach were used as the standard for deciding biopsy. The four false negative lesions maintained the original BI-RADS 4b classification for the combined method were 2 IDC, 1 aenocarcinoma, and 1 mixed carcinoma. They all enhanced slower than the surrounding tissue.

There were several limitations in this study. First, this was a study with a limited sample size. The malignancy seems to be quite high (58%, 61/105), so further studies with a larger sample size are required to validate the present results and to reduce the deviation from the sample. Second, quantitative index analysis of CEUS is limited, and there is a need to further analyze the significance of other quantitative indicators. Third, there is no uniform standard for qualitative and quantitative analysis of CEUS, further emphasizing the need to be confirmed by large sample studies.

5 Conclusions

In conclusion, CEUS was useful to differentiate benign from malignant small breast lesions, and the combination of CEUS and BI-RADS-US can improve the early diagnosis of breast cancers.

Footnotes

Acknowledgments

This work was supported by the National Natural Science Foundation of China [grants 81470079 and 81172078], the Scientific research project of the Shanghai municipal health and Planning Committee [grands 20164Y0162].

References

Torre

, Bray

, Siegel

, Ferlay

, Lortet-Tieulent

, Jemal

. Global cancer statistics, 2012. CA Cancer J Clin. 2015;65(2):87–108.

Siegel

, Miller

, Jemal

. Cancer statistics, 2015. CA Cancer J Clin. 2015;65(1):5–29.

NCCN Clinical Practice Guidelines in Oncology- Acute Lymphoblastic Leukemia (2016 Version I) [DB/OL]

Uehiro

, Horii

, Iwase

, Tanabe

, Sakai

, Morizono

, et al. Validation study of the UICC TNM classification of malignant tumors, seventh edition, in breast cancer. Breast Cancer. 2014;21(6):748–53.

Zhou

, Wang

, Ren

, Li

, He

, Liu

, et al. Value of shear wave arrival time contour display in shear wave elastography for breast masses diagnosis. Sci Rep. 2017;7(1):7036.

Ren

, Li

, He

, Li

, Wang

, Zhao

, et al. Two-dimensional shear wave elastography of breast lesions: Comparison of two different systems. Clin Hemorheol Microcirc. 2017;66(1):37–46.

, Zhao

, Yao

, Xu

, Liu

, Xu

, et al. Conventional US combined with acoustic radiation force impulse (ARFI) elastography for prediction of triple-negative breast cancer and the risk of lymphatic metastasis. Clin Hemorheol Microcirc. 2017;65(4):335–47.

, Xu

, Guo

, Bo

, Li

, 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. 2016;26(9):3290–300.

Claudon

, Dietrich

, Choi

, Cosgrove

, Kudo

, Nolsøe

, et al. World Federation for Ultrasound in Medicine; European Federation of Societies for Ultrasound. Guidelines and good clinical practice recommendations for contrast enhanced ultrasound (CEUS) in the liver-update 2012 a WFUMB-EFSUMB initiative in cooperation with representatives of AFSUMB, AIUM, ASUM, FLAUS and ICUS. Ultrasound Med Biol. 2013;39(2):187–210.

10.

Zhao

, Xu

, Ouyang

, Chen

, Dong

, Huihua

. Contrast-enhanced ultrasound is helpful in the differentiation of malignant and benign breast lesions. Eur J Radiol. 2010;73(2):288–93.

11.

Raza

, Baum

. Solid breast lesions: Evaluation with power Doppler US. Radiology. 1997;203(1):164–8.

12.

Less

, Skalak

, Sevick

, Jain

. Microvascular architecture in a mammary carcinoma: Branching patterns and vessel dimensions. Cancer Res. 1991;51(1):265–73.

13.

Kedar

, Cosgrove

, McCread

, Bamber

, Carter

. Microbubble contrast agent for color Doppler US: Effect on breast masses-work in progress. Radiology. 1996;198(3):679–86.

14.

, Wang

, Wan

, Hua

, Fang

, Chen

, et al. Differentiating benign from malignant solid breast lesions: Combined utility of conventional ultrasound and contrast-enhanced ultrasound in comparison with magnetic resonance imaging. Eur J Radiol. 2012;81(12):3890–99.

15.

Xiang

, Yao

, Xu

, Pu

, Liu

, Fang

, Wu

. Diagnostic value of contrast-enhanced ultrasound and shear-wave elastography for breast lesions of sub-centimeter. Clin Hemorheol Microcirc. 2017;67(1):69–80.

16.

Caproni

, Marchisio

, Pecchi

, Canossi

, Battista

, D’Alimonte

, et al. Contrast-enhanced ultrasound in the characterisation of breast masses: Utility of quantitative analysis in comparison with MRI. Eur Radiol. 2010;20(6):1384–95.

17.

, Li

, He

, Li

, Gu

, Xu

. Quantitative parameters of contrast-enhanced ultrasound in breast invasive ductal carcinoma: The correlation with pathological prognostic factors. Clin Hemorheol Microcirc. 2017;66(4):333–45.

18.

Tang

, Xu

, Bo

, Liu

, Li

, Wu

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

19.

Agarwal

, Pradeep

, Aggarwal

, Yip

, Cheung

. Spectrum of breast cancer in Asian women. World J Surg. 2007;31:1031–40.

20.

Abdullah

, Mesurolle

, El-Khoury

, Kao

. Breast imaging reporting and data system lexicon for US: Interobserver agreement for assessment of breast masses. Radiology. 2009;252(3):665–72.

21.

Zhang

, Cai

, Chen

, Dai

, Song

. CEUS helps to rerate small breast tumors of BI-RADS category 3 and category 4. Biomed Res Int. 2014;2014:572532.

22.

Wan

, Du

, Fang

, Li

, Wang

. Evaluation of breast lesions by contrast enhanced ultrasound: Qualitative and quantitative analysis. Eur J Radiol. 2012;81(4):e444–e50.

23.

Saracco

, Szabó

, Aspelin

, Leifland

, Wilczek

, Celebioglu

, et al. Differentiation between benign and malignant breast tumors using kinetic features of real-time harmonic contrast-enhanced ultrasound. Acta Radiol. 2012;53(4):382–8.

24.

Liu

, Jiang

, Liu

, Zhu

, Sun

, Chang

. Contrast-enhanced breast ultrasonography: Imaging: Features with histopathologic correlation. J Ultrasound Med. 2009;28(7):911–20.

25.

Wang

, Fan

, Zhao

, Zhang

, et al. Qualitative, quantitative and combination score systems in differential diagnosis of breast lesions by contrast-enhanced ultrasound. Eur J Radiol. 2016;85(1):48–54.

26.

Weskott

. Emerging roles for contrast-enhanced ultrasound. Clin Hemorheol Microcirc. 2008;40(1):51–71.

27.

, Wen

, Yang

. Inspection on angiogenesis in malignant transformation of breast tumor by ultrasound contrast and quantitative analysis. Zhonghua Yi Xue Za Zhi. 2009;89(9):587–91.

28.

Kook

, Park

, Lee

, Pae

, Park

. Evaluation of solid breast lesions with power doppler sonography. J Clin ultrasound. 1999;27(5):231–7.

29.

Nakopoulou

, Stefanaki

, Panayotopoulou

, Giannopoulou

, Athanassiadou

, Gakiopoulou-Givalou

, et al. Expression of the vascular endothelial growth factor receptor-2/Flk-1 in breast carcinomas: Correlation with proliferation. Hum Pathol. 2002;33(9):863–70.

	Before adjustment			After adjustment
	Overall	Malignant	Benign	Overall	Malignant	Benign
BI-RADS 3	2	1 (1.6%)	1 (2.2%)	1	0 (0%)	1 (2.2%)
BI-RADS 4a	35	7 (11.4%)	28 (63.6%)	32	4 (6.8%)	28 (63.6%)
BI-RADS 4b	22	15 (24.6%)	7 (15.9%)	19	12 (19.7%)	7 (15.9%)
BI-RADS 4c	31	25 (40.9%)	6 (13.6%)	25	20 (32.8%)	5 (11.4%)
BI-RADS 5	15	13 (21.3%)	2 (4.5%)	28	25 (40.9%)	3 (6.8%)