The value of multimodal ultrasound imaging in differentiating HER2-positive breast cancer

Objectives: This study investigated vascular differences between human epidermal growth factor receptor 2 (HER2)-positive and HER2-negative breast cancers using multimodal ultrasound and evaluated their diagnostic utility.

Methods: Breast masses were examined with conventional ultrasound (CUS), color Doppler flow imaging (CDFI), contrast-enhanced ultrasound (CEUS), and super-resolution ultrasound (SRUS). Quantitative and qualitative parameters were extracted. Binary logistic regression was used to identify independent predictors of HER2-positive breast cancer. The diagnostic performance was evaluated by receiver operating characteristic (ROC) curve.

Results: 72 breast masses were divided into HER2-positive (n = 23) and HER2-negative (n = 49) groups according to immunohistochemical results. Univariate analysis showed that 12 parameters of four ultrasound modalities were significantly different between the groups (p < 0.05), whereas multivariate analysis showed differences in 8 parameters. Maximum diameter of the lump, microcalcification, peak intensity and max flow velocity of HER2-positive group were significantly greater than HER2-negative group (p < 0.05). HER2-positive group showed greater lesion extent after CEUS, less mean tortuosity, more uniform microvasculature distribution and microvasculature flow direction was away from the transducer (p < 0.05). The combination of CUS, CEUS, and SRUS achieved the highest area under the curve (AUC) of 0.976, with sensitivity, specificity and accuracy of 95.7%, 93.9% and 94.4%, respectively.

Conclusions: Multimodal ultrasound parameters from CUS, CEUS, and SRUS can predict HER2-positive breast cancer, and their combination provides superior diagnostic efficacy.

1 Introduction

Breast cancer is the leading cause of cancer death among women worldwide, with approximately 2.3 million new cases in 2022, accounting for 25% of all female cancers (1). Different molecular subtypes of breast cancer vary in malignancy and treatment, making early identification of molecular subtypes critical for patient management (2). The human epidermal growth factor receptor 2 (HER2) is a tyrosine kinase receptor (3). When the HER2 gene is amplified and its related kinase receptor protein is overexpressed, the cell proliferation regulation process is dysregulated, thereby promoting cancer cell proliferation and invasion, and then leading to the occurrence of malignant tumors. Compared to HER2-negative breast cancer, HER2-positive breast cancer has the characteristics of fast growth, strong invasiveness, easy visceral metastasis, and poor prognosis, accounting for 25% of all breast cancer cases (4, 5). Therefore, early identification and treatment of HER2-positive breast cancer is essential to improve patient prognosis.

Angiogenesis is a critical hallmark of breast cancer invasiveness and plays a central role in tumor growth and metastasis (6). Studies have shown that the generation of tumor angiogenesis has been considered an early sign of cancer, and it is closely related to the expression of vascular endothelial growth factor (VEGF) (7). Previous studies have shown that HER2 can upregulate VEGF, promoting neovascularization within tumor tissues (8). Related research results show that there are marked differences in neovascular among breast cancers with different HER2 expression levels (9). Therefore, early noninvasive evaluation of tumor microvasculature has significant clinical value for HER2 detection and prognosis assessment.

Common imaging techniques, including magnetic resonance imaging (MRI) and contrast-enhanced ultrasound (CEUS), provide information on breast tumor vascularity. MRI offers high sensitivity but limited spatial resolution, restricting microvasculature assessment (10, 11). Conventional ultrasound (CUS), with advantages of real-time imaging, radiation-free operation, accurate localization, and simplicity, is widely used for breast screening in Chinese women (12). CUS provides morphological information on breast masses, including size, shape, margin, and internal echotexture. Color Doppler flow imaging (CDFI) evaluates tissue perfusion, visualizing large vessels and rapid flow within and around the mass (13–16). CEUS uses microbubble contrast agents to real-time dynamic display of tumor blood perfusion, including small, low-flow vessels (17–19). Super-resolution ultrasound (SRUS) breaks through the acoustic diffraction limit, enabling visualization of microvasculature with spatial resolutions on the order of tens of micrometers. This technique enables noninvasive visualization of the tumor microvasculature network and provides detailed information on both its morphology and hemodynamic characteristics (20–22). Meanwhile, SRUS has been validated in preclinical tumour models and clinical pilot studies (23), and our research group has previously confirmed that it can clearly display the microvasculature system in human breast and thyroid lesions (24, 25). However, further research is needed in areas such as molecular typing of breast cancer.

In this study, we used CUS, CDFI, CEUS, and SRUS to extract morphological and vascular features of breast masses. We assessed the correlation between multimodal ultrasound parameters and HER2-positive breast cancer and developed a predictive model to aid individualized treatment and improve breast cancer management.

2 Method

2.1 Patient population

We conducted a prospective, dual-center study including 105 patients with suspected malignant breast tumors from China Resources & Wisco General Hospital and Hubei Cancer Hospital between January and December 2023. All patients underwent multimodal ultrasound examinations, including CUS, CDFI, and CEUS. Inclusion criteria: (1) surgical resection with pathologically confirmed malignancy; (2) comprehensive immunohistochemistry (IHC) results clearly distinguishing HER2-positive and HER2-negative cases; (3) surgery performed within one month after ultrasound; (4) no prior neoadjuvant chemotherapy or other treatments before ultrasonography. Exclusion criteria: (1) previous therapy; (2) contraindications for SonoVue; (3) incomplete clinical data; (4) pregnancy; (5) refusal to participate. The required sample size was calculated a priori using PASS 2025 Efficacy Analysis and Sample Size Software (Version 2025). A two-tailed, two-sample t-test (unequal variances) was employed with a significance level of 0.05 and a test power of 0.80. Based on the anticipated effect size, a minimum of 50 patients is required. Considering a potential 20% dropout rate, at least 63 patients should be enrolled. IHC results of post-surgical breast masses served as the gold standard. This study was approved by the Medical Ethics Committee of hospitals, and all participants provided written informed consent. Figure 1 illustrates patient registration and data processing.

Figure 1

Figure 1. Patient enrollment and data processing pipeline.

2.2 HER2-positive breast cancer diagnosis

HER2 expression was assessed using the EnVision™ IHC method and classified as four different levels (-, +, ++, and +++). HER2 (-) and HER2 (+) are classified as HER2-negative, while HER2 (+++) is classified as HER2-positive. HER2 (++) is considered a borderline result. In this case, fluorescence in situ hybridization (FISH) is performed to check for amplification of HER2. If HER2 amplification is not found, HER2 (++) is classified as HER2-negative. If HER2 amplification occurs, HER2 (++) is classified as HER2-positive. In this study, three cases initially showed IHC 2+ results. Further FISH testing revealed no HER2 amplification in any of these cases, leading to their reclassification as HER2-negative.

2.3 Ultrasonic data acquisition

All ultrasounds were performed by physicians with over 20 years of experience using Mindray Resona R9 with an L11-3U linear probe (3–10 MHz). Patients were supine with arms raised. CUS was performed to assess the gray-scale characteristics and maximum diameter of the lesion. CDFI was used to evaluate vascularity, and the plane showing the richest blood flow, including both the lesion and surrounding tissue, was selected. CEUS was performed using a mechanical index of 0.08. Subsequently, 0.5 mL of SonoVue (Bracco, Milan, Italy) microbubble contrast agent was injected via the cubital peripheral vein, immediately followed by a 5 mL saline flush. More than 1000 frames of ultrasound contrast images were collected at a frame rate of approximately 80 Hz for further processing and analysis by ultrasound super-resolution imaging. Real-time dual-mode images (CUS and CEUS) were used to guide the imaging plane and monitor the microbubble signal after injection. The imaging plane remained unchanged to ensure that all images were acquired in the same area. The remaining 4.3 mL of SonoVue microbubble contrast agent was injected through the cubital peripheral vein and immediately flushed with 5 mL of normal saline. After the injection, the timer and dynamic storage function of the ultrasound instrument were turned on to collect conventional ultrasound contrast images for subsequent analysis.

2.4 Ultrasonic multimodal parameter extraction

Maximum diameter of the lump, boundary, morphology, internal echo, orientation, microcalcification, and posterior echo attenuation were obtained from greyscale images. CDFI images were independently graded by a radiologist using the Adler grading system (grade 0: no detectable flow; grade 1: 1–2 punctate signals; grade 2: 3–4 punctate signals or a short vessel-like flow; grade 3: ≥5 punctate signals or ≥2 vessel-like flows). CEUS data were analyzed using the system software, with regions of interest (ROIs) placed in the most enhanced lesion area and corresponding normal tissue to generate Time–intensity curves (TICs) parameters including time to peak, peak intensity (PI), wash-in rate, wash-out rate and area under the curve (AUC). Each CEUS quantitative parameter was measured three times using the system software, and the average value was used for subsequent analysis. Qualitative CEUS features were assessed based on standardized visual evaluation and include: enhancement time (rapid, simultaneous, or delayed), enhancement intensity (hypoenhancement, isoenhancement, or hyperenhancement), enhancement pattern (centripetal, centrifugal, or diffuse), contrast agent distribution (homogeneous or heterogeneous), enhancement margin (clear or blurred), filling defects (presence or absence of non-enhanced areas), peripheral vessels penetrating the lesion (present or absent), and area of enhancement (unchanged or larger). High-frame-rate CEUS data were exported from the ultrasound device in DICOM format and processed offline using MATLAB (MathWorks Inc., Natick, MA, USA) for super-resolution ultrasound (SRUS) analysisreconstruction (26). SRUS required singular value decomposition (SVD) to separate tissue and background signals from microbubble (MB) signals. Following SVD filtering, a pixel intensity threshold is applied to isolate MB signals, and the area, intensity, and shape/eccentricity of each MB are measured to eliminate artifacts. The spatial coordinates of individual MBs are determined using an intensity-weighted centroid method, which locates each MB signal by calculating the center of its intensity distribution. Localization results from all frames are aggregated to reconstruct a super-resolution microvasculature map. Microvasculature density (MVD) is calculated as the total tracked MB area divided by the ROI. The ROI is manually delineated in MATLAB to align with lesions visible on CUS images and corresponding super-resolution velocity maps. Microvasculature flow velocity is obtained through continuous inter-frame MB tracking. For each MB in frame F, the best-matched MB in frame F + 1 is identified within a 700 μm search window. At an 80 Hz frame rate, velocities up to 40 mm/s can be tracked. MB pairs are accepted if the maximum normalized cross-correlation between frames exceeds a threshold of 0.9; pairs outside the search window or failing the threshold are excluded. Using this method, multiple quantitative parameters—including mean and max flow velocity, mean and max tortuosity, mean and largest diameter, fractal dimension—are calculated by tracking microbubbles (MBs). Results within ROIs are averaged to comprehensively characterize the tumor microvasculature system. In addition to quantitative analysis, qualitative SRUS parameters, including microvasculature distribution and microvasculature flow direction, were assessed based on the reconstructed super-resolution microvasculature maps. The Adler grade and all qualitative parameters were assessed independently by two doctors. In case of disagreement, a third doctor was consulted for assessment. Detailed definitions and calculation methods of CEUS and SRUS parameters are summarized in Table 1, while the interobserver agreement results are presented in Table 2.

Table 1

Table 1. Definitions and measurement methods of CEUS and SRUS parameters.

Table 2

Table 2. The measures of interobserver agreement between the two observers.

2.5 Statistical analysis

SPSS 27.0 software was used for statistical analysis of all data. Interobserver agreement for qualitative parameters was assessed using Cohen’s kappa statistics. Continuous variables were tested for normality and homogeneity of variance. Two-sample t-tests were applied for normally distributed data with equal variance; Mann–Whitney U tests for non-normal or unequal variance data. Categorical variables were compared using chi-square or Fisher’s exact tests. Logistic regression was used to select independent variables, and significant parameters were incorporated to build a HER2-positive prediction model. IHC results served as the gold standard to calculate sensitivity and specificity, and the receiver operating characteristic curve was drawn. The receiver operating characteristic (ROC) curve and AUC were used to evaluate diagnostic performance of CUS, CDFI, CEUS, SRUS, and their combinations. Statistical significance was defined as p < 0.05, with significance levels noted at 0.05 (*) and 0.01 (**) (two-tailed).

3 Result

3.1 Ultrasound images of breast masses

We ultimately included 72 patients, comprising 23 HER2-positive patients (mean age, 52.04 ± 8.93 years) and 49 HER2-negative patients (mean age, 53.55 ± 9.16 years). Figure 2 showed CUS, CDFI, CEUS, and SRUS images of a patient with HER2-positive and HER2-negative breast masses, respectively.

Figure 2

Figure 2. Representative multimodal ultrasound images of breast masses with HER2-positive and HER2- negative. (A–D) correspond to a HER2-positive case: (A) conventional ultrasound image; (B) color Doppler flow image; (C) contrast-enhanced ultrasound image; (D) super-resolution ultrasound image. (E–H) correspond to a HER2-negative case: (E) conventional ultrasound image; (F) color Doppler flow image; (G) contrast-enhanced ultrasound image; (H) super-resolution ultrasound image.

3.2 Comparison of clinicopathological characteristics, CUS and CDFI parameters between HER2-positive and HER2-negative breast cancers

Significant differences between HER2-positive and HER2-negative breast cancers were observed in maximum diameter of the lump, orientation, and microcalcification. HER2-positive breast cancers exhibited a significantly larger maximum diameter compared to HER2-negative tumors (3.78 ± 1.38 cm vs. 3.07 ± 1.34 cm, p = 0.043). In terms of orientation, a higher proportion of HER2-positive breast cancer showed a parallel growth pattern relative to the skin (87.0% vs. 59.2%, p = 0.018). Additionally, microcalcification was more frequently detected in the HER2-positive group (65.2% vs. 32.7%, p = 0.009). No significant associations were found for patient age, histologic grade, lymph node status, boundary, morphology, internal echo, posterior echo attenuation, or Adler grade. The results are shown in Table 3.

Table 3

Table 3. Clinicopathological characteristics and CUS and CDFI ultrasound parameters in HER2-positive and HER2-negative breast cancers.

3.3 Comparison of CEUS and SRUS parameters between HER2-positive and HER2-negative breast cancers

Differences in qualitative and quantitative CEUS and SRUS parameters were evaluated between HER2-positive and HER2-negative breast cancers (Tables 4, 5). The PI value was significantly higher in the HER2-positive group (29.47 ± 1.51 dB vs. 17.91 ± 1.14 dB, p < 0.001; Figures 3A–C). In terms of CEUS qualitative analysis, area of enhancement showed distinct differences as most HER2-positive group exhibited an enlarged enhancement area, whereas the majority of HER2-negative group demonstrated remained unchanged in size following enhancement (Figures 3D–F). Compared to the HER2-negative group, the HER2-positive group exhibited significantly higher max (20.0 cm/s [19.2-34.1 cm/s] vs. 18.0 cm/s [16.8-19.9 cm/s], p = 0.001) and mean flow velocity (6.0 cm/s [4.1-6.9 cm/s] vs. 4.1 cm/s [3.5-5.5 cm/s], p = 0.018) (Figures 4A–E). Regarding microvasculature distribution, peripheral distribution was predominant in both groups, but it was more frequently observed in HER2-negative breast cancer (Figures 4F–J). Conversely, vascular tortuosity was significantly lower in HER2-positive tumors, indicating straighter vessels (Figure 5).

Table 4

Table 4. Comparison of CEUS parameters between HER2-positive and HER2-negative breast cancers.

Table 5

Table 5. Comparison of SRUS parameters between HER2-positive and HER2-negative breast cancers.

Figure 3

Figure 3. Representative images of peak intensity and area of enhancement in HER2-positive and HER2-negative breast masses. (A, D) show contrast-enhanced ultrasound images of HER2-positive breast masses; (B, E) show contrast-enhanced ultrasound images of HER2-negative breast masses. The violin plot (C) compares peak intensity between HER2-positive and HER2-negative groups. Corresponding bar graph (F) illustrates the quantitative comparison of area of enhancement between the two groups. * indicates P-value < 0.05, **** indicates P-value < 0.0001.

Figure 4

Figure 4. Representative images of microvasculature flow velocity and microvasculture distribution in HER2-positive and HER2-negative breast masses. (A, B) are conventional ultrasound image and microvasculature flow velocity image of a HER2-positive breast mass. (C, D) are conventional ultrasound image and microvasculature flow velocity image of a HER2-negative breast mass. The violin plot (E) illustrates the comparison of max flow velocity between HER2-positive and HER2-negative groups. (F, G) are conventional ultrasound image and microvasculature distribution image of a HER2-positive breast mass. (H, I) are conventional ultrasound image and microvasculature distribution image of a HER2-negative breast mass. The bar graph (J) illustrates the comparison of microvasculature distribution between HER2-positive and HER2-negative groups. **p < 0.01; ****p < 0.0001.

Figure 5

Figure 5. Representative images of microvasculature tortuosity in HER2 positive and HER2 negative breast masses. (A, B) are conventional ultrasound image and microvasculature tortuosity image of HER2-positive breast mass. (C, D) are conventional ultrasound image and microvasculature tortuosity image of HER2-negative breast mass. (E) is the corresponding violin plot. *p < 0.05.

3.4 Binary logistic regression analysis of CUS, CEUS, and SRUS parameters with HER2-positive and HER2-negative breast cancers

Binary logistic regression was conducted to evaluate the predictive value of ultrasound parameters for HER2-positive breast cancer. Maximum diameter of the lump, microcalcification, PI, area of enhancement, mean tortuosity, maximum flow velocity, microvasculature distribution, and microvasculature direction were eight risk factors for ultrasound prediction of HER2-positive breast cancer, with microvasculature flow direction exhibiting the highest odds ratio (Table 6).

Table 6

Table 6. Binary logistic regression results of ultrasound parameters for differentiating HER2-positive and HER2- negative breast cancers.

3.5 Diagnostic performance of CUS, CEUS, and SRUS alone and in combination

ROC analysis evaluated the diagnostic performance of seven imaging approaches in distinguishing HER2-positive from HER2-negative breast cancers, including CUS, CEUS, SRUS, CUS+CEUS, CUS+SRUS, CEUS+SRUS, and CUS+CEUS+SRUS. Among the individual modalities, CEUS showed the highest specificity (95.9%), accuracy (90.3%), and AUC (0.914), whereas SRUS achieved the highest sensitivity (95.7%) but a comparatively lower AUC of 0.874. The three-modality combination (CUS + CEUS + SRUS) yielded an AUC of 0.976, with a sensitivity of 95.7% and specificity of 93.9%, outperforming dual-modality combinations (Table 7). Figure 6 shows the ROC curves.

Table 7

Table 7. The diagnostic efficacy of CUS, CEUS, SRUS and their combination in differentiating HER2-negative and HER2-positive breast cancers.

Figure 6

Figure 6. The receiver operating characteristic curves for diagnosis of HER2-positive and HER2-negative breast cancers by conventional ultrasound, enhanced ultrasound, super-resolution ultrasound imaging alone and in combination.

4 Discussion

In this study, we extracted feature parameters from multiple ultrasound modalities and analyzed their differences and diagnostic performance in distinguishing HER2-positive from HER2-negative breast cancers. Binary logistic regression analysis revealed that maximum diameter of the lump, microcalcification, PI, area of enhancement, max flow velocity, mean tortuosity, microvasculature distribution, and microvasculature flow direction were significantly associated with HER2 status, with microvasculature flow direction showing the strongest association. In addition, the ROC analysis demonstrated that the combination of CUS, CEUS, and SRUS achieved the highest AUC compared to individual or dual-modality approaches, providing valuable information for early identification of HER2-positive breast cancer and aiding clinical decision-making.

Compared with HER2-negative breast cancer, HER2-positive breast cancer has higher malignancy, more aggressive tumor growth, increased risk of metastasis and recurrence, and worse prognosis (27). Currently HER2-targeted therapies, including monoclonal antibodies and tyrosine kinase inhibitors, have been shown to significantly improve patient outcomes (28, 29). CUS is a primary imaging modality for breast cancer, and previous studies have demonstrated its potential in predicting the molecular subtype and HER2 status (30, 31). In this study, HER2-positive tumors were larger, with microcalcifications and parallel growth, consistent with prior findings. Previous studies have shown that mass size and microcalcification were key predictors of HER2 status (32, 33). HER2 promotes tumor cell proliferation, causing local ischemia that induces cell death and microcalcification. Related studies have shown that HER2-positive tumors alter immune cell activity, increasing BMP-2 levels and contributing to calcification (34, 35). Consequently, HER2-positive breast cancer typically exhibits a larger size, with microcalcifications and parallel growth.

Breast cancer is an angiogenesis-dependent tumor with high heterogeneity, and its molecular subtypes exhibit distinct vascular patterns. Related studies have shown that HER2 overexpression promotes angiogenesis by upregulating VEGF and stimulating endothelial cell proliferation, resulting in higher MVD in HER2-positive tumors (36, 37). In this study, CDFI showed no significant difference in adler grade between HER2-positive and HER2-negative groups, differing from Zhu et al. and Wang et al., who reported richer blood flow in HER2-positive tumors (37, 38). The discrepancy may be attributed to differences in grouping strategies. Zhu et al. classified breast cancer into four molecular subtypes, Wang et al. divided adler grade into two groups, while this study divided breast cancer into HER2-positive and HER2-negative groups, and divided adler grade into four groups. Differences in grouping strategies may partially account for the inconsistencies in findings. Further multicenter studies with larger cohorts are needed to clarify the relationship between CDFI features and HER2 status.

CEUS is widely used and safe diagnostic tool in breast cancer, allowing continuous visualization of blood flow and microcirculation, thereby improving lesion detection and diagnostic accuracy (18, 39). It has also been applied to assess HER2 expression (30, 40). In this study, PI derived from the TIC analysis on CEUS was significantly higher in HER2-positive tumors, which also exhibited faster wash-in and wash-out perfusion and greater lesion extent following contrast enhancement. The results of this study are consistent with the results of previous studies. The study of Li et al. showed that HER2-positive breast cancer had a larger PI (40). Wang et al. proved that HER2-positive breast cancers had greater AS and the overranging phenomenon (41). Li et al. demonstrated that the HER2 overexpression subtype showed rapid enhancement (95.65%) and larger lesion size after enhancement (100%), and higher PI compared to the Luminal A and Luminal B subtypes (42). Park et al. performed low-dose perfusion computed tomography (CT) in 70 breast cancer patients and found that HER2-overexpressing breast cancers had significantly higher perfusion and PI (43). This phenomenon may be related to VEGF-mediated angiogenic patterns reported in previous studies (44–47), although direct molecular validation was not performed in the present study. Such vascular characteristics could contribute to faster perfusion and changes in tumor blood flow. In addition, PI rises with increasing tumor neovascularization, and high-perfusion breast cancers tend to show greater lesion extent after CEUS (47, 48).

SRUS overcomes the ultrasound diffraction limit to visualize tumor microvasculature, enabling quantitative and qualitative assessment of morphological and functional parameters of microcirculation (49). Our previous studies have shown that SRUS can provide valuable microvasculature information for distinguishing between benign and malignant breast masses (25, 50), but its role in predicting molecular subtypes remains unexplored. Zheng et al. proposed a protocol to evaluate the relationship between SRUS parameters and histologic biomarkers in breast cancer, but the study is still in progress (51). This study demonstrated faster microvasculature flow velocity in HER2-positive breast cancer. And qualitative analysis showed that HER2-positive breast cancers had more uniform microvasculature distribution, smaller tortuosity, and flow directed away from the transducer. These findings are supported by Park et al., who reported increased perfusion function and blood volume permeability in HER2-positive tumors using low-dose perfusion CT, suggesting faster flow and uniform distribution (52). Kim et al. evaluated the correlation between tumor vascularity and prognostic biomarkers in breast cancer patients using CT and MRI. The results showed that HER2 expression was significantly associated with blood flow perfusion, indicating increased angiogenesis and higher flow rates (53). A study using incoherent motion diffusion-weighted imaging showed that HER2-positive breast cancer had higher blood flow velocity and capillary number (54). Mean vessel tortuosity is a method to quantify uncontrolled vasculature, measuring the abnormal twisting of vessels and indicating the complexity of vascular pathways (55). Previous studies have shown that tumor neovascular tortuosity can block blood supply (56). This may further affect blood flow velocity. The result of this study suggested that HER2-positive breast cancers exhibited faster blood flow, suggesting lower intratumoral tortuosity. Previous studies have shown that CDFI can detect blood flow direction (57–59), but it cannot resolve microvessels with small diameter or low velocity. In this study, we used SRUS to track the direction of microbubble movement and found that microvasculature flow in HER2-positive breast cancer was away from the transducer. These results provided more information on microcirculation for noninvasive prediction of HER2-positive breast cancer, and provided a basis for clinicians to make treatment decisions in advance.

HER2 expression has been reported to positively correlate with VEGF expression and tumor angiogenesis, thereby promoting neovascular formation and increased vascular permeability in breast cancer (34, 43, 60, 61). In this study, we evaluated the value of multimodal ultrasound in predicting HER2-positive breast cancer. As a pure blood pool imaging technique, CEUS enables real-time monitoring of the entire tissue perfusion process (62, 63). SRUS overcomes the ultrasound diffraction limit to display the morphology and structure of microvessels in tumors (64). The results of the study showed that the combination of CUS, CEUS, and SRUS achieved the highest diagnostic performance, outperforming individual or dual-modality methods. The combination of CEUS and SRUS enabled visualization of millimeter and micron level tumor vessels, integrating CUS features and enhancing diagnostic efficacy for predicting HER2-positive breast cancer, thereby providing a robust basis for clinical decision-making.

However, this study has several limitations. First, it focused solely on differences in multimodal ultrasound parameters between HER2-positive and HER2-negative breast cancers. Future research could incorporate four molecular subtypes to further elucidate the imaging characteristics specific to different molecular subtypes. Second, the sample size of this study is small, and the findings require validation in larger cohorts. Third, HER2 status was determined by immunohistochemistry using the EnVision™ system, with equivocal cases confirmed by FISH. However, variability in IHC sensitivity related to differences in antibody clones, detection systems, and staining protocols may influence classification, particularly in cases with low HER2 expression. Fourth, the blood flow parameters extracted by vascular imaging are relatively limited. In future studies, more high-throughput parameters can be extracted from dynamic CEUS video and 3D ultrasound SRUS through radiomics.

5 Conclusion

In conclusion, the multimodal ultrasound model (CUS, CEUS, and SRUS) has shown good predictive efficacy in distinguishing HER2-positive and HER2-negative breast cancer. As a non-invasive method, it provides a scientific basis for treatment planning and is crucial for improving patient prognosis.

Data availability statement

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

Ethics statement

The studies involving humans were approved by Institutional Review Board of China Resources and Wisco General Hospital. The studies were conducted in accordance with the local legislation and institutional requirements. The participants provided their written informed consent to participate in this study.

Author contributions

Y-ML: Data curation, Methodology, Writing – original draft, Writing – review & editing. CLiu: Data curation, Formal analysis, Writing – original draft, Writing – review & editing. B-SL: Data curation, Writing – original draft, Writing – review & editing. CLiao: Data curation, Visualization, Writing – review & editing. NZ: Data curation, Writing – review & editing. QW: Data curation, Writing – review & editing. S-EZ: Data curation, Funding acquisition, Validation, Writing – review & editing. GZ: Validation, Writing – review & editing. H-RY: Data curation, Funding acquisition, Resources, Validation, Writing – review & editing.

Funding

The author(s) declared that financial support was received for this work and/or its publication. This study was supported by the Scientific Research Projects of Wuhan Municipal Health Commission (Nos. WX23B16, WX23Q39, WX23Q41), the Hubei Province Key Laboratory of Occupational Hazards Identification and Control, Wuhan University of Science and Technology (No. JF2023-K02), national key clinical specialty construction discipline (HBCHBCC-D06).

Conflict of interest

The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Abbreviations

CDFI, Color Doppler flow imaging; CEUS, Contrast-enhanced ultrasound; CT, Computed tomography; CUS, Conventional ultrasound; ER, Estrogen receptor; HER2, Human epidermal growth factor receptor 2; IHC, Immunohistochemistry; MRI, Magnetic resonance imaging; MVD, Microvasculature density; PI, Peak intensity; ROC, Receiver operating characteristic; ROI, Region of interest; SRUS, Super-resolution ultrasound; TIC, Time–intensity curve; VEGF, Vascular endothelial growth factor.

References

1. Bray F, Laversanne M, Sung H, Ferlay J, Siegel RL, Soerjomataram I, et al. Global cancer statistics 2022: globocan estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA: Cancer J Clin. (2024) 74:229–63. doi: 10.3322/caac.21834

PubMed Abstract | Crossref Full Text | Google Scholar

2. Testa U, Castelli G, and Pelosi E. Breast cancer: A molecularly heterogenous disease needing subtype-specific treatments. Med Sci (Basel Switzerland). (2020) 8:18. doi: 10.3390/medsci8010018

PubMed Abstract | Crossref Full Text | Google Scholar

3. Le Du F, Diéras V, and Curigliano G. The role of tyrosine kinase inhibitors in the treatment of her2+ Metastatic breast cancer. Eur J Cancer (Oxford England: 1990). (2021) 154:175–89. doi: 10.1016/j.ejca.2021.06.026

PubMed Abstract | Crossref Full Text | Google Scholar

4. Uslu H, Önal T, Tosun M, Arslan AS, Ciftci E, and Utkan NZ. Intravoxel incoherent motion magnetic resonance imaging for breast cancer: A comparison with molecular subtypes and histological grades. Magnetic Resonance Imaging. (2021) 78:35–41. doi: 10.1016/j.mri.2021.02.005

PubMed Abstract | Crossref Full Text | Google Scholar

5. Schlam I and Swain SM. Her2-positive breast cancer and tyrosine kinase inhibitors: the time is now. NPJ Breast Cancer. (2021) 7:56. doi: 10.1038/s41523-021-00265-1

PubMed Abstract | Crossref Full Text | Google Scholar

6. Teleanu RI, Chircov C, Grumezescu AM, and Teleanu DM. Tumor angiogenesis and anti-angiogenic strategies for cancer treatment. J Clin Med. (2019) 9:84. doi: 10.3390/jcm9010084

PubMed Abstract | Crossref Full Text | Google Scholar

7. Elebiyo TC, Rotimi D, Evbuomwan IO, Maimako RF, Iyobhebhe M, Ojo OA, et al. Reassessing vascular endothelial growth factor (Vegf) in anti-angiogenic cancer therapy. Cancer Treat Res Commun. (2022) 32:100620. doi: 10.1016/j.ctarc.2022.100620

PubMed Abstract | Crossref Full Text | Google Scholar

8. Wu M, Chen L, Qi Y, Ci H, Mou S, Yang J, et al. Human umbilical cord mesenchymal stem cell promotes angiogenesis via integrin β1/erk1/2/hif-1α/vegf-a signaling pathway for off-the-shelf breast tissue engineering. Stem Cell Res Ther. (2022) 13:99. doi: 10.1186/s13287-022-02770-x

PubMed Abstract | Crossref Full Text | Google Scholar

9. Chen S, Cui W, Dong J, Chen W, Dong H, and Zhao R. A novel signature based on angiogenesis-related genes predicts the prognosis and immunotherapy response in her2-positive breast cancer. J Cancer. (2024) 15:4731–48. doi: 10.7150/jca.94120

PubMed Abstract | Crossref Full Text | Google Scholar

10. Lo G, McLaughlin A, Jacques A, Dhillon R, Porter G, Jayaratne T, et al. Does medicare-eligible high-risk breast cancer screening mri target the right women? J Med Imaging Radiat Oncol. (2020) 64:220–8. doi: 10.1111/1754-9485.13009

PubMed Abstract | Crossref Full Text | Google Scholar

11. Iranmakani S, Mortezazadeh T, Sajadian F, Ghaziani MF, Ghafari A, Khezerloo D, et al. A review of various modalities in breast imaging: technical aspects and clinical outcomes. Egyptian J Radiol Nucl Med. (2020) 51:57. doi: 10.1186/s43055-020-00175-5

Crossref Full Text | Google Scholar

12. Zhang X, Li H, Wang C, Cheng W, Zhu Y, Li D, et al. Evaluating the accuracy of breast cancer and molecular subtype diagnosis by ultrasound image deep learning model. Front Oncol. (2021) 11:623506. doi: 10.3389/fonc.2021.623506

PubMed Abstract | Crossref Full Text | Google Scholar

13. Luo R, Zhang Y, Jiang W, Wang Y, and Luo Y. Value of micro-flow imaging and high-definition micro-flow imaging in differentiating Malignant and benign breast lesions. Clin Radiol. (2024) 79:e48–56. doi: 10.1016/j.crad.2023.10.007

PubMed Abstract | Crossref Full Text | Google Scholar

14. Liu W, Wang D, Liu L, and Zhou Z. Assessing the influence of B-us, cdfi, se, and patient age on predicting molecular subtypes in breast lesions using deep learning algorithms. J Ultrasound Med: Off J Am Institute Ultrasound Med. (2024) 43:1375–88. doi: 10.1002/jum.16460

PubMed Abstract | Crossref Full Text | Google Scholar

15. Zhu JY, He HL, Jiang XC, Bao HW, and Chen F. Multimodal ultrasound features of breast cancers: correlation with molecular subtypes. BMC Med Imaging. (2023) 23:57. doi: 10.1186/s12880-023-00999-3

PubMed Abstract | Crossref Full Text | Google Scholar

16. Huang C, Zhang W, Gong P, Lok UW, Tang S, Yin T, et al. Super-resolution ultrasound localization microscopy based on a high frame-rate clinical ultrasound scanner: an in-human feasibility study. Phys Med Biol. (2021) 66:10.1088/1361-6560/abef45. doi: 10.1088/1361-6560/abef45

PubMed Abstract | Crossref Full Text | Google Scholar

17. Sun J, Shen X, Zhang N, Zhang Q, Xing K, and Liu Y. Combination of conventional ultrasound with quantitative and qualitative analyses of ceus for the differentiation of benign and Malignant breast solid lesions: A modified breast cancer model. Asian J Surg. (2024) S1015-9584(24)01844-X. doi: 10.1016/j.asjsur.2024.08.104

PubMed Abstract | Crossref Full Text | Google Scholar

18. Ito T, Manabe H, Kubota M, and Komoike Y. Current status and future perspectives of contrast-enhanced ultrasound diagnosis of breast lesions. J Med Ultrasonics (2001). (2024) 51:611–25. doi: 10.1007/s10396-024-01486-0

PubMed Abstract | Crossref Full Text | Google Scholar

19. Brloznik M, Boc N, Cemazar M, Bosnjak M, Savarin M, Kejzar N, et al. Contrast-enhanced ultrasound for evaluation of tumor perfusion and outcome following treatment in a murine melanoma model. Bioelectrochem (Amsterdam Netherlands). (2021) 142:107932. doi: 10.1016/j.bioelechem.2021.107932

PubMed Abstract | Crossref Full Text | Google Scholar

20. van Sloun RJG, Solomon O, Bruce M, Khaing ZZ, Wijkstra H, Eldar YC, et al. Super-resolution ultrasound localization microscopy through deep learning. IEEE Trans Med Imaging. (2021) 40:829–39. doi: 10.1109/tmi.2020.3037790

PubMed Abstract | Crossref Full Text | Google Scholar

21. Chen Q, Song H, Yu J, and Kim K. Current development and applications of super-resolution ultrasound imaging. Sensors (Basel). (2021) 21:2417. doi: 10.3390/s21072417

PubMed Abstract | Crossref Full Text | Google Scholar

22. Christensen-Jeffries K, Couture O, Dayton PA, Eldar YC, Hynynen K, Kiessling F, et al. Super-resolution ultrasound imaging. Ultrasound Med Biol. (2020) 46:865–91. doi: 10.1016/j.ultrasmedbio.2019.11.013

PubMed Abstract | Crossref Full Text | Google Scholar

23. Opacic T, Dencks S, Theek B, Piepenbrock M, Ackermann D, Rix A, et al. Motion model ultrasound localization microscopy for preclinical and clinical multiparametric tumor characterization. Nat Commun. (2018) 9:1527. doi: 10.1038/s41467-018-03973-8

PubMed Abstract | Crossref Full Text | Google Scholar

24. Zhang G, Yu J, Lei YM, Hu JR, Hu HM, Harput S, et al. Ultrasound super-resolution imaging for the differential diagnosis of thyroid nodules: A pilot study. Front Oncol. (2022) 12:978164. doi: 10.3389/fonc.2022.978164

PubMed Abstract | Crossref Full Text | Google Scholar

25. Zhang G, Lei YM, Li N, Yu J, Jiang XY, Yu MH, et al. Ultrasound super-resolution imaging for differential diagnosis of breast masses. Front Oncol. (2022) 12:1049991. doi: 10.3389/fonc.2022.1049991

PubMed Abstract | Crossref Full Text | Google Scholar

26. Silva I and Moody GB. An open-source toolbox for analysing and processing physionet databases in matlab and octave. J Open Res Softw. (2014) 2:e27. doi: 10.5334/jors.bi

PubMed Abstract | Crossref Full Text | Google Scholar

27. McGale J, Khurana S, Huang A, Roa T, Yeh R, Shirini D, et al. Pet/ct and spect/ct imaging of her2-positive breast cancer. J Clin Med. (2023) 12:4882. doi: 10.3390/jcm12154882

PubMed Abstract | Crossref Full Text | Google Scholar

28. Swain SM, Shastry M, and Hamilton E. Targeting her2-positive breast cancer: advances and future directions. Nat Rev Drug Discov. (2023) 22:101–26. doi: 10.1038/s41573-022-00579-0

PubMed Abstract | Crossref Full Text | Google Scholar

29. Hayes DF. Her2 and breast cancer - a phenomenal success story. New Engl J Med. (2019) 381:1284–6. doi: 10.1056/NEJMcibr1909386

PubMed Abstract | Crossref Full Text | Google Scholar

30. Gong X, Li Q, Gu L, Chen C, Liu X, Zhang X, et al. Conventional ultrasound and contrast-enhanced ultrasound radiomics in breast cancer and molecular subtype diagnosis. Front Oncol. (2023) 13:1158736. doi: 10.3389/fonc.2023.1158736

PubMed Abstract | Crossref Full Text | Google Scholar

31. Gumowska M, Mączewska J, Prostko P, Roszkowska-Purska K, and Dobruch-Sobczak K. Is there a correlation between multiparametric assessment in ultrasound and intrinsic subtype of breast cancer? J Clin Med. (2021) 10:5394. doi: 10.3390/jcm10225394

PubMed Abstract | Crossref Full Text | Google Scholar

32. Zhuo X, Lv J, Chen B, Liu J, Luo Y, Liu J, et al. Combining conventional ultrasound and ultrasound elastography to predict her2 status in patients with breast cancer. Front Physiol. (2023) 14:1188502. doi: 10.3389/fphys.2023.1188502

PubMed Abstract | Crossref Full Text | Google Scholar

33. Zhang T, Tan T, Han L, Appelman L, Veltman J, Wessels R, et al. Predicting breast cancer types on and beyond molecular level in a multi-modal fashion. NPJ Breast Cancer. (2023) 9:16. doi: 10.1038/s41523-023-00517-2

PubMed Abstract | Crossref Full Text | Google Scholar

34. Cui H, Sun Y, Zhao D, Zhang X, Kong H, Hu N, et al. Radiogenomic analysis of prediction her2 status in breast cancer by linking ultrasound radiomic feature module with biological functions. J Trans Med. (2023) 21:44. doi: 10.1186/s12967-022-03840-7

PubMed Abstract | Crossref Full Text | Google Scholar

35. Scimeca M, Giannini E, Antonacci C, Pistolese CA, Spagnoli LG, and Bonanno E. Microcalcifications in breast cancer: an active phenomenon mediated by epithelial cells with mesenchymal characteristics. BMC Cancer. (2014) 14:286. doi: 10.1186/1471-2407-14-286

PubMed Abstract | Crossref Full Text | Google Scholar

36. Bulut IN, Kayadibi Y, Deger E, Kurt SA, Velidedeoglu M, Onur I, et al. Preoperative role of superb microvascular imaging and shear-wave elastography for prediction of axillary lymph node metastasis in patients with breast cancer. Ultrasound Q. (2024) 40:111–8. doi: 10.1097/ruq.0000000000000671

PubMed Abstract | Crossref Full Text | Google Scholar

37. Zhao Y, Yang N, Wang X, Huang Y, Zhou X, and Zhang D. Potential roles of lymphovascular space invasion based on tumor characteristics provide important prognostic information in T1 tumors with er and her2 positive breast cancer. Clin Trans Oncol Off Publ Fed Spanish Oncol Societies Natl Cancer Institute Mexico. (2020) 22:2275–85. doi: 10.1007/s12094-020-02369-9

PubMed Abstract | Crossref Full Text | Google Scholar

38. Wang K, Yang X, Yang S, Du X, Shi R, Bai W, et al. A diagnostic test of two-dimensional ultrasonic feature extraction based on artificial intelligence combined with blood flow adler classification and contrast-enhanced ultrasound for predicting her-2-positive breast cancer. Trans Cancer Res. (2025) 14:640–50. doi: 10.21037/tcr-24-2182

PubMed Abstract | Crossref Full Text | Google Scholar

39. Chen Q, Xiao H, Hu L, Huang Y, Cao Z, Shuai X, et al. 19f mri/ceus dual imaging-guided sonodynamic therapy enhances immune checkpoint blockade in triple-negative breast cancer. Adv Sci (Weinheim Baden-Wurttemberg Germany). (2024) 11:e2401182. doi: 10.1002/advs.202401182

PubMed Abstract | Crossref Full Text | Google Scholar

40. Li H, Zhang CT, Shao HG, Pan L, Li Z, Wang M, et al. Prediction models of breast cancer molecular subtypes based on multimodal ultrasound and clinical features. BMC Cancer. (2025) 25:886. doi: 10.1186/s12885-025-14233-6

PubMed Abstract | Crossref Full Text | Google Scholar

41. Wang XY, Hu Q, Fang MY, He Y, Wei HM, Chen XX, et al. The correlation between her-2 expression and the ceus and arfi characteristics of breast cancer. PloS One. (2017) 12:e0178692. doi: 10.1371/journal.pone.0178692

PubMed Abstract | Crossref Full Text | Google Scholar

42. Li X, Zhang J, Zhang G, Liu J, Tang C, Chen K, et al. Contrast-enhanced ultrasound and conventional ultrasound characteristics of breast cancer with different molecular subtypes. Clin Breast Cancer. (2024) 24:204–14. doi: 10.1016/j.clbc.2023.11.005

PubMed Abstract | Crossref Full Text | Google Scholar

43. Park EK, Seo BK, Kwon M, Cho KR, Woo OH, Song SE, et al. Low-dose perfusion computed tomography for breast cancer to quantify tumor vascularity: correlation with prognostic biomarkers. Invest Radiol. (2019) 54:273–81. doi: 10.1097/rli.0000000000000538

PubMed Abstract | Crossref Full Text | Google Scholar

44. Sood N, Maurya R, Gautam S, Patel P, Das Gupta G, and Das Kurmi B. Comprehensive advances in her2-positive and her2-negative breast cancer: unveiling molecular mechanisms and exploring cutting-edge targeted therapies for enhanced patient outcomes. Naunyn-Schmiedeberg’s Arch Pharmacol. (2025) 398:14877–98. doi: 10.1007/s00210-025-04204-w

PubMed Abstract | Crossref Full Text | Google Scholar

45. Wen B, Kong W, Zhang Y, Xue H, Wu M, and Wang F. Association between contrast-enhanced ultrasound characteristics and molecular subtypes of breast cancer. J Ultrasound Med: Off J Am Institute Ultrasound Med. (2022) 41:2019–31. doi: 10.1002/jum.15886

PubMed Abstract | Crossref Full Text | Google Scholar

46. Asgari-Karchekani S, Aryannejad A, Mousavi SA, Shahsavarhaghighi S, and Tavangar SM. The role of her2 alterations in clinicopathological and molecular characteristics of breast cancer and her2-targeted therapies: A comprehensive review. Med Oncol (Northwood London England). (2022) 39:210. doi: 10.1007/s12032-022-01817-6

PubMed Abstract | Crossref Full Text | Google Scholar

47. Leng X, Huang G, Li S, Yao M, Ding J, and Ma F. Correlation of breast cancer microcirculation construction with tumor stem cells (Cscs) and epithelial-mesenchymal transition (Emt) based on contrast-enhanced ultrasound (Ceus). PloS One. (2021) 16:e0261138. doi: 10.1371/journal.pone.0261138

PubMed Abstract | Crossref Full Text | Google Scholar

48. Shin I. “Her2 Signaling in Breast Cancer.” In: Advances in Experimental Medicine and Biology. Singapore: Springer (2021). pp. 53–79. doi: 10.1007/978-981-32-9620-6_3

PubMed Abstract | Crossref Full Text | Google Scholar

49. Xia S, Zheng Y, Hua Q, Wen J, Luo X, Yan J, et al. Super-resolution ultrasound and microvasculomics: A consensus statement. Eur Radiol. (2024) 34:7503–13. doi: 10.1007/s00330-024-10796-3

PubMed Abstract | Crossref Full Text | Google Scholar

50. Lei YM, Liu C, Hu HM, Li N, Zhang N, Wang Q, et al. Combined use of super-resolution ultrasound imaging and shear-wave elastography for differential diagnosis of breast masses. Front Oncol. (2024) 14:1497140. doi: 10.3389/fonc.2024.1497140

PubMed Abstract | Crossref Full Text | Google Scholar

51. Li J, Chen L, Wang R, Zhu J, Li A, Li J, et al. Ultrasound localization microscopy in the diagnosis of breast tumors and prediction of relevant histologic biomarkers associated with prognosis in humans: the protocol for a prospective, multicenter study. BMC Med Imaging. (2025) 25:13. doi: 10.1186/s12880-024-01535-7

PubMed Abstract | Crossref Full Text | Google Scholar

52. Park EK, Lee KS, Seo BK, Cho KR, Woo OH, Son GS, et al. Machine learning approaches to radiogenomics of breast cancer using low-dose perfusion computed tomography: predicting prognostic biomarkers and molecular subtypes. Sci Rep. (2019) 9:17847. doi: 10.1038/s41598-019-54371-z

PubMed Abstract | Crossref Full Text | Google Scholar

53. Kim HY, Bae MS, Seo BK, Lee JY, Cho KR, Woo OH, et al. Comparison of ct- and mri-based quantification of tumor heterogeneity and vascularity for correlations with prognostic biomarkers and survival outcomes: A single-center prospective cohort study. Bioengineering (Basel Switzerland). (2023) 10:504. doi: 10.3390/bioengineering10050504

PubMed Abstract | Crossref Full Text | Google Scholar

54. Ma Y, Shan D, Wei J, and Chen A. Application of intravoxel incoherent motion diffusion-weighted imaging in differential diagnosis and molecular subtype analysis of breast cancer. Am J Trans Res. (2021) 13:3034–43.

PubMed Abstract | Google Scholar

55. Xu Z, Escalona G, Schrack I, Zhang W, Zhai T, Shea LD, et al. Detecting metastatic potential of cancer through longitudinal vasculature imaging of biomaterial scaffold using non-invasive in vivo photoacoustic microscopy and optical coherence tomography. Theranostics. (2025) 15:509–20. doi: 10.7150/thno.101685

PubMed Abstract | Crossref Full Text | Google Scholar

56. Szabó A and Merks RMH. Blood vessel tortuosity selects against evolution of aggressive tumor cells in confined tissue environments: A modeling approach. PloS Comput Biol. (2017) 13:e1005635. doi: 10.1371/journal.pcbi.1005635

PubMed Abstract | Crossref Full Text | Google Scholar

57. Deng L, Qi X, Wang G, Wang L, and Liu S. The diagnostic value of ultra-microangiography for biliary atresia. Quantitative Imaging Med Surg. (2025) 15:1784–92. doi: 10.21037/qims-24-1586

PubMed Abstract | Crossref Full Text | Google Scholar

58. Wang Y, Zheng C, Zhou Y, Li L, Peng H, and Zhang C. Novel method for fetal and maternal heart rate measurements using 2-D ultrasound color doppler flow images. Ultrasound Med Biol. (2022) 48:2029–39. doi: 10.1016/j.ultrasmedbio.2022.05.027

PubMed Abstract | Crossref Full Text | Google Scholar

59. Murai Y, Nakagawa S, Matano F, Shirokane K, Teramoto A, and Morita A. The feasibility of detecting cerebral blood flow direction using the indocyanine green video angiography. Neurosurg Rev. (2016) 39:685–90. doi: 10.1007/s10143-016-0726-7

PubMed Abstract | Crossref Full Text | Google Scholar

60. Saputra TA, Indra I, Syamsu SA, Sampepajung E, Nelwan BJ, Hamid F, et al. Vascular endothelial growth factor-a expression is significantly correlated with her2 expression in late-stage breast cancer patients. Breast Dis. (2022) 41:433–8. doi: 10.3233/bd-229006

PubMed Abstract | Crossref Full Text | Google Scholar

61. Schoppmann SF, Tamandl D, Roberts L, Jomrich G, Schoppmann A, Zwrtek R, et al. Her2/neu expression correlates with vascular endothelial growth factor-C and lymphangiogenesis in lymph node-positive breast cancer. Ann Oncol Off J Eur Soc Med Oncol. (2010) 21:955–60. doi: 10.1093/annonc/mdp532

PubMed Abstract | Crossref Full Text | Google Scholar

62. Ding W, Meng Y, Ma J, Pang C, Wu J, Tian J, et al. Contrast-enhanced ultrasound-based ai model for multi-classification of focal liver lesions. J Hepatol. (2025) 83:426–39. doi: 10.1016/j.jhep.2025.01.011

PubMed Abstract | Crossref Full Text | Google Scholar

63. Chen F, Han H, Wan P, Chen L, Kong W, Liao H, et al. Do as sonographers think: contrast-enhanced ultrasound for thyroid nodules diagnosis via microvascular infiltrative awareness. IEEE Trans Med Imaging. (2024) 43:3881–94. doi: 10.1109/tmi.2024.3405621

PubMed Abstract | Crossref Full Text | Google Scholar

64. Zhang G, Liao C, Hu JR, Hu HM, Lei YM, Harput S, et al. Nanodroplet-based super-resolution ultrasound localization microscopy. ACS Sens. (2023) 8:3294–306. doi: 10.1021/acssensors.3c00418

PubMed Abstract | Crossref Full Text | Google Scholar