Diagnosis of axillary lymph node metastasis in breast cancer: a systematic review and meta-analysis of the literature on ultrasound and magnetic resonance imaging published from 2014 to 2025

Introduction

Breast cancer is among the most prevalent of malignant neoplasms in females (1). In the United States, breast cancer ranks second among malignancies in cancer-related mortality among women, while in China, there has been a significant and rapid increase in both the occurrence frequency and cancer-related mortality of breast cancer among women (2,3). Axillary lymph node (ALN) metastasis serves as a significant reference indicator for patients diagnosed with breast cancer in deciding upon a treatment plan (4,5). In clinical practice, sentinel lymph node biopsy (SLNB) is commonly employed as an initial surgical approach for axillary staging and the diagnosis of ALN metastasis; when found to have sentinel lymph node metastasis, patients typically proceed to further surgery, namely, axillary lymph node dissection (ALND) (6). ALND can not only help with treatment to a certain extent but can also improve the accuracy of prognosis evaluation (7). However, the above-mentioned diagnostic methods are invasive and may carry certain risks and complications, such as lymphedema, sensory abnormalities, and reduced range of motion in the ipsilateral shoulder (8). Therefore, there is a clear need to develop noninvasive methods that can effectively diagnose ALN metastasis in the early stages.

Magnetic resonance imaging (MRI), which has greater contrast in soft tissue imaging than does computed tomography (CT), is a nonionizing and noninvasive in vivo imaging technique (9,10). Consequently, MRI holds substantial value in the early detection and diagnosis, preoperative staging, and treatment evaluation of patients with breast cancer (11,12). During the appraisal of ALN, MRI is capable of not only revealing the characteristics of lymph nodes but also providing information regarding tumor cell diffusion and perfusion via its functional imaging modalities, such as diffusion-weighted imaging (DWI) and dynamic contrast-enhanced MRI (DCE MRI) (13). This helps distinguish between malignant and benign lymph nodes, thus providing a more comprehensive assessment of lymph node status. However, MRI involves certain limitations: Some patients with implanted devices, claustrophobia, or limited access cannot undergo MRI (14). Furthermore, MRI examinations entail high costs, consume a significant amount of time, and require sophisticated equipment and skilled operators (15).

In contrast, ultrasound (US) is a convenient, low-cost, real-time dynamic imaging method that can assess ALN status in patients with breast cancer without the need for ionizing radiation (16). In addition to basic conventional US scanning, techniques such as contrast-enhanced US, among other modalities, have demonstrated considerable value in diagnosis. Obtaining images of suspicious lymph nodes is one of the most flexible methods for assessing ALN status (17). However, US also involves certain limitations: Some ALNs are anatomically positioned posterior to the pectoralis minor muscle, which presents an obstacle to their detection and observation via US (18). In addition, US quality is significantly dependent upon the experience and proficiency of the operator, and a portion of benign lesions (such as inflammation and reactive hyperplasia) can sonographically resemble malignant metastases, which may potentially lead to misdiagnosis (19,20).

Despite their disadvantages, US and MRI are the two most commonly used noninvasive imaging modalities for preoperative axillary staging due to their accessibility (US) and superior soft-tissue contrast and functional sequences (MRI) and nonionizing nature. Meanwhile, although positron emission tomography-computed tomography (PET/CT) may be used in select scenarios, it is less widely applied for routine axillary staging (21). Previous reviews [e.g., Le Boulc’h et al.’s study (22)] summarized imaging modalities (e.g., US, MRI and FDG-PET) for ALN assessment in breast cancer but either predated the publication of more recent contrast-enhanced ultrasound (CEUS) and DWI/DCE-MRI studies or used broad inclusion criteria (For example, methods that affect histologic findings, such as US-guided core or fine-needle aspiration, are included as the Reference Standard). Therefore, we conducted an updated, modality-stratified meta-analysis, systematically evaluating and comparing the efficacy of US and MRI in determining ALN metastasis in breast cancer through a review of the relevant literature published between 2014 and 2025. It is hoped this review can aid radiologists in achieving greater diagnostic precision in cases of breast cancer. We present this article in accordance with the PRISMA reporting checklist (available at https://qims.amegroups.com/article/view/10.21037/qims-2024-2634/rc).

Methods

Before literature search was started, the protocol for this review was registered with the PROSPERO (International Prospective Register of Systematic Reviews) registry (No. CRD42024589342).

Search strategy

This meta-analysis compiled and synthesized the extant literature on the application of US and MRI techniques in assessing ALN status, including lymph node metastasis, in patients with breast cancer. Given that this analysis was based on previously published data, ethical authorization and patient permission were not required. The relevant literature was independently obtained from three major sources: PubMed database, Embase, and Web of Science databases. The earliest time node for the publication of the included literature was January 2014. An update of this literature collection was carried out in September 2025. The keywords included the following subject words and free words: “breast tumor”, “sentinel lymph nodes”, “lymph node metastasis”, “axillary lymph nodes”, “ultrasound”, and “magnetic resonance imaging”. These keywords were combined to generate the retrieval strategy. The details on the literature search strategy can be found in Appendix 1. In addition, a manual search was conducted of the references in systematic reviews related to the topic, which was followed by the retrieval and evaluation of other relevant studies to augment the comprehensiveness of the search.

Literature selection

In this study, two researchers (C.C., J.W.) separately examined the titles and abstracts of papers to identify those potentially suitable for inclusion and then carefully read through the full texts to determine the final papers for inclusion. Discrepancies were resolved via discussion and, when needed, through consultation with a third reviewer (W.L.).

The inclusion criteria were as follows: (I) published in the English language; (II) a cohort study design; (III) a publication date between January 2014 and September 2025; (IV) a focus on imaging examinations of ALN involvement in breast cancer; (V) the inclusion of US and/or MRI; (VI) pathological examination of ALN tissue acquired via SLNB or ALND; and (VII) the reporting of true-positive (TP), false-positive (FP), true-negative (TN), and false-negative (FN) values or the information from which such values could be derived when a fourfold table was not directly provided.

Meanwhile, the exclusion criteria were as follows: (I) research in which neoadjuvant chemotherapy was administered either in the period between imaging and axillary surgery or before imaging and axillary surgery; (II) research on patients with breast cancer with accessible ipsilateral ALNs; (III) absence of a histopathological reference benchmark; (IV) studies not involving patients with breast cancer; (V) inadequate data for computing the TP, FP, TN, and FN values; (VI) imaging studies with a sole focus on sentinel ALN detection; (VII) patients already included in other studies; (VIII) studies with animal or in vitro subjects; (IX) case-control studies, literature reviews, case reports, editorial correspondence, etc. (X); and unavailability of a full-text document.

It should be noted that we chose to include cohort studies to ensure temporal consistency between the imaging and histopathology and to limit spectrum and selection bias that are more common in cross-sectional case-control designs.

Quality assessment and data extraction

The extraction of data from each study was conducted independently by the same two investigators and separately from the process of research selection. Information recorded included name of the leading author, the region, year of publication, diagnostic approach, trial protocol, number of cases, gold standard used, sensitivity, specificity, fourfold table data, average age of the diagnosed population, and the trial period. Discrepancies were resolved via discussion and, when needed, through consultation with a third reviewer. Quality assessment in this meta-analysis was conducted via the Diagnostic Accuracy Study Quality Assessment-2 (QUADAS-2) tool, which was separately filled out by the same two observers mentioned above. The outcomes of the QUADAS-2 were generated and presented using RevMan 5.3 software (Cochrane, London, UK), specifically developed by the Cochrane Collaboration Network (23).

Statistical analysis

All data were systematically organized with Excel 2019 (Microsoft Corp., Redmond, WA, USA). The statistical evaluation of sensitivity, specificity, and diagnostic odds ratio (DOR) with their associated 95% confidence intervals (CIs) for the two diagnostic modalities examined was performed via STATA software version 14.0 (StataCorp., College Station, TX, USA). Moreover, the area under the receiver operating characteristic curve (AUC) and its 95% CI were calculated to assess the diagnostic value of the methods, with a greater AUC indicating greater diagnostic capacity. For the assessment of statistical heterogeneity among studies, both the I² statistical measure and the Cochrane Q test were applied. When the I² index fell below the 50% threshold, the analysis of sensitivity and specificity included a fixed effects model. Conversely, when the I² index exceeded 50%, the analysis was conducted with a random effects model (24). Subsequently, STATA 14.0 was used to generate Deeks’ funnel plots for US and MRI, respectively. When significant statistical heterogeneity was identified, a meta-regression analysis was initiated to investigate potential contributing factors. Statistical significance was ascertained through application of a P value threshold of less than 0.05. Finally, we used a bivariate random effects model to derive pooled sensitivity (PSEN) and pooled specificity (PSPE) and their corresponding 95% CIs. Heterogeneity was assessed via the I² and Q test. We prespecified sensitivity analyses (leave one out, subgroup by study design, etc.). Fagan plots were constructed to translate pooled likelihood ratios to posttest probabilities for pretest probabilities of 25%, 50%, and 75%. The covariates included in the univariate meta-regression were study design (prospective vs. retrospective), sample size (≥200 vs. <200 total metastatic nodes), and publication year (≥2019 vs. <2019). For studies on contrast-enhanced US, the covariates further included the injection method (percutaneous vs. intravenous) and type of contrast agent (SonoVue vs. Sonazoid); meanwhile, for studies on MRI, the covariates further included the Tesla number (T; 1.5 T vs. 3.0 T) and single- or multi-Tesla number (single vs. 1.5 T or 3.0 T). It should be noted that the 2019 cutoff was chosen as it approximates the median publication year in our set.

Results

Literature search

A comprehensive literature for documents published as of September 2025 was carried out, with 855 records being retrieved for the analysis. Among them, 176 were from PubMed, 324 from Embase, and 355 from Web of Science. After duplicate articles were removed, two investigators independently examined the titles and abstracts of the remaining articles, with literature reviews, case studies, news articles, and other formats being excluded. Subsequently, a more in-depth, full-text review was carried out on the acquired articles. Ultimately, 41 studies that were in accordance with criteria of the meta-analysis were included. The details of the selection procedure are provided in Figure 1.

Figure 1 Research flowchart. The meta-analysis included 41 studies.

Research characteristics

The inclusion of the studies was conducted according to the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) diagnostic research selection protocol. All included studies were published between 2014 and 2024, with their publication dates uniformly determined by the publication date displayed on the record pages of PubMed, Embase, and Web of Science; for studies with both online early access and formal publication dates, the latter was preferred. Among these studies, there were 19 retrospective investigations (25-43), 16 prospective studies (44-59), and 6 (60-65) that did not clarify the research type. However, in the process of data extraction, we determined that the methodological characteristics of those articles without an explicitly stated design were cohort studies and were thus included. Among the 17 studies on US, two studies were from Europe (Spain and the Netherlands) (28,44), one study was from Africa (Egypt) (45), and the remaining were from Asia (Korea, Turkey, and China) (25-27,29-35,46-49). Among the 17 studies on CEUS, 2 were from Europe (Italy and Greece) (50,60), and the remaining were from Asia (China and Japan) (35,36,49,51-57,61-65). Among the 16 studies on MRI, 2 were from Europe (both from the Netherlands) (28,59), 1 was from Africa (Egypt) (45), and the remaining were from Asia (Korea, Turkey, and China) (25-27,37-43,46,48,58). Guney et al. (46) and Dulgeroglu et al. (39) did not specify the Tesla number of the MRI machines used in their studies. All studies used histological biopsy as the gold standard.

A comprehensive overview of the studies can be found in Tables 1-3.

Method quality assessment

The methodological quality of the included studies was assessed with the QUADAS-2 tool, as shown in Figure 2. This tool evaluates the risk of bias across four domains—patient selection, index test, reference standard, and flow and timing—and examines applicability concerns in three domains—patient selection, index test, and reference standard. Overall, the majority of studies demonstrated a low risk of bias, indicating generally high methodological quality. Regarding risk of bias, 3 studies (44,49,50) were deemed to have high risk in the patient selection domain, 15 studies (25,27,28,30-32,40,41,51,56,60-63,65) were rated as unclear risk, and the remaining studies were assessed as low risk. In the index test domain, the studies were considered low risk, except for those of Li et al. (34), Wang et al. (48), and Su et al. (35), which were deemed to be high risk, whereas the studies by Chen et al. (30), Liu et al. (52), and Zhuang et al. (57) were assessed as unclear risk. In terms of reference standard, most studies exhibited a low risk of bias, with only the study by Hu et al. (49) rated as high risk, while those of Agliata et al. (50), Dellaportas et al. (60), Li et al. (51), Liu et al. (52), Qiao et al. (55), Yang et al. (65), Zhang et al. (36), Zheng et al. (56), and Zhuang et al. (57) were rated as having unclear risk. The flow and timing domain also indicated an overall low risk, with only the studies by Li et al. (51) and Liu et al. (52) being rated as high risk and several studies (26,33,36,40,42,44,56,57) being rated as unclear risk. Regarding applicability concerns, only one study, that by Agliata et al. (50), was considered to be as high risk in the patient selection domain, while the studies by Qiao et al. (55) and Sohn et al. (32) showed unclear risk; the remaining studies were rated as low risk. The index test and reference standard domains demonstrated uniformly low applicability concerns, with only the studies by Ma et al. (63) and Niu et al. (53) being judged as high risk and several studies (36,52,55-57,61,65) as unclear; this suggests that the included studies employed diagnostic methods and reference standards highly consistent with the review item criteria. Overall, most studies exhibited a low risk of bias and low applicability concerns, with only a few showing high or unclear risk in patient selection or procedural aspects. In summary, the included studies were of high methodological quality, supporting the robustness and reliability of the pooled results.

Figure 2 Bias risk of the included studies (QUADAS 2 criteria). Judgments of review authors for each domain of each included study. QUADAS-2, Quality Assessment of Diagnostic Accuracy Studies-2.

Sensitivity and specificity of conventional ultrasound (CUS), CEUS, and MRI

For identifying ALN metastasis in breast cancer, CUS demonstrated a PSEN of 0.69 (95% CI: 0.56–0.80) and a PSPE of 0.85 (95% CI: 0.77–0.90) (Figure 3). The heterogeneity, as quantified by the Higgins I² index, was significant for both sensitivity (P<0.05; I²=95.97%) and specificity (P<0.05; I²=94.58%). This prompted the selection of a random effects model for further analysis. The DOR associated with CUS was 11.21 (95% CI: 8.36–15.05; Figure 4). The AUC for CUS was 0.85 (95% CI: 0.82–0.88; Figure 5). The Spearman correlation coefficient for CUS (r=−0.91; P=0.83) did not indicate significant threshold effect, given that P value was above 0.05, suggesting that other factors could be the source of heterogeneity.

Figure 3 The forest plot of sensitivity and specificity for diagnosing axillary lymph node metastasis of breast cancer based on CUS. CI, confidence interval; CUS, conventional ultrasound.

Figure 4 The DOR for diagnosing axillary lymph node metastasis of breast cancer based on CUS. CI, confidence interval; CUS, conventional ultrasound; DOR, diagnostic odds ratios.

Figure 5 The SROC of CUS. Values in brackets represent 95% confidence intervals. AUC, area under the summary receiver operating characteristic curve; CUS, conventional ultrasound; SENS, sensitivity; SPEC, specificity; SROC, summary receiver operating characteristic curve.

The PSEN of the CEUS-based diagnosis of ALN metastasis in breast cancer was 0.86 (95% CI: 0.79–0.91) and the PSPE was 0.88 (95% CI: 0.84–0.91) (Figure 6). Higgins I² statistical results indicated heterogeneity for the sensitivity (P<0.05; I²=84.76%) and specificity (P<0.05; I²=75.24%) of CEUS. Therefore, we chose the random effects model to analyze the sensitivity and specificity. The DOR of CEUS was 42.35 (95% CI: 22.86–78.45; Figure 7). The AUC based on CEUS was 0.93 (95% CI: 0.90–0.95) (Figure 8). The Spearman correlation coefficient for CEUS (r=0.14; P=0.02) showed that there was a significant threshold effect (P<0.05).

Figure 6 The forest plot of sensitivity and specificity for diagnosing axillary lymph node metastasis of breast cancer based on CEUS. CEUS, contrast-enhanced ultrasound; CI, confidence interval.

Figure 7 The DOR for diagnosing axillary lymph node metastasis of breast cancer based on CEUS. CEUS, contrast-enhanced ultrasound; CI, confidence interval; DOR, diagnostic odds ratios.

Figure 8 The SROC of CEUS. Values in brackets represent 95% confidence intervals. AUC, area under the summary receiver operating characteristic curve; CEUS, contrast-enhanced ultrasound; SENS, sensitivity; SPEC, specificity; SROC, summary receiver operating characteristic curve.

For application of MRI in the detection of ALN metastasis in breast cancer, the PSEN was 0.62 (95% CI: 0.53–0.71), and the PSPE was 0.85 (95% CI: 0.79–0.89) (Figure 9). According to the Higgins I² index, there was marked heterogeneity for both sensitivity (P<0.05; I²=88.28%) and specificity (P<0.05; I²=86.03%). This led to the adoption of a random effects model for the analysis. The DOR derived from the MRI data was 8.51 (95% CI: 5.72–12.67; Figure 10). The AUC for MRI was 0.82 (95% CI: 0.78–0.85; Figure 11). The Spearman correlation coefficient for MRI (r=−0.56; P=0.31) did not indicate a significant threshold effect, with the P value exceeding 0.05, suggesting that the heterogeneity could stem from factors other than the threshold.

Figure 9 The forest plot of sensitivity and specificity for diagnosing axillary lymph node metastasis of breast cancer based on MRI. CI, confidence interval; MRI, magnetic resonance imaging.

Figure 10 The DOR for diagnosing axillary lymph node metastasis of breast cancer based on MRI. CI, confidence interval; DOR, diagnostic odds ratios; MRI, magnetic resonance imaging.

Figure 11 The SROC of MRI. Values in brackets represent 95% confidence intervals. AUC, area under the summary receiver operating characteristic curve; MRI, magnetic resonance imaging; SENS, sensitivity; SPEC, specificity; SROC, summary receiver operating characteristic curve.

Publication bias

The Deeks’ funnel plots for CUS, CEUS, and MRI are shown in Figures 12-14. The plots of CUS and MRI exhibited significant asymmetry, with corresponding P values of 0.28 and 0.57 (both P values >0.05), as shown in Figures 12,14. This implies an absence of a significant likelihood of publication bias. However, Figure 13 shows asymmetry in the plot of CEUS, and the corresponding P values are 0.03 (<0.05), which means that there is the possibility of publication bias, and thus further analysis is warranted.

Figure 12 The publication bias of the included studies on CUS. No significant publication bias was found in the meta-analysis. Each circle represents an eligible article. CUS, conventional ultrasound; ESS, effective sample size.

Figure 13 The publication bias of the included studies on CEUS. Significant publication bias was found in the meta-analysis (P<0.05). Each circle represents an eligible article. CEUS, contrast-enhanced ultrasound; ESS, effective sample size.

Figure 14 The publication bias of the included studies of MRI. No significant publication bias was found in the meta-analysis. Each circle represents an eligible article. ESS, effective sample size; MRI, magnetic resonance imaging.

Heterogeneity detection

Due to the heterogeneity present in the studies included, we used regression analysis to examine the potential sources of heterogeneity related to CUS, CEUS, and MRI.

In the meta-analysis of CUS, specificity showed statistically significant between-study heterogeneity (P<0.05), and this heterogeneity was correlated with sample size (P<0.05). For CEUS, the heterogeneity for specificity was significant and was correlated with experimental design, sample size, the injection method, and type of contrast agent (P<0.05); the heterogeneity for sensitivity was also significant and was correlated with sample size (P<0.05). For studies on DWI, specificity showed heterogeneity, which was associated with publication year (P<0.05). Studies on DCE MRI also demonstrated considerable heterogeneity in terms of specificity, which was associated with experimental design, sample size, publication year, and single- or multiple Tesla number (all P values <0.05). The results of the meta-regression analysis are shown in Tables 4-7.

Among the studies on CUS, concerning the sample size of the studies, the PSEN of studies with a number of metastatic lymph nodes <200 and those with ≥200 was 0.73 (95% CI: 0.58–0.89) and 0.64 (95% CI: 0.46–0.83). There was no statistically significant difference in PSEN of the sample size in the PSEN (P>0.05). Meanwhile, the PSPE of studies with a number of metastatic lymph nodes <200 and those with ≥200 was 0.83 (95% CI: 0.74–0.9) and 0.86 (95% CI: 0.77–0.95) and was statistically significant (P<0.05) in PSPE of the sample size.

Among the studies on CEUS, the PSEN in prospective and retrospective studies was 0.91 (95% CI: 0.86–0.95) and 0.85 (95% CI: 0.71–0.99), respectively. There was no statistically significant difference in the PSEN (P>0.05); meanwhile, the PSPE was 0.88 (95% CI: 0.84–0.92) and 0.83 (95% CI: 0.75–0.92), respectively, representing a significant difference (P<0.05) in PSPE of study design. The PSEN of studies with a number of metastatic lymph nodes ≥200 and those with <200 was 0.9 (95% CI: 0.81–0.98) and 0.85 (95% CI: 0.78–0.91), respectively, representing a significant difference in the PSEN (P<0.05); meanwhile, the PSPE was 0.90 (95% CI: 0.85–0.90) and 0.87 (95% CI: 0.82–0.91), respectively, also representing a significant difference (P<0.05). The PSEN of studies on percutaneous puncture and that on those on intravenous injection was 0.87 (95% CI: 0.81–0.93) and 0.84 (95% CI: 0.73–0.95), respectively, representing a significant difference in the PSEN (P>0.05); meanwhile, the PSPE was 0.89 (95% CI: 0.85–0.92) and 0.84 (95% CI: 0.76–0.92), respectively, also representing a significant difference (P<0.05). Moreover, the PSEN of studies on SonoVue and those on Sonazoid was 0.85 (95% CI: 0.79–0.92) and 0.89 (95% CI: 0.77–1.00), representing, which did not represent a significant difference in the PSEN (P>0.05); meanwhile, the PSPE was 0.88 (95% CI: 0.84–0.92) and 0.89 (95% CI: 0.81–0.98), respectively, representing a significant difference (P<0.05).

For the studies on DWI, the PSEN of studies published during or after 2019 and those published before 2019 was 0.69 (95% CI: 0.51–0.87) and 0.59 (95% CI: 0.40–0.78), respectively, which did not represent a significant difference in the PSEN (P>0.05); meanwhile, the PSPE was 0.78 (95% CI: 0.63–0.94) and 0.91 (95% CI: 0.85–0.98), respectively, representing a statistically significant difference (P<0.05).

For the studies on DCE MRI, the PSEN of prospective and retrospective studies was 0.68 (95% CI: 0.50–0.87) and 0.61 (95% CI: 0.48–0.74), respectively, which did not indicate a significant difference in the PSEN (P>0.05); meanwhile, the PSPE was 0.82 (95% CI: 0.73–0.92) and 0.84 (95% CI: 0.79–0.90), respectively, representing a significant difference (P<0.05). Additionally, the PSEN of studies published during and after 2019 and those published before 2019 was 0.58 (95% CI: 0.41–0.75) and 0.67 (95% CI: 0.53–0.82), respectively, which did not indicate a significant difference in the PSEN (P>0.05); meanwhile, the PSPE was 0.83 (95% CI: 0.76–0.90) and 0.84 (95% CI: 0.78–0.91), respectively, representing a significant difference (P<0.05). Furthermore, the PSEN of studies with a number of metastatic lymph nodes ≥200 and those with <200 was 0.70 (95% CI: 0.53–0.86) and 0.59 (95% CI: 0.46–0.73), respectively, which indicated no significant difference in the PSEN (P>0.05); meanwhile, the PSPE was 0.81 (95% CI: 0.73–0.90) and 0.85 (95% CI: 0.80–0.90), respectively, indicating a significant difference (P<0.05). Finally, the PSEN of studies with a single Tesla number and those with multiple Tesla numbers was 0.59 (95% CI: 0.47–0.71) and 0.63 (95% CI: 0.51–0.76), respectively, which did not indicate a significant difference in the PSEN (P>0.05); meanwhile, the PSPE was 0.86 (95% CI: 0.81–0.92) and 0.85 (95% CI: 0.78–0.91), respectively, representing a significant difference (P<0.05).

Sensitivity analysis

To determine whether there were studies that influenced the stability of the PSEN and PSPE, we adopted a leave-one-out analysis. The findings of the sensitivity and specificity analyses are displayed in Tables 8-10. It was found that after the exclusion of each individual paper on CUS, no significant differences were observed in the PSEN or PSPE. Among the literature on CEUS, when the papers by Li et al. (51) and Yang et al. (65) were excluded, there was a difference in the Higgins I² values for PSEN as compared to when other studies were excluded; meanwhile, when the papers by Li et al. (51), Liu et al. (52), and Zheng et al. (56), were excluded there was a difference in the Higgins I² value for PSPE as compared to when other papers were excluded; however, there was no significant difference in PSEN or PSPE when other articles were excluded (both P values <0.05). For studies on MRI, when the studies by Guney et al. (46) and van Nijnatten et al. (the data of the second group) (28) were excluded, there was difference in the Higgins I² values of PSEN as compared to when other studies were excluded; meanwhile, when the studies by Sae-Lim et al. (42) and van Nijnatten et al. (the data of the second group) (28) were excluded, there was a difference in the Higgins I² value for PSPE as compared to when other articles were excluded; however, there was no significant difference in PSEN or PSPE when other articles were excluded (both P values <0.05).

Fagan diagram analysis

Fagan diagram analysis revealed that CUS, CEUS, and MRI are capable of providing a certain level of assistance to radiologists when evaluating ALN status in patients with breast cancer (as shown in Figures 15-17). In the context of US, when the pretest probability was 50%, the posttest probability for correctly categorizing metastatic lymph nodes as positive reached 85%, and the posttest probability of erroneously classifying them as negative was 24%; when the pretest probabilities were 25% and 75%, the positive posttest probabilities were 65% and 94%, respectively, while the negative posttest probabilities were 10% and 49%, respectively. For CEUS, when the pretest probability was 50%, the posttest probability of correctly identifying metastatic lymph nodes as positive was 82%, while the posttest probability of identifying erroneously classifying them as negative was 27%; meanwhile, when the pretest probability was 25% and 75%, the posttest probability for a correct positive classification was 60% and 93%, respectively, while the posttest probability of incorrect negative classification was 11% and 52%, respectively. For MRI, when the pretest probability was 50%, the posttest probability of correctly identifying metastatic lymph nodes as positive was 80%, and the posttest probability of misidentifying them as negative was 31%; meanwhile, when the pretest probabilities were 25% and 75%, the posttest probabilities were 57% and 92%, respectively, and the negative posttest probabilities were 13% and 58%, respectively.

Figure 15 Fagan plot analysis for the ability of CUS to diagnose axillary lymph node metastasis of breast cancer. (A) Pretest probability at 25%. (B) Pretest probability at 50%. (C) Pretest probability at 75%. The Fagan plot is composed of the left vertical axis representing the pretest probability, the middle vertical axis representing the likelihood ratio, and the right vertical axis representing the posttest probability. CUS, conventional ultrasound; LR, likelihood ratio.

Figure 16 Fagan plot analysis for the ability of CEUS to diagnose axillary lymph node metastasis of breast cancer. (A) Pretest probability at 25%. (B) Pretest probability at 50%. (C) Pretest probability at 75%. The Fagan plot is composed of the left vertical axis representing the pretest probability, the middle vertical axis representing the likelihood ratio, and the right vertical axis representing the posttest probability. CEUS, contrast-enhanced ultrasound; LR, likelihood ratio.

Figure 17 Fagan plot analysis for the ability of MRI to diagnose axillary lymph node metastasis of breast cancer. (A) Pretest probability at 25%. (B) Pretest probability at 50%. (C) Pretest probability at 75%. The Fagan plot is composed of the left vertical axis representing the pretest probability, the middle vertical axis representing the likelihood ratio, and the right vertical axis representing the posttest probability. LR, likelihood ratio; MRI, magnetic resonance imaging.

Discussion

This meta-analysis conducted a comprehensive assessment of the performance of CUS, CEUS, and MRI in the determination of ALN metastasis in patients with breast cancer. The results indicated that both US and MRI possess relatively satisfactory capabilities in diagnosing ALN in this setting. CUS had a PSEN and PSPE of 0.69 (95% CI: 0.56–0.80) and 0.85 (95% CI: 0.77–0.90), respectively; a DOR of 11.21 (95% CI: 8.36–15.05); and an AUC of 0.85 (95% CI: 0.82–0.88). CEUS had a PSEN and PSPE of 0.86 (95% CI: 0.79–0.91) and 0.88 (95% CI: 0.84–0.91), respectively; a DOR of 42.35 (95% CI: 22.86–78.45); and an AUC of 0.93 (95% CI: 0.90–0.95). Finally, MRI had a PSEN and PSPE of 0.62 (95% CI: 0.53–0.71) and 0.85 (95% CI: 0.79–0.89), respectively; a DOR of 8.51 (95% CI: 5.72–12.67); and an AUC of 0.82 (95% CI: 0.78–0.85).

Previous meta-analyses have typically focused on a single technique or direct or indirect data comparisons of different techniques [e.g., the studies of Boulc’h et al. (22) and Li et al. (66)]. Our study differed in that we performed modality-stratified pooled analyses and modality-specific subgroup/meta-regression for literature published in the 2014–2025 period. In addition to CUS and MRI, we further conducted a portion of the meta-analysis on CEUS, which is more widely used.

The reviews by Liu et al. (67) and Boulc’h et al. (22) suggest that MRI has higher sensitivity than does US. However, taking into account the publication times of the included studies, the relatively early date of these reviews, and the disparities in the quality evaluation of the included studies, there might be concerns regarding the currency and quality of evidence. Therefore, they can only be used as references for clinicians to make imaging decisions in the earlier years.

Most of the studies integrated into our analysis exhibited relatively good quality. However, it should be noted that there was threshold effect and publication bias in the CEUS group (both P values <0.05). Similarly, a number of the studies failed to furnish particulars concerning the continuity and randomness of case selection or the blinding design of the experiments. This led to a somewhat inferior quality of the reports, which might have introduced bias and significant heterogeneity. As a result, we employed sensitivity analysis and meta-regression to elucidate this high heterogeneity.

In the sensitivity analysis on CUS, conducted to identify the sources of heterogeneity, we systematically excluded each study one by one to assess its influence on the pooled estimates, and the PSEN, PSPE, and Higgins I² of the CUS studies did not show significant changes. However, for MRI, when the studies by Guney et al. (46), van Nijnatten et al. (the data of the second group) (28), and Sae-Lim et al. (42) were excluded, Higgins I² showed a certain degree of difference. When the studies by Guney et al. (46) and Sae-Lim et al. (42) were excluded, there were differences in the Higgins I² values of PSEN and PSPE, respectively. When the study by van Nijnatten et al. (the data of the second group) (28) was excluded, there were differences in the Higgins I² values for both PSEN and PSPE. This suggests that these three studies may be an important factor contributing to the high heterogeneity between studies. However, when the results from these three studies were separately excluded, their PSEN and PSPE did not show significant differences, indicating relatively good consistency and stability between the studies. For these three studies, we found that the patients they included were all patients with breast cancer diagnosed by biopsy, which may be the source of heterogeneity, meaning that some of the patients with breast cancer that might not have been diagnosed were excluded. However, for other factors such as sample size and study design, this heterogeneity did not exert a significant impact on the final PSEN and PSPE.

When the studies by Li et al. (51),Yang et al. (65), Liu et al. (52), and Zheng et al. (56) were excluded from the CEUS studies, the Higgins I² showed a certain degree of difference. When the studies by Li et al. (51) and Yang et al. (65) were excluded, discrepancies were observed in the Higgins I² values for PSEN. Meanwhile, when the studies by Li et al. (51), Liu et al. (52), and Zheng et al. (56) were excluded, differences emerged in the Higgins I² values for PSPE. This suggests that these four studies may be important factors leading to the heterogeneity among the studies. However, when the results of these four studies were excluded, the PSEN and PSPE differed significantly, indicating good consistency and stability between the studies. For CEUS, regardless of whether the contrast agent was injected intravenously (65) or subcutaneously (51,52,56), the subjective dependence of the physician on its operation and diagnosis could be ignored. The most important aspect of CEUS is diagnosis and interpretation based on the dynamic blood flow perfusion of microbubbles under US, which also places higher demands on the understanding of dynamic manifestations by US physicians than does CUS. Therefore, the subjectivity of US physicians cannot be ignored in their interpretation. Studies on subcutaneous injection (51,52,56) standardized the procedure by injecting at the 12, 3, 6, and 9 o’clock positions. However, uncontrolled variables such as variations in patient anatomy (e.g., areola size) and operator technique (e.g., injection depth or post-injection massage) could still introduce inconsistencies that affect the interpretation of dynamic CEUS. These factors could have contributed to the heterogeneity between studies.

As a noninvasive diagnostic method, CUS has long been a routine imaging modality for the preoperative determination of breast lesion characteristics and ALN status (68). In our meta-analysis, 17 studies on CUS were included, among which there were 6 prospective studies and 11 retrospective studies, which demonstrated no statistically significant difference in specificity (P>0.05). Similarly, no significant differences were found in this regard related to the year of publication (7 studies published in 2019 or later and 10 studies published before 2019). Moreover, there were 8 studies with ≥200 metastatic lymph nodes and 9 studies with <200. Regression analysis revealed that there was no significant difference in the research outcomes (P>0.05).

There are certain issues regarding the diagnosis of lymph nodes via CUS. First, the sensitivity of CUS to ALN metastasis varies greatly, with a mean sensitivity of only about 60% (69). Meanwhile, the spatial resolution of CUS is limited, and it cannot be used to identify micrometastases or isolated tumor cells, which account for 10–20% of SLNB lesions (70). Therefore, although CUS has important auxiliary value in the preliminary screening of ALN metastasis of breast cancer, it is still limited by its low sensitivity, marked subjectivity, and insensitivity to micrometastasis.

CEUS is a US technique that demands the injection of a contrast agent to facilitate dynamic and real-time observation of the lesion, which can also play a crucial role in distinguishing between metastatic and nonmetastatic lymph nodes in patients with breast cancer (71). For our meta-analysis on CEUS, we included 11 prospective studies and 3 retrospective studies, which demonstrated statistically significant differences in specificity (P<0.05). This difference may stem from a selective bias caused by a retrospective study design. Regarding the number of metastatic lymph nodes, there were 4 studies with ≥200 metastatic lymph nodes and 16 studies with <200. Regression analysis revealed that there were statistically significant differences in specificity between these study types (P<0.05). In the subgroup of contrast injection methods (13 studies on percutaneous puncture and 7 on intravenous injection), significant differences were also observed in specificity (P<0.05). Similarly, significant differences (P<0.05) in specificity were observed in relation to US contrast agents (3 on using Sonazoid and 16 studies on SonoVue). Percutaneous puncture, as a method for the injection of contrast agent, is more in line with the physiological pathway of breast lymphatic drainage. After percutaneous puncture, contrast agents can simulate physiological drainage pathways through the lymphatic plexus below the areola, making it easier to display true sentinel lymph nodes (SLNs). However, intravenous injection is not related to lymph drainage and cannot determine whether lymph nodes are SLNs (72). Meanwhile, percutaneous lymphatic CEUS can dynamically display the lymphatic drainage pathway and SLN position in real time, while venous CEUS cannot display lymphatic vessel structure and can only evaluate the perfusion status of individual lymph nodes (57).

Among the included DWI studies, 7 were included in the meta-analysis. First, there were no statistically significant differences observed in subgroups for research design, the number of metastatic lymph nodes, or Tesla number. However, between the 3 studies published in 2019 and the 4 published before 2019, there were significant differences in specificity (P<0.05). This difference may stem from the optimization of DWI sequence technology and the improvement in image quality control in the decade before and after 2019. Moreover, the imaging diagnostic physicians in these 3 studies all had more than 7 years of experience, and with the accumulation of clinical experience, their ability to diagnose negative images could also have been enhanced.

A total of 11 studies on DCE MRI were included. The meta-analysis indicated that there were significant statistical differences in specificity between the subgroups of trial design (3 studies had prospective designs, and 8 studies had retrospective designs), the number of metastatic lymph nodes (3 studies ≥200 and 8 studies <200), publication year (5 studies published in 2019 or later, and 6 studies published before 2019), and Tesla number (4 studies on 1.5 T or 3.0 T and 5 on a single Tesla number) (P<0.05). This heterogeneity may be related to the different staging systems for ALNs used across the different studies, optimization of DCE MRI technology, and improvement of image quality control, as well as possible selection bias caused by experimental design. The research findings of Li et al. (73) indicated that there was no significant difference in the outcomes related to the experimental design, but the inclusion scope of the study of Li et al. was expanded compared to our study. The study of Li et al. included cases after neoadjuvant chemotherapy, which not only expands the universal applicability of DCE MRI in clinical applications but also affects the research results due to the higher number of included studies, methods, and experimental design factors. Therefore, further research is necessary to obtain results that better reflect real-world practice.

Moreover, a Fagan diagram was used to examine the clinical application of CUS, CEUS, and MRI in assessing ALN status in patients with breast cancer. During an US examination, if a patient is estimated to have a 50% likelihood of ALN following the initial clinical assessment, if the CUS examination results are positive, the odds of ALN metastasis increase from 50% to 82%. Conversely, if the CUS examination result is negative, the possibility of ALN metastasis in the patient decreases to 27%. In the context of CEUS, if a patient is deemed to have a 50% likelihood of ALN metastasis after the initial clinical evaluation and if the MRI result is positive, the odds of ALN metastasis increases from 50% to 87%. Conversely, if the MRI result is negative, the possibility of ALN metastasis in the patient decreases to 14%. In the context of MRI, if a patient is deemed to have a 50% likelihood of ALN metastasis after the initial clinical evaluation, and if the MRI result is positive, the odds of ALN metastasis increases from 50% to 80%. Conversely, if the MRI result is negative, the likelihood of ALN metastasis in the patient decreases to 31%. These outcomes also indicate that these methods exhibit good diagnostic capabilities for patients who have ALN metastasis and can assist clinicians in making a diagnosis.

Owing to the high spatial resolution for superficial tissues and real-time guidance capability, CUS enables precise evaluation of cortical morphology and hilar structure (74). The addition of CEUS further improves visualization of nodal perfusion and microvascular changes, which facilitates the early detection of micro-metastases (75). In contrast, MRI performance can be affected by variations in sequence design or magnetic-field strength and limited coverage of certain anatomical regions such as the retropectoral nodes (76). From a clinical perspective, these differences suggest that CUS and CEUS may serve as first-line modalities for preoperative axillary staging, while MRI can be valuable in complex or equivocal cases. Accurate imaging evaluation directly informs the surgical strategy—particularly decisions regarding SLNB versus axillary dissection—and thus contributes to individualized treatment planning.

Although the advantages of applying all three technologies are obvious, in practical clinical practice, the costs and economic benefits associated with them must be considered. The study by Meng et al. (77) suggests that MRI may be a cost-effective alternative to SLNB in detecting axillary metastasis, with lower expected total cost and higher expected total mass adjusted life years. The findings from the study by Boughey et al. suggest that preoperative routine axillary US and US-guided fine needle puncture, which are employed for guiding the surgical plan, can reduce the overall cost of care for patients with invasive breast cancer (78). In addition, given that the cost of US is lower than that of MRI in practical clinical application and the operation is relatively more convenient, US is used more frequently than MRI for assessing ALN metastasis. Of course, besides US and MRI, there exist many other examination techniques for evaluating ALN metastasis, such as mammography, CT, and PET/CT. Their costs, economic benefits, and diagnostic capabilities are also different from those of US and MRI discussed in this paper, and further research is needed to evaluate them.

This study involved certain limitations that should be addressed. To begin, the number of studies incorporated into the meta-analysis was relatively small. In the case of high heterogeneity between the two techniques examined, although random effects models were used for analysis, we identified several sources of heterogeneity through heterogeneity testing, subgroup analysis, and sensitivity analysis. It should also be noted that a large proportion of the studies included in this study were from Asia, which might have resulted in a population concentration of patients from this region in the analysis. This uneven geographical distribution may limit the generalizability of research results, especially for populations with significant differences in genetic background, disease spectrum, medical practice, and environmental factors. In addition, the lack of a sufficient number of prospective studies may weaken the strength of the evidence concerning the detection of metastatic ALNs. In order to produce a more comprehensive assessment of the diagnostic efficacy of US and MRI, it is essential to carry out multicenter, large-scale, prospective studies across different regions.

Conclusions

CUS, CEUS, and MRI possess distinct advantages in the diagnosis of ALN metastasis in patients with breast cancer. Therefore, in the selection of diagnostic methods based on actual clinical conditions, careful consideration of factors such as cost and technological requirements is needed.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the PRISMA reporting checklist. Available at https://qims.amegroups.com/article/view/10.21037/qims-2024-2634/rc

Funding: This work was supported by the Tianshan Young Talent Scientific and Technological Innovation Team: Innovative Team for Research on Prevention and Treatment of High-incidence Diseases in Central Asia (No. 2023TSYCTD0020), the National Natural Science Foundation of China (Nos. 82060318, 82460353, and 82260105), and the Corps Science and Technology Key Project (No. 2022CB002-04).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://qims.amegroups.com/article/view/10.21037/qims-2024-2634/coif). All authors report that this work was supported by the Tianshan Young Talent Scientific and Technological Innovation Team: Innovative Team for Research on Prevention and Treatment of High-incidence Diseases in Central Asia (No. 2023TSYCTD0020), the National Natural Science Foundation of China (Nos. 82060318, 82460353, and 82260105), and the Corps Science and Technology Key Project (No. 2022CB002-04). The authors have no other conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. US Preventive Services Task Force. Screening for Breast Cancer. JAMA 2024;331:1973-4. [Crossref] [PubMed]
2. Siegel RL, Giaquinto AN, Jemal A. Cancer statistics, 2024. CA Cancer J Clin 2024;74:12-49. [Crossref] [PubMed]
3. Xia C, Dong X, Li H, Cao M, Sun D, He S, Yang F, Yan X, Zhang S, Li N, Chen W. Cancer statistics in China and United States, 2022: profiles, trends, and determinants. Chin Med J (Engl) 2022;135:584-90. [Crossref] [PubMed]
4. Chang JM, Leung JWT, Moy L, Ha SM, Moon WK. Axillary Nodal Evaluation in Breast Cancer: State of the Art. Radiology 2020;295:500-15. [Crossref] [PubMed]
5. Wang B, Yang J, Tang YL, Chen YY, Luo J, Cui XW, Dietrich CF, Yi AJ. The value of microvascular Doppler ultrasound technique, qualitative or quantitative shear-wave elastography of breast lesions for predicting axillary nodal burden in patients with breast cancer. Quant Imaging Med Surg 2024;14:408-20. [Crossref] [PubMed]
6. Morrison S, Han D. Re-evaluation of Sentinel Lymph Node Biopsy for Melanoma. Curr Treat Options Oncol 2021;22:22. [Crossref] [PubMed]
7. Houssami N, Ciatto S, Turner RM, Cody HS 3rd, Macaskill P. Preoperative ultrasound-guided needle biopsy of axillary nodes in invasive breast cancer: meta-analysis of its accuracy and utility in staging the axilla. Ann Surg 2011;254:243-51. [Crossref] [PubMed]
8. Wernicke AG, Shamis M, Sidhu KK, Turner BC, Goltser Y, Khan I, Christos PJ, Komarnicky-Kocher LT. Complication rates in patients with negative axillary nodes 10 years after local breast radiotherapy after either sentinel lymph node dissection or axillary clearance. Am J Clin Oncol 2013;36:12-9. [Crossref] [PubMed]
9. Cooper KL, Meng Y, Harnan S, Ward SE, Fitzgerald P, Papaioannou D, Wyld L, Ingram C, Wilkinson ID, Lorenz E. Positron emission tomography (PET) and magnetic resonance imaging (MRI) for the assessment of axillary lymph node metastases in early breast cancer: systematic review and economic evaluation. Health Technol Assess 2011;15:iii-iv, 1-134. [Crossref] [PubMed]
10. Harnan SE, Cooper KL, Meng Y, Ward SE, Fitzgerald P, Papaioannou D, Ingram C, Lorenz E, Wilkinson ID, Wyld L. Magnetic resonance for assessment of axillary lymph node status in early breast cancer: a systematic review and meta-analysis. Eur J Surg Oncol 2011;37:928-36. [Crossref] [PubMed]
11. Mann RM, Cho N, Moy L. Breast MRI: State of the Art. Radiology 2019;292:520-36. [Crossref] [PubMed]
12. Panico C, Ferrara F, Woitek R, D'Angelo A, Di Paola V, Bufi E, Conti M, Palma S, Cicero SL, Cimino G, Belli P, Manfredi R. Staging Breast Cancer with MRI, the T. A Key Role in the Neoadjuvant Setting. Cancers (Basel) 2022;14:5786. [Crossref] [PubMed]
13. Wu PQ, Guo FL, Wang J, Gao Y, Feng ST, Chen SL, Ma J, Liu YB. Development and validation of a dynamic contrast-enhanced magnetic resonance imaging-based habitat and peritumoral radiomic model to predict axillary lymph node metastasis in patients with breast cancer: a retrospective study. Quant Imaging Med Surg 2024;14:8211-26. [Crossref] [PubMed]
14. Yalon M, Sae-Kho T, Khanna A, Chang S, Andrist BR, Weber NM, Hoodeshenas S, Ferrero A, Glazebrook KN, McCollough CH, Baffour FI. Staging of breast cancer in the breast and regional lymph nodes using contrast-enhanced photon-counting detector CT: accuracy and potential impact on patient management. Br J Radiol 2024;97:93-7. [Crossref] [PubMed]
15. Sardanelli F, Boetes C, Borisch B, Decker T, Federico M, Gilbert FJ, et al. Magnetic resonance imaging of the breast: recommendations from the EUSOMA working group. Eur J Cancer 2010;46:1296-316. [Crossref] [PubMed]
16. Feig S. Cost-effectiveness of mammography, MRI, and ultrasonography for breast cancer screening. Radiol Clin North Am 2010;48:879-91. [Crossref] [PubMed]
17. Guo R, Lu G, Qin B, Fei B. Ultrasound Imaging Technologies for Breast Cancer Detection and Management: A Review. Ultrasound Med Biol 2018;44:37-70. [Crossref] [PubMed]
18. Kyriacou H, Khan YS. Anatomy, Shoulder and Upper Limb, Axillary Lymph Nodes. Treasure Island (FL): StatPearls Publishing; 2025.
19. Li JM, Shao YH, Sun XM, Shi J. Ultrasonic features of automated breast volume scanner (ABVS) and handheld ultrasound (HHUS) combined with molecular biomarkers in predicting axillary lymph node metastasis of clinical T1-T2 breast cancer. Quant Imaging Med Surg 2024;14:1359-68. [Crossref] [PubMed]
20. Alvarez S, Añorbe E, Alcorta P, López F, Alonso I, Cortés J. Role of sonography in the diagnosis of axillary lymph node metastases in breast cancer: a systematic review. AJR Am J Roentgenol 2006;186:1342-8. [Crossref] [PubMed]
21. Cooper KL, Harnan S, Meng Y, Ward SE, Fitzgerald P, Papaioannou D, Wyld L, Ingram C, Wilkinson ID, Lorenz E. Positron emission tomography (PET) for assessment of axillary lymph node status in early breast cancer: A systematic review and meta-analysis. Eur J Surg Oncol 2011;37:187-98. [Crossref] [PubMed]
22. Le Boulc'h M, Gilhodes J, Steinmeyer Z, Molière S, Mathelin C. Pretherapeutic Imaging for Axillary Staging in Breast Cancer: A Systematic Review and Meta-Analysis of Ultrasound, MRI and FDG PET. J Clin Med 2021;10:1543. [Crossref] [PubMed]
23. Whiting PF, Rutjes AW, Westwood ME, Mallett S, Deeks JJ, Reitsma JB, Leeflang MM, Sterne JA, Bossuyt PM. QUADAS-2: a revised tool for the quality assessment of diagnostic accuracy studies. Ann Intern Med 2011;155:529-36. [Crossref] [PubMed]
24. Maitra S. Fixed-effect versus random-effect model in meta-analysis: How to decide? Indian J Anaesth 2025;69:143-6. [Crossref] [PubMed]
25. An YS, Lee DH, Yoon JK, Lee SJ, Kim TH, Kang DK, Kim KS, Jung YS, Yim H. Diagnostic performance of 18F-FDG PET/CT, ultrasonography and MRI. Detection of axillary lymph node metastasis in breast cancer patients. Nuklearmedizin 2014;53:89-94. [Crossref] [PubMed]
26. Park HL, Yoo IR. O JH, Kim H, Kim SH, Kang BJ. Clinical utility of 18F-FDG PET/CT in low 18F-FDG-avidity breast cancer subtypes: comparison with breast US and MRI. Nucl Med Commun 2018;39:35-43. [Crossref] [PubMed]
27. Shao M, Zi J, Wen G. 2-fluoro-2-deoxy-D-glucose positron emission tomography versus conventional imaging for the diagnosis of breast cancer and lymph node metastases. J Cancer Res Ther 2018;14:S661-6. [Crossref] [PubMed]
28. van Nijnatten TJA, Ploumen EH, Schipper RJ, Goorts B, Andriessen EH, Vanwetswinkel S, Schavemaker M, Nelemans P, de Vries B, Beets-Tan RGH, Smidt ML, Lobbes MBI. Routine use of standard breast MRI compared to axillary ultrasound for differentiating between no, limited and advanced axillary nodal disease in newly diagnosed breast cancer patients. Eur J Radiol 2016;85:2288-94. [Crossref] [PubMed]
29. Chang W, Jia W, Shi J, Yuan C, Zhang Y, Chen M. Role of Elastography in Axillary Examination of Patients With Breast Cancer. J Ultrasound Med 2018;37:699-707. [Crossref] [PubMed]
30. Chen J, Li CX, Shao SH, Yao MH, Su YJ, Wu R. The association between conventional ultrasound and contrast-enhanced ultrasound appearances and pathological features in small breast cancer. Clin Hemorheol Microcirc 2022;80:413-22. [Crossref] [PubMed]
31. He X, Sun L, Huo Y, Shao M, Ma C. A comparative study of 18F-FDG PET/CT and ultrasonography in the diagnosis of breast cancer and axillary lymph node metastasis. Q J Nucl Med Mol Imaging 2017;61:429-37. [Crossref] [PubMed]
32. Sohn YM, Hong IK, Han K. Role of [18F]fluorodeoxyglucose positron emission tomography-computed tomography, sonography, and sonographically guided fine-needle aspiration biopsy in the diagnosis of axillary lymph nodes in patients with breast cancer: comparison of diagnostic performance. J Ultrasound Med 2014;33:1013-21.
33. Zhang YN, Wang CJ, Xu Y, Zhu QL, Zhou YD, Zhang J, Mao F, Jiang YX, Sun Q. Sensitivity, Specificity and Accuracy of Ultrasound in Diagnosis of Breast Cancer Metastasis to the Axillary Lymph Nodes in Chinese Patients. Ultrasound Med Biol 2015;41:1835-41. [Crossref] [PubMed]
34. Li L, Zhao J, Li F, Pan Z. Comparison of MRI and Ultrasound for Evaluation of Axillary Lymph Node Status in Early Breast Cancer. Cancer Manag Res 2024;16:1685-92. [Crossref] [PubMed]
35. Su S, Ye J, Ke H, Zhong H, Lyu G, Xu Z. Multimodal ultrasound imaging: a method to improve the accuracy of sentinel lymph node diagnosis in breast cancer. Front Oncol 2024;14:1366876. [Crossref] [PubMed]
36. Zhang Q, Agyekum EA, Zhu L, Yan L, Zhang L, Wang X, Yin L, Qian X. Clinical Value of Three Combined Ultrasonography Modalities in Predicting the Risk of Metastasis to Axillary Lymph Nodes in Breast Invasive Ductal Carcinoma. Front Oncol 2021;11:715097. [Crossref] [PubMed]
37. Ahn HS, Jang M, Kim SM, La Yun B, Lee SH. Usefulness of preoperative breast magnetic resonance imaging with a dedicated axillary sequence for the detection of axillary lymph node metastasis in patients with early ductal breast cancer. Radiol Med 2019;124:1220-8. [Crossref] [PubMed]
38. Baran MT, Gundogdu H, Demiral G, Kupik O, Arpa M, Pergel A. PET-CT and MR Imaging in the Management of Axillary Nodes in Early Stage Breast Cancer. J Coll Physicians Surg Pak 2020;30:946-50. [Crossref] [PubMed]
39. Dulgeroglu O, Arikan AE, Capkinoglu E, Kara H, Uras C. Comparison of PET-CT and MRI for evaluation of axillary lymph nodes in early breast cancer patients. Ann Ital Chir 2022;93:648-55.
40. Jung NY, Kim SH, Kim SH, Seo YY, Oh JK, Choi HS, You WJ. Effectiveness of Breast MRI and (18)F-FDG PET/CT for the Preoperative Staging of Invasive Lobular Carcinoma versus Ductal Carcinoma. J Breast Cancer 2015;18:63-72. [Crossref] [PubMed]
41. Kim SH, Shin HJ, Shin KC, Chae EY, Choi WJ, Cha JH, Kim HH. Diagnostic Performance of Fused Diffusion-Weighted Imaging Using T1-Weighted Imaging for Axillary Nodal Staging in Patients With Early Breast Cancer. Clin Breast Cancer 2017;17:154-63. [Crossref] [PubMed]
42. Sae-Lim C, Wu WP, Chang MC, Lai HW, Chen ST, Chou CT, Liao CY, Huang HI, Chen ST, Chen DR, Hung CL. Reliability of predicting low-burden (≤ 2) positive axillary lymph nodes indicating sentinel lymph node biopsy in primary operable breast cancer - a retrospective comparative study with PET/CT and breast MRI. World J Surg Oncol 2024;22:12. [Crossref] [PubMed]
43. Yun SJ, Sohn YM, Seo M. Differentiation of benign and metastatic axillary lymph nodes in breast cancer: additive value of MRI computer-aided evaluation. Clin Radiol 2016;71:403.e1-7. [Crossref] [PubMed]
44. Barco I, Chabrera C, García-Fernández A, Fraile M, González S, Canales L, Lain JM, González C, Vidal MC, Vallejo E, Deu J, Pessarrodona A, Giménez N, García Font M. Role of axillary ultrasound, magnetic resonance imaging, and ultrasound-guided fine-needle aspiration biopsy in the preoperative triage of breast cancer patients. Clin Transl Oncol 2017;19:704-10. [Crossref] [PubMed]
45. Elmesidy DS, Badawy EAMO, Kamal RM, Khallaf ESE, AbdelRahman RW. The additive role of diffusion-weighted magnetic resonance imaging to axillary nodal status evaluation in cases of newly diagnosed breast cancer. Egypt J Radiol Nucl Med 2021;52:97.
46. Guney IB, Dalci K, Teke ZT, Kucuker KA. A prospective comparative study of ultrasonography, contrast-enhanced MRI and 18F-FDG PET/CT for preoperative detection of axillary lymph node metastasis in breast cancer patients. Ann Ital Chir 2020;91:458-64.
47. Zhao QL, Xia XN, Zhang Y, He JJ, Sheng W, Ruan LT, Yin YM, Hou HL. Elastosonography and two-dimensional ultrasonography in diagnosis of axillary lymph node metastasis in breast cancer. Clin Radiol 2018;73:312-8. [Crossref] [PubMed]
48. Wang LJ, Yu JC, Hong ZJ. The predict value of lymph node status pre-operation by ultrasound, mammography and MRI in early breast cancer. J Formos Med Assoc 2025;S0929-6646(25)00457-7.
49. Hu Y, Li Q, Huang Y, Peng C, Yang L, Guo Z, Zheng W, Li L, Zhou J. Post-vascular phase of contrast-enhanced ultrasound with perfluorobutane for preoperative evaluation of axillary lymph node status in early-stage breast cancer. Radiol Med 2025;130:991-1002. [Crossref] [PubMed]
50. Agliata G, Valeri G, Argalia G, Tarabelli E, Giuseppetti GM. Role of Contrast-Enhanced Sonography in the Evaluation of Axillary Lymph Nodes in Breast Carcinoma: A Monocentric Study. J Ultrasound Med 2017;36:505-11. [Crossref] [PubMed]
51. Li J, Lu M, Cheng X, Hu Z, Li H, Wang H, Jiang J, Li T, Zhang Z, Zhao C, Ma Y, Tan B, Liu J, Yu Y. How Pre-operative Sentinel Lymph Node Contrast-Enhanced Ultrasound Helps Intra-operative Sentinel Lymph Node Biopsy in Breast Cancer: Initial Experience. Ultrasound Med Biol 2019;45:1865-73. [Crossref] [PubMed]
52. Liu YB, Xia M, Li YJ, Li S, Li H, Li YL. Contrast-Enhanced Ultrasound in Locating Axillary Sentinel Lymph Nodes in Patients with Breast Cancer: A Prospective Study. Ultrasound Med Biol 2021;47:1475-83. [Crossref] [PubMed]
53. Niu Z, Gao Y, Xiao M, Mao F, Zhou Y, Zhu Q, Jiang Y. Contrast-enhanced lymphatic US can improve the preoperative diagnostic performance for sentinel lymph nodes in early breast cancer. Eur Radiol 2023;33:1593-602. [Crossref] [PubMed]
54. Sun Y, Cui L, Wang S, Shi T, Hao Y, Lei Y. Comparative study of two contrast agents for intraoperative identification of sentinel lymph nodes in patients with early breast cancer. Gland Surg 2021;10:1638-45. [Crossref] [PubMed]
55. Qiao J, Li J, Wang L, Guo X, Bian X, Lu Z. Predictive risk factors for sentinel lymph node metastasis using preoperative contrast-enhanced ultrasound in early-stage breast cancer patients. Gland Surg 2021;10:761-9. [Crossref] [PubMed]
56. Zheng Y, Sun J, Zhu L, Hu MS, Hou LZ, Liu JX, Dong FL. Diagnosing sentinel lymph node metastasis of T1/T2 breast cancer with conventional ultrasound combined with double contrast-enhanced ultrasound: a preliminary study. Quant Imaging Med Surg 2023;13:3451-63. [Crossref] [PubMed]
57. Zhuang L, Ming X, Liu J, Jia C, Jin Y, Wang J, Shi Q, Wu R, Jin L, Du L. Comparison of lymphatic contrast-enhanced ultrasound and intravenous contrast-enhanced ultrasound in the preoperative diagnosis of axillary sentinel lymph node metastasis in patients with breast cancer. Br J Radiol 2022;95:20210897. [Crossref] [PubMed]
58. Ergul N, Kadioglu H, Yildiz S, Yucel SB, Gucin Z, Erdogan EB, Aydin M, Muslumanoglu M. Assessment of multifocality and axillary nodal involvement in early-stage breast cancer patients using 18F-FDG PET/CT compared to contrast-enhanced and diffusion-weighted magnetic resonance imaging and sentinel node biopsy. Acta Radiol 2015;56:917-23. [Crossref] [PubMed]
59. Schipper RJ, Paiman ML, Beets-Tan RG, Nelemans PJ, de Vries B, Heuts EM, van de Vijver KK, Keymeulen KB, Brans B, Smidt ML, Lobbes MB. Diagnostic Performance of Dedicated Axillary T2- and Diffusion-weighted MR Imaging for Nodal Staging in Breast Cancer. Radiology 2015;275:345-55. [Crossref] [PubMed]
60. Dellaportas D, Koureas A, Contis J, Lykoudis PM, Vraka I, Psychogios D, Kondi-Pafiti A, Voros DK. Contrast-Enhanced Color Doppler Ultrasonography for Preoperative Evaluation of Sentinel Lymph Node in Breast Cancer Patients. Breast Care (Basel) 2015;10:331-5. [Crossref] [PubMed]
61. Li JT, Zhao HM, Guo XH, Tian PQ, Lü MH, Li LF, Liu ZZ, Cui SD, Zhang HW. Preoperative evaluation of sentinel lymph node biopsy using contrast-enhanced ultrasonography in early breast cancer patients and the involved disturbing factors. Zhonghua Yi Xue Za Zhi 2019;99:1086-9. [Crossref]
62. Matsuzawa F, Omoto K, Einama T, Abe H, Suzuki T, Hamaguchi J, Kaga T, Sato M, Oomura M, Takata Y, Fujibe A, Takeda C, Tamura E, Taketomi A, Kyuno K. Accurate evaluation of axillary sentinel lymph node metastasis using contrast-enhanced ultrasonography with Sonazoid in breast cancer: a preliminary clinical trial. Springerplus 2015;4:509. [Crossref] [PubMed]
63. Ma S, Xu Y, Ling F. Preoperative evaluation and influencing factors of sentinel lymph node detection for early breast cancer with contrast-enhanced ultrasonography: What matters. Medicine (Baltimore) 2021;100:e25183. [Crossref] [PubMed]
64. Zhu Y, Fan X, Yang D, Dong T, Jia Y, Nie F. Contrast-Enhanced Ultrasound for Precise Sentinel Lymph Node Biopsy in Women with Early Breast Cancer: A Preliminary Study. Diagnostics (Basel) 2021;11:2104. [Crossref] [PubMed]
65. Yang SL, Tang KQ, Tao JJ, Ao HF, Wan AH, Shen ZY. Clinical study of different parts of ultrasound contrast agent in the diagnosis of sentinel lymph node in breast cancer. Journal of Clinical and Experimental Medicine 2016;15:1773-6.
66. Li Z, Gao Y, Gong H, Feng W, Ma Q, Li J, Lu X, Wang X, Lei J. Different Imaging Modalities for the Diagnosis of Axillary Lymph Node Metastases in Breast Cancer: A Systematic Review and Network Meta-Analysis of Diagnostic Test Accuracy. J Magn Reson Imaging 2023;57:1392-403. [Crossref] [PubMed]
67. Liu CQ, Guo Y, Shi JY, Sheng Y. Late morbidity associated with a tumour-negative sentinel lymph node biopsy in primary breast cancer patients: a systematic review. Eur J Cancer 2009;45:1560-8. [Crossref] [PubMed]
68. Lee B, Lim AK, Krell J, Satchithananda K, Coombes RC, Lewis JS, Stebbing J. The efficacy of axillary ultrasound in the detection of nodal metastasis in breast cancer. AJR Am J Roentgenol 2013;200:W314-20. [Crossref] [PubMed]
69. Rezkallah EMN, Elsaify A, Tin SMM, Elsaify W. Diagnostic Accuracy of Ultrasonography in Axillary Staging in Breast Cancer Patients. J Med Ultrasound 2023;31:293-7. [Crossref] [PubMed]
70. Diaz-Ruiz MJ, Arnau A, Montesinos J, Miguel A, Culell P, Solernou L, Tortajada L, Vergara C, Yanguas C, Salvador-Tarrasón R. Diagnostic Accuracy and Impact on Management of Ultrasonography-Guided Fine-Needle Aspiration to Detect Axillary Metastasis in Breast Cancer Patients: A Prospective Study. Breast Care (Basel) 2016;11:34-9. [Crossref] [PubMed]
71. Ouyang Q, Chen L, Zhao H, Xu R, Lin Q. Detecting metastasis of lymph nodes and predicting aggressiveness in patients with breast carcinomas. J Ultrasound Med 2010;29:343-52. [Crossref] [PubMed]
72. Pang W, Zhu Y, Nie F. Research progress of percutaneous contrast-enhanced ultrasound in the diagnosis of sentinel lymph node. Zhongliu Yingxiangxue 2023;32:189-93.
73. Li Z, Ma Q, Gao Y, Qu M, Li J, Lei J. Diagnostic performance of MRI for assessing axillary lymph node status after neoadjuvant chemotherapy in breast cancer: a systematic review and meta-analysis. Eur Radiol 2024;34:930-42. [Crossref] [PubMed]
74. Roberto S, Valeria B. Roberto del V, Raffaella M, Chiara FA, Leopoldo R. Analysis by high resolution ultrasound of superficial lymph nodes: anatomical, morphological and structural variations. Clin Imaging 2014;38:96-9. [Crossref] [PubMed]
75. Xiao L, Zhou J, Tan W, Liu Y, Zheng H, Wang G, Zheng W, Pei X, Yang A, Liu L. Contrast-enhanced US with Perfluorobutane to Diagnose Small Lateral Cervical Lymph Node Metastases of Papillary Thyroid Carcinoma. Radiology 2023;307:e221465. [Crossref] [PubMed]
76. Zhang X, Liu Y, Luo H, Zhang J. PET/CT and MRI for Identifying Axillary Lymph Node Metastases in Breast Cancer Patients: Systematic Review and Meta-Analysis. J Magn Reson Imaging 2020;52:1840-51. [Crossref] [PubMed]
77. Meng Y, Ward S, Cooper K, Harnan S, Wyld L. Cost-effectiveness of MRI and PET imaging for the evaluation of axillary lymph node metastases in early stage breast cancer. Eur J Surg Oncol 2011;37:40-6. [Crossref] [PubMed]
78. Boughey JC, Moriarty JP, Degnim AC, Gregg MS, Egginton JS, Long KH. Cost modeling of preoperative axillary ultrasound and fine-needle aspiration to guide surgery for invasive breast cancer. Ann Surg Oncol 2010;17:953-8. [Crossref] [PubMed]