Imaging examinations, especially molybdenum target radiography, magnetic resonance imaging, and ultrasonography play an important role in the early screening and diagnosis of breast diseases. Over the past three years, studies outside of China have suggested that large-scale breast cancer screening programs can reduce overall breast cancer mortality by 20% and early breast cancer mortality by 60% [11]. However, in comparison to women of other ethnicities, Chinese women generally have relatively smaller breast volumes with higher proportions of dense breast tissue; these characteristics can reduce the sensitivity of mammography for identifying malignant breast lesions. In these cases, ultrasound is a more effective method for characterizing suspicious lesions [12].
Conventional ultrasound can clearly show a variety of breast lesion characteristics, including location, number, size, shape, orientation, perimeter, internal echo, rear features, calcification, blood supply, and associated axillary lymph nodes. In clinical practice, however, two-dimensional ultrasound manifestations of benign and malignant nodules sometimes share overlapping characteristics, especially in patients with small breast tumors. The BI-RADS classification system, which standardizes breast ultrasonography reporting worldwide [13], has been shown to improve the sensitivity of identifying malignant masses; however, the false-positive rate remains high [14].
Based on years of clinical experience, we believe that the biological characteristics of small breast cancers are less commonly manifested and that signs of malignancy in these small tumors are more difficult to observe using two-dimensional ultrasound imaging. Even color Doppler imaging cannot adequately detect obvious differences in blood flow for benign nodules, making differential diagnosis even more difficult. In addition, the pathological characteristics of breast cancer tissues are diverse, which can lead to both false positives and false negatives. For example, some in situ cancers with small and atypical nodules do not display hyperechoic halos or burr signs. Moreover, some small invasive ductal carcinomas, mucinous carcinomas, and papillary carcinomas manifest with clear boundaries, few lobes, and parallel growth, leading to false negatives. Finally, some sclerosing adenoses, intraductal papillomas with inflammation, and other benign lesions exhibit obscure boundaries, leading to false infiltration and false positives. Therefore, the use of conventional ultrasonography, which relies on the recognition of distinctive morphological features, can be problematic during the evaluation of small breast nodules.
In addition to morphological features, other tumor characteristics have been evaluated for their applicability to imaging-based diagnoses. For example, previous studies have shown that angiogenesis plays an important role in tumor growth and metastasis [15]. Tumor blood flow has been evaluated using Doppler ultrasound with some success; however, this modality can only identify a blood flow signal above the threshold of large vessels and the wall filter [16]. On the other hand, real-time CEUS can provide more accurate information about the morphology and distribution of the blood vessels associated with tumors [15]. Moreover, SWE can quantitatively evaluate the hardness of breast nodules [17,18]. Therefore, conventional ultrasound combined with CEUS and SWE techniques can comprehensively evaluate nodules based on the shape, micro-blood flow, and hardness of nodules to improve diagnostic efficacy.
CEUS is an advanced technique for microperfusion imaging using a safe contrast medium. It can dynamically display the microperfusion of tumors, lymph nodes, and surrounding tissues in real-time, and it plays significant roles in the diagnosis and targeted therapy of breast diseases, detection of metastatic sentinel lymph nodes, and evaluation of neoadjuvant chemotherapy efficacy. CEUS can qualitatively characterize benign and malignant lesions by identifying the homogeneous and centripetal enhancement of benign lesions. In addition, it can quantitatively characterize lesions by creating time-intensity curves and obtaining a series of quantitative parameters, including rise time, peak time (time to peak), peak intensity, mean transit time, and the AUC under the ROC curve.
The present study found that most fibroadenomas and intraductal papillomas manifested with true capsules or pseudocapsules caused by expansive growth; accordingly, if there was high enhancement on CEUS, the boundary was clear and the enhancement was well distributed. Malignant lesions of the breast, on the other hand, often lacked a capsule, and the boundary and size of the tumor, as well as the shape of nutrient angiography, could only be seen after enhancement. Since angiogenesis is a risk factor for invasion and metastasis of solid tumors, there are generally many proliferative and active cells at the perimeter of malignant tumors, as well as numerous abnormal capillary networks with disordered structures and increased microvessel densities. At the same time, perfusion disruptions can lead to increased microcirculation flow and velocity [19]. In addition, the peripheral regions of malignant breast tumors are often associated with breast hyperplasia and precancerous lesions at different stages. With the progression of these precancerous lesions to breast cancer, neovascularization and blood vessel density also increase [20].
In this study, adenoses generally manifested without pseudocapsules and exhibited an irregular shape after enhancement, which was often low or equally distributed, with a focus that either remained unchanged, became narrowed, or was completely integrated with the surrounding glands after enhancement. Further, few nutrient vessels and filling defects were manifested. These features can be helpful for differential diagnosis; however, some special adenoses exhibit uneven enhancement and unclear boundaries and can, therefore, easily be misdiagnosed as malignant nodules. Some inflammatory lesions also showed malignant signs after enhancement, including uneven high enhancement, irregular shapes, and unclear boundaries, which likely resulted from irregular infiltration of inflammatory cells into the surrounding tissue.
Real-time SWE can indicate differences in tissue hardness by measuring the propagation velocity of shear waves in tissues and can also qualitatively and quantitatively distinguish benign from malignant lesions. The elastic coefficient, or hardness, of the tissue is closely associated with the biological characteristics of lesions. The stroma of benign tumors, such as breast fibroadenomas, are rich in loose mucopolysaccharides; thus, their hardness is reduced in comparison to malignant tumors like invasive ductal carcinomas, in which the stroma is denser and harder due to fibrous tissue components [21]. In general, tumors are softer if they have a higher proportion of parenchyma than stroma. Tumor tissues can also soften when undergoing necrosis and harden when calcareous deposition or bone formation occurs. Real-time SWE is a relatively easy, noninvasive, and objective method to evaluate tissue hardness, and this study, as well as others, has demonstrated that it can effectively be used to differentiate benign from malignant breast tumors.
Consistent with our previous study, we found that SWE had a high diagnostic sensitivity and specificity for BI-RADS category 4 nodules and small breast tumors. These findings are also consistent with other studies, both based in China and elsewhere [22,23]. Many studies have shown that tissue density information obtained by SWE can predict the degree of vascular infiltration, which is a predictor of lymph node metastasis [24,25]. Tumor cells can infiltrate the surrounding stroma and cause changes to connective tissue, increasing collagen cross-linking and the corresponding density of surrounding tissues. Using color gradients, SWE can demonstrate the hardness of areas surrounding lesions, with high-density areas displaying "hard ring signs" [26]. Unfortunately, this technique can also lead to misdiagnosis, as some malignant lesions undergo liquefactive necrosis, which results in decreased SWV values and false-negative diagnoses. The SWV values for intraductal carcinomas with low tumor cell atypia, reduced proliferation of fibrous tissue, and no obvious infiltration into surrounding tissues have also been shown to be below the cut-off value. Some sclerosing benign breast lesions may also lead to an increase in the elastic modulus of tissues, resulting in false-positive diagnoses.
In our study, the combined use of CEUS, SWE, and BI-RADS led to adjustments in the BI-RADS classification for some breast nodules, with improvements in the diagnostic accuracies for benign and malignant breast nodules. Sclerosing adenoses were most commonly identified to be false positives, with some cases showing irregular morphologies due to interstitial fiber hyperplasia that was sometimes mixed with inflammation. In these instances, the SWV was higher than the cut-off value and CEUS showed uneven regions of high contrast enhancement. The malignant features of small breast carcinomas identified as false negatives were not as obvious on two-dimensional ultrasonography, showing only large lobulation, homogeneous echo patterns, and no changes or enhancement in posterior features. The ultrasound images resembled those of benign tumors, with SWE values lower than the cut-off value and CEUS showing only uniformly low contrast enhancement. Therefore, it is necessary to comprehensively analyze many images for nodules with these features, with close follow-up recommended. In addition, patients who are less than 60 years of age with a higher risk of malignancy due to family history should be followed closely, and further puncture biopsies should be carried out if necessary.
The BI-RADS classification system is not without flaws, especially in patients with category 4 nodules (with a malignant probability greater than 2% but less than 95%) [27]. In the United States, most BI-RADS category 4 lesions (69%–95%) undergo puncture biopsies, even though only 22% to 33% of these lesions are malignant. A meta-analysis of studies from Europe and the United States showed that the overdiagnosis rate of breast cancer lesions using molybdenum target mammography was 52% with the current BI-RADS classification system [26]. In the present study, 120 nodules were initially evaluated as BI-RADS category 4A, with 67 cases adjusted to BI-RADS category 3 after consideration of the CEUS and SWE results. This finding indicates that 55.8% (67/120) of BI-RADS 4A nodules would have undergone unnecessary puncture biopsies after conventional ultrasound if not combined with the CEUS and SWE results.
The current study had several limitations. First, as a retrospective study, the potential bias in the selection of the participants could not be eliminated. Second, the sample size was small; therefore, studies on larger datasets with long-term follow-up are required.