Prognostic value of ultrasound Breast Imaging Reporting and Data System for disease-free survival in early-stage breast cancer: a retrospective study (2011–2018)

Introduction

Breast cancer (BC) is the most commonly diagnosed malignancy in women globally (1). Owing to advances in screening and treatment, the outcomes of early-stage BC have improved substantially. To guide individualized treatment, a number of well-established prognostic factors are routinely assessed, including patient age at diagnosis, axillary lymph node (ALN) involvement, presence of lymphovascular invasion (LVI), tumor histologic or nuclear grade, hormone receptor (estrogen and progesterone) status, human epidermal growth factor receptor 2 (HER2) expression levels, and Ki-67 proliferation index (2-4). In current standard care, surgical resection is typically performed, followed by adjuvant therapies such as chemotherapy, anti-HER2 therapy, radiotherapy, and endocrine therapy. Recently, cryotherapy has emerged as a surgical de-escalation strategy for patients with small, solid breast tumors (5).

Ultrasound (US), mammography (MG), and magnetic resonance imaging (MRI) serve as the primary preoperative diagnostic modalities for BC. Although conventional imaging examinations focus on diagnostic accuracy, growing evidence suggests that imaging features of the primary tumor are associated with histopathological characteristics (6-8). For instance, MG frequently reveals spiculated margins in luminal A BCs, a feature observed significantly more frequently in this subtype than in others (9,10). In contrast, triple-negative breast cancer (TNBC), which typically exhibits aggressive biological behavior, commonly presents smooth margins, absence of calcification, and increased hardness on US (11,12). However, current evidence is limited by a lack of sufficient long-term follow-up data, precluding the establishment of direct correlations between imaging characteristics and the biological outcomes of BC.

The Breast Imaging Reporting and Data System (BI-RADS) classifies suspicious lesions based on the presence of malignant features. The categories 3, 4A, 4B, 4C, and 5 correspond to distinct probabilities of malignancy: ≤2%, >2% to ≤10%, >10% to ≤50%, >50% to <95%, and ≥95%, respectively. Patients with low-risk categories typically qualify for imaging follow-up, whereas those with high-risk categories often require biopsy for pathological confirmation to guide treatment. A previous US study reported that lesions categorized as BI-RADS 4A are associated with a higher risk of recurrence compared to those in categories 4B–5 (13). Currently, there is no widely accepted consensus on whether BI-RADS category can reliably predict aggressive tumor biology.

Therefore, this study was designed to evaluate the potential of US BI-RADS category as a prognostic imaging phenotype marker for early-stage BC. Furthermore, we combined US BI-RADS categories with established prognostic factors to construct an exploratory nomogram for predicting 5‑ and 10‑year disease‑free survival (DFS), and compared its performance with a clinical model based on standard clinicopathological factors. We present this article in accordance with the STROBE reporting checklist (available at https://qims.amegroups.com/article/view/10.21037/qims-2026-0506/rc).

Methods

Patients

We retrospectively identified patients with clinically early-stage BC (defined as tumors ≤2 cm) who underwent surgical treatment at Fudan University Shanghai Cancer Center between August 2011 and January 2018 (Figure 1). All patients underwent preoperative examination of breast US. The exclusion criteria included suspicious findings on preoperative axillary US, history of malignancy or anticancer treatment, and incomplete clinicopathological records. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Ethics Committee of Fudan University Shanghai Cancer Center (No. 1612167-18). Given its retrospective design, the requirement for informed consent was waived.

Figure 1 Flowchart of the patient selection. Early‑stage breast cancer was defined as tumors ≤2 cm. BI-RADS, Breast Imaging Reporting and Data System; US, ultrasound.

Data collection

Clinicopathological and imaging data were retrieved from the institutional electronic medical records. Collected variables included: patient age at diagnosis, personal history of BC, BC treatments (types of surgery and adjuvant treatment), preoperative breast and axillary US findings, and MG BI-RADS category. Pathological reports provided information on tumor size, histological type and grade, ALN status, LVI, and expression status of estrogen receptor (ER), progesterone receptor (PR), HER2, and Ki-67 index.

Variable definitions

Preoperative axillary US findings were deemed suspicious if lymph nodes exhibited any malignant features. Histological grading of invasive BC followed the Scarff-Bloom-Richardson system, whereas ductal carcinoma in situ was graded based on nuclear morphology (7). ER/PR-positivity was defined as ≥1% nuclear staining (14). HER2-positivity was defined as an immunohistochemistry score of 3+ or 2+ with confirmation by fluorescence in situ hybridization (FISH) (15). A Ki-67 index ≥20% indicated high proliferative activity (16). According to the international guidelines, BC molecular subtypes are classified into four categories: Luminal A, Luminal B (including HER2-positive and HER2-negative), HER2 overexpressing, and TNBC (17).

Follow-up

Follow-up data were obtained from the institutional cancer registry. Recurrence was defined as loco-regional recurrence (ipsilateral or contralateral breast, chest wall, or regional lymph nodes) or distant metastasis.

Overall survival (OS) and DFS were the primary endpoints. OS was defined as the time from surgery to death from any cause; DFS as the time from surgery to first event of recurrence or death from any cause. Surviving patients were censored at the administrative study end (June 2025).

Statistical analysis

Continuous variables were summarized as median with interquartile range (IQR) and categorized based on clinical relevance; categorical variables were presented as frequencies (%). Group comparisons for continuous variables were assessed using the Mann-Whitney U test. For categorical variables, χ² or Fisher’s exact tests were used as appropriate. Survival curves were generated using the Kaplan-Meier method and compared via log-rank tests. Missing data were handled using multiple imputation. Univariate and multivariate Cox proportional hazards models were used to identify independent predictors of OS and DFS, with results expressed as hazard ratios (HRs) and 95% confidence intervals (CIs). The proportional hazards assumption was assessed using Schoenfeld residuals. In cases where the assumption was violated, the reported HR was interpreted as the weighted average of the time-varying HR over the entire follow-up period (18).

Spearman’s correlation analysis was performed to identify associations among variables. Given the limited number of DFS events and their established prognostic importance, we included only US BI‑RADS category, lymph node metastasis (LNM), and Ki‑67 in the multivariable model and subsequently used them to construct a nomogram to predict 5- and 10-year DFS. We also built a clinical model using standard clinicopathological factors including LNM, Ki‑67, and histological grade. The discriminatory performance of both models was assessed using the bootstrap-corrected concordance index (C‑index) and the area under the receiver operating characteristic (ROC) curve (AUC). Calibration curves were constructed to compare the predicted 5- and 10-year DFS with the observed outcomes, respectively. All statistical analyses were conducted using RStudio (version 2025.5.0; Posit Software, PBC, Boston, MA, USA). All statistical tests were two-sided, with P<0.05 considered significant.

Results

Patient characteristics

A total of 681 patients with early-stage BC (age range, 28–85 years; median age, 51 years; IQR, 44–60) were finally included (Figure 1). Of the 681 lesions, 4 (0.6%) were classified as BI-RADS category 3, 137 (20.1%) as category 4A, 192 (28.2%) as category 4B, 168 (24.7%) as category 4C, and 180 (26.4%) as category 5 at US. The median pathologic tumor size was 1.5 cm (range, 1.3–2.0 cm), and the distribution of pathological T stage was as follows: Tis, 35 (5.1%); T1a, 2 (0.3%); T1b, 101 (14.8%); T1c, 543 (79.8%). Most cancers were invasive ductal carcinoma (88.6%), ALN-negative (79.6%), without LVI (80.9%), of low or intermediate histological grade (59.8%), luminal B-like (47.1%), ER-positive (74.6%), HER2-negative (78.6%), and high Ki-67 proliferation (64.2%). Some 249 (36.6%) patients underwent breast-conserving surgery, whereas 432 (63.4%) underwent mastectomy. Regarding adjuvant therapy, 469 (68.9%) received endocrine therapy, 236 (34.7%) received radiotherapy, 449 (65.9%) received chemotherapy, and 70 (10.3%) received anti‑HER2 therapy (Table 1). Among them, 408 patients (59.9%) received more than one type of adjuvant therapy.

Based on US BI-RADS category, 333 lesions (48.9%) were classified as non-high-suspicion (categories 3–4B) and 348 (51.1%) as high-suspicion (categories 4C–5). Table 1 compares the clinicopathological characteristics between the two groups. Patients in the high-suspicion group (HSG) were significantly older than those in the non-high-suspicion group (non-HSG) (P<0.001) and had larger tumors (P<0.001). MG BI-RADS categories also differed substantially: the non-HSG showed a higher proportion of low-risk findings (categories 4A–4B) (38.7% vs. 23.3%, P<0.001), whereas the HSG demonstrated a greater frequency of high-risk lesions (categories 4C–5) (52.9% vs. 9.6%, P<0.001). Non-HSG patients had significantly higher proportions of breast-conserving surgery (43.8% vs. 29.6%, P<0.001) and adjuvant radiotherapy (39.0% vs. 30.5%, P=0.019) than did HSG patients. Although the proportion of invasive BC was significantly higher in the HSG (98.3% vs. 91.3%, P<0.001), the proportion of high-grade tumors in this group was comparatively lower (32.8% vs. 37.5%, P<0.001) (Table 1).

No significant differences were observed in traditional prognostic factors, including axillary lymph node metastasis (ALNM) (P=0.088), LVI (P=0.062), HER2 status (P=0.881), or Ki-67 index (P=0.596). Similarly, no significant differences were found between the two groups in the administration of chemotherapy (P=0.461) or anti‑HER2 therapy (P=0.655). However, ER status differed significantly (83.6% vs. 65.2%, P<0.001), and endocrine therapy was correspondingly more frequent in HSG patients (77.0% vs. 60.4%, P<0.001). Among HSG patients, 196 out of 348 (56.3%) were classified as luminal B-like (P<0.001). Compared to the HSG, the non‑HSG contained a higher proportion of TNBC (26.2% vs. 9.5%, P<0.001) (Table 1).

Patient outcomes

The median follow-up period was 8.1 years (IQR, 6.9–9.8 years) for the non-HSG and 8.5 years (IQR, 6.9–10.0 years) for the HSG. Loco-regional recurrence or distant metastasis occurred more frequently in HSG patients (21/348 vs. 8/333, P=0.022). Among these 29 recurrences, 25 were loco-regional, 9 were distant, and 5 patients had concurrent loco-regional recurrence and distant metastasis. No significant difference was observed in mortality between the groups (20/348 vs. 10/333, P=0.093) (Table 1).

Factors associated with patient outcomes

Univariate and multivariate Cox regression analyses were performed to assess DFS in the entire patient cohort. The univariate analysis identified several factors that were significantly associated with DFS, including US BI-RADS category, histologic or nuclear grade, and Ki-67 index. Multivariate analysis identified US BI-RADS categories 4C–5 (HR =2.56; 95% CI: 1.32–4.99; P=0.006) and Ki-67 index ≥20% (HR =3.08; 95% CI: 1.37–6.91; P=0.006) as independent predictors of less favorable DFS (Table 2).

For OS, univariate analysis indicated that Ki-67 index ≥20% and grade 3 were significantly associated with worse OS. No independent predictor was identified in the multivariable analysis (Table 3).

Survival analysis

Among 681 women with early-stage BC, the 5-year OS and DFS rates were 97.5% (95% CI: 96.3–98.7%) and 95.6% (95% CI: 94.0–97.2%), respectively. At 10 years, the corresponding OS and DFS rates were 94.9% (95% CI: 93.0–96.8%) and 92.9% (95% CI: 90.8–95.0%), respectively.

Survival curves stratified by US BI-RADS category showed that non-HSG patients demonstrated significantly superior DFS outcomes than HSG patients at both 5 years (97.5% vs. 93.7%) and 10 years (95.5% vs. 90.3%) (P=0.004) (Figure 2). In contrast, no significant difference in OS was observed between the groups at 5 years (98.1% vs. 97.0%) or 10 years (96.4% vs. 93.5%) (P=0.098) (Figure 3).

Figure 2 Kaplan-Meier survival curves for DFS across different US BI-RADS categories. BI-RADS, Breast Imaging Reporting and Data System; DFS, disease-free survival; US, ultrasound.

Figure 3 Kaplan-Meier survival curves for OS across different US BI-RADS categories. BI-RADS, Breast Imaging Reporting and Data System; OS, overall survival; US, ultrasound.

Construction and evaluation of the nomogram model

The heatmap demonstrated the association of US BI‑RADS category, LNM, and Ki‑67 with DFS status (Figure 4A). Spearman’s correlation analysis showed no significant collinearity among US BI-RADS category, LNM, and Ki-67 (Figure 4B). In our US-based combined nomogram (Figure 4C), Ki-67 index held the highest predictive weight, followed by the US BI-RADS category and LNM. The model achieved a C-index of 0.701 (95% CI: 0.611–0.770), and its calibration curves showed excellent agreement for 5- and 10-year DFS (Figure 4D). The AUC of this nomogram was 0.720 (95% CI: 0.624–0.801) at 5 years and 0.647 (95% CI: 0.549–0.738) at 10 years (Figure 5A). In contrast, the clinical model achieved a C‑index of 0.649 (95% CI: 0.556–0.728), with a 5‑year AUC of 0.691 (95% CI: 0.553–0.790) and a 10‑year AUC of 0.617 (95% CI: 0.523–0.703) (Figure 5B). Figure 6 shows representative examples of DFS prediction for individual early-stage BC patients using our US‑based nomogram.

Figure 4 US-based nomogram performance and variable associations. (A) The heatmap depicts the correlation between DFS status and key clinicopathological characteristics, including US BI-RADS category, LNM, and Ki-67 index. (B) Spearman’s correlation analysis among candidate variables. (C) US-based nomogram for predicting 5- and 10-year DFS. (D) Calibration curves for predicting 5- and 10-year DFS. BI-RADS, Breast Imaging Reporting and Data System; DFS, disease-free survival; LNM, lymph node metastasis; US, ultrasound.

Figure 5 Time‑dependent AUC of the model for predicting 5‑ and 10‑year DFS. (A) US‑based model. (B) Clinical model. AUC, area under the curve; CI, confidence interval; DFS, disease-free survival; US, ultrasound.

Figure 6 Representative examples of DFS prediction using the US‑based nomogram. (A) US image of a BI‑RADS 4A lesion (Ki‑67 80%, node‑negative). (B) Nomogram prediction for case A: 5-/10-year DFS 97.1%/95.3%. (C) US image of a BI‑RADS 4C lesion (Ki‑67 25%, node‑positive). (D) Nomogram prediction for case B: 5-/10-year DFS 88.3%/81.6%. BI-RADS, Breast Imaging Reporting and Data System; DFS, disease-free survival; LNM, lymph node metastasis; US, ultrasound.

Discussion

In this retrospective study, we analyzed a cohort of 681 women diagnosed with clinically early-stage BC. Consistent with previous studies, the majority of tumors were ALN-negative invasive ductal carcinoma with low or intermediate grade, HR-positivity, and HER2-negativity (13). Patients with small tumors in our cohort generally achieved satisfactory survival outcomes with surgical resection and appropriate adjuvant therapy.

Our comprehensive analysis identified US BI-RADS categories 4C–5 as independent predictors of less favorable DFS. This suggests that patients presenting with these imaging phenotypes may warrant closer surveillance and consideration of more intensive treatment. HSG patients exhibited significantly higher recurrence rates and worse DFS compared with non-HSG patients. These results suggest more aggressive disease biology or advanced presentation in the HSG population at baseline. Notably, despite these differences in recurrence patterns, no significant difference in OS was observed between the HSG and non‑HSG following appropriate treatment. This observation highlights the effectiveness of current treatment protocols in mitigating the initial prognostic disadvantage observed in the HSG cohort. The lack of significant OS difference suggests that timely detection and management of recurrences may have compensated for the higher recurrence risk in the HSG.

These findings have important implications for clinical practice, particularly in risk stratification and patient counseling. The identification of US BI-RADS categories 4C–5 as prognostic imaging phenotype markers may be useful for preoperative risk assessment and adjuvant treatment consideration. The differential DFS outcomes between groups may warrant consideration of more intensive surveillance protocols for high-risk patients, whereas the equivalent OS provides reassurance regarding the ultimate effectiveness of current treatment paradigms.

Although patients with suspicious axillary US findings were excluded from the study, postoperative pathology confirmed ALNM in 139 patients (20.4%), demonstrating the limitations of axillary US in accurately assessing ALN status. This finding aligns with findings reported in previous literature (19,20). Previous studies have demonstrated that the US features of primary breast tumors also hold significant predictive value for ALNM (21,22). Recent advances involving deep learning technology integrating imaging and clinical features have shown improved accuracy in predicting ALN status (23,24), which warrants further investigation into the relationship between US BI-RADS category and ALNM.

In our study, 17.2% of the lesions were classified as BI-RADS category 0 on MG, reflecting a higher rate of incomplete mammographic assessment that required additional imaging. This limitation was particularly evident in women with dense breast tissue or those with a history of breast surgery, as supported by previous research findings (25-27). Consequently, US may serve as a useful complementary diagnostic tool for this particular patient population. It is important to note that MG maintains its clinical value by demonstrating relatively high sensitivity in detecting calcifications which are usually missed by US examination (28). Multiple studies confirm that combining breast US and MG is an efficient, cost-effective BC diagnostic strategy for diverse populations (29,30). This integrated approach may also help reduce interval BCs (31).

Interestingly, despite being associated with better DFS, the non-HSG contained a higher prevalence of TNBC and grade 3 tumors. This may be explained by the characteristic US presentation of TNBC: despite frequently being high-grade, these tumors often exhibit benign-like US features: circumscribed margins, oval shape, and posterior acoustic enhancement. These deceptive morphological characteristics often result in lower BI-RADS scores (32-34). The higher recurrence risk of BI-RADS 4A lesions reported by Kim et al. may also be attributable to the predominance of TNBC within this category (13). Although TNBC is well-established as the molecular subtype with poorest overall prognosis among BCs (35), emerging clinical evidence suggests that timely diagnosis combined with appropriate systemic chemotherapy regimens can significantly improve outcomes of TNBC patients (36,37). In our cohort, the administration of chemotherapy showed no significant difference between the HSG and non-HSG, which may have contributed to favorable survival in small TNBCs. These findings collectively emphasize the crucial need for both improved diagnostic strategies to overcome the imaging challenges posed by TNBC and the development of subtype-specific treatment protocols.

In our cohort, the non-HSG (BI-RADS 3–4B) had a significantly higher proportion of TNBC (26.2% vs. 9.5%) and grade 3 tumors, yet paradoxically showed better DFS. This apparent paradox actually reinforces our main conclusion: the low-suspicion imaging phenotype (lower BI-RADS score) appears to reflect a less aggressive tumor biology that may override the inherent high-risk nature of TNBC. In other words, even among TNBCs, those with benign-looking US features carry a better prognosis, which aligns with growing evidence that imaging phenotype correlate with tumor biological properties (38,39).

Some limitations should be considered when interpreting the results of our study. First, as a retrospective single-center investigation, the findings may be influenced by selection bias. Second, the small number of DFS events limited statistical power and prevented us from adjusting for adjuvant therapies. Third, variability in radiologists’ interpretation of BI-RADS criteria may reduce reliability. Finally, the small number of BI-RADS category 3 lesions limited comprehensive analysis across all categories. Future studies with larger, multi-center cohorts are needed to validate these findings.

Conclusions

In this exploratory study, preoperative US BI‑RADS category was associated with DFS in early‑stage BC. Compared with a clinical model based on standard clinicopathological factors, our nomogram integrating this imaging marker demonstrated improved predictive performance. These preliminary findings suggest that the BI-RADS system may be useful beyond diagnosis, potentially aiding in prognostic stratification and treatment planning. However, external validation in larger cohorts is needed before clinical application.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://qims.amegroups.com/article/view/10.21037/qims-2026-0506/rc

Data Sharing Statement: Available at https://qims.amegroups.com/article/view/10.21037/qims-2026-0506/dss

Funding: None.

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://qims.amegroups.com/article/view/10.21037/qims-2026-0506/coif). The authors have no 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. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Ethics Committee of Fudan University Shanghai Cancer Center (No. 1612167-18). Given its retrospective design, the requirement for informed consent was waived.

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. Siegel RL, Kratzer TB, Giaquinto AN, Sung H, Jemal A. Cancer statistics, 2025. CA Cancer J Clin 2025;75:10-45. [Crossref] [PubMed]
2. Partridge AH, Hughes ME, Warner ET, Ottesen RA, Wong YN, Edge SB, Theriault RL, Blayney DW, Niland JC, Winer EP, Weeks JC, Tamimi RM. Subtype-Dependent Relationship Between Young Age at Diagnosis and Breast Cancer Survival. J Clin Oncol 2016;34:3308-14. [Crossref] [PubMed]
3. Elston CW, Ellis IO. Pathological prognostic factors in breast cancer. I. The value of histological grade in breast cancer: experience from a large study with long-term follow-up. Histopathology 1991;19:403-10.
4. Houvenaeghel G, Cohen M, Classe JM, Reyal F, Mazouni C, Chopin N, et al. Lymphovascular invasion has a significant prognostic impact in patients with early breast cancer, results from a large, national, multicenter, retrospective cohort study. ESMO Open 2021;6:100316. [Crossref] [PubMed]
5. Lee KL, Sule AA, Lai HX, Ng QX, Goh SSN. Cryotherapy as a Surgical De-Escalation Strategy in Breast Cancer: Techniques, Complications, and Oncological Outcomes. Biomedicines 2025;13:2987. [Crossref] [PubMed]
6. Boisserie-Lacroix M, Bullier B, Hurtevent-Labrot G, Ferron S, Lippa N, Mac Grogan G. Correlation between imaging and prognostic factors: molecular classification of breast cancers. Diagn Interv Imaging 2014;95:227-33. [Crossref] [PubMed]
7. Kim SH, Seo BK, Lee J, Kim SJ, Cho KR, Lee KY, Je BK, Kim HY, Kim YS, Lee JH. Correlation of ultrasound findings with histology, tumor grade, and biological markers in breast cancer. Acta Oncol 2008;47:1531-8. [Crossref] [PubMed]
8. Tamaki K, Ishida T, Miyashita M, Amari M, Ohuchi N, Tamaki N, Sasano H. Correlation between mammographic findings and corresponding histopathology: potential predictors for biological characteristics of breast diseases. Cancer Sci 2011;102:2179-85. [Crossref] [PubMed]
9. Killelea BK, Chagpar AB, Bishop J, Horowitz NR, Christy C, Tsangaris T, Raghu M, Lannin DR. Is there a correlation between breast cancer molecular subtype using receptors as surrogates and mammographic appearance? Ann Surg Oncol 2013;20:3247-53. [Crossref] [PubMed]
10. Liu S, Wu XD, Xu WJ, Lin Q, Liu XJ, Li Y. Is There a Correlation between the Presence of a Spiculated Mass on Mammogram and Luminal A Subtype Breast Cancer? Korean J Radiol 2016;17:846-52. [Crossref] [PubMed]
11. Men DX, Li HZ, Dong J, Xue MH, Wang ZF, Xiao WL, Xue JP, Jia MH. Correlation between ultrasonography and elastography parameters and molecular subtypes of breast cancer in young women. Ann Med 2025;57:2443041. [Crossref] [PubMed]
12. Wang C, Zhao X, Wang D, Wu J, Lin J, Huang W, Shen Y, Chen Q. Analysis of prognostic factors and nomogram construction for postoperative survival of triple-negative breast cancer. Front Immunol 2025;16:1561563. [Crossref] [PubMed]
13. Kim SY, Han BK, Kim EK, Choi WJ, Choi Y, Kim HH, Moon WK. Breast Cancer Detected at Screening US: Survival Rates and Clinical-Pathologic and Imaging Factors Associated with Recurrence. Radiology 2017;284:354-64. [Crossref] [PubMed]
14. Hammond ME, Hayes DF, Dowsett M, Allred DC, Hagerty KL, Badve S, et al. American Society of Clinical Oncology/College Of American Pathologists guideline recommendations for immunohistochemical testing of estrogen and progesterone receptors in breast cancer. J Clin Oncol 2010;28:2784-95. [Crossref] [PubMed]
15. Wolff AC, Hammond ME, Schwartz JN, Hagerty KL, Allred DC, Cote RJ, et al. American Society of Clinical Oncology/College of American Pathologists guideline recommendations for human epidermal growth factor receptor 2 testing in breast cancer. J Clin Oncol 2007;25:118-45. [Crossref] [PubMed]
16. Stuart-Harris R, Caldas C, Pinder SE, Pharoah P. Proliferation markers and survival in early breast cancer: a systematic review and meta-analysis of 85 studies in 32,825 patients. Breast 2008;17:323-34. [Crossref] [PubMed]
17. Goldhirsch A, Winer EP, Coates AS, Gelber RD, Piccart-Gebhart M, Thürlimann B, Senn HJ. Personalizing the treatment of women with early breast cancer: highlights of the St Gallen International Expert Consensus on the Primary Therapy of Early Breast Cancer 2013. Ann Oncol 2013;24:2206-23. [Crossref] [PubMed]
18. Stensrud MJ, Hernán MA. Why Test for Proportional Hazards? JAMA 2020;323:1401-2. [Crossref] [PubMed]
19. Rattay T, Muttalib M, Khalifa E, Duncan A, Parker SJ. Clinical utility of routine pre-operative axillary ultrasound and fine needle aspiration cytology in patient selection for sentinel lymph node biopsy. Breast 2012;21:210-4. [Crossref] [PubMed]
20. Kwak HY, Chae BJ, Bae JS, Kim EJ, Chang EY, Kim SH, Jung SS, Song BJ. Feasibility of sentinel lymph node biopsy in breast cancer patients clinically suspected of axillary lymph node metastasis on preoperative imaging. World J Surg Oncol 2013;11:104. [Crossref] [PubMed]
21. Xiong J, Zuo W, Wu Y, Wang X, Li W, Wang Q, Zhou H, Xie M, Qin X. Ultrasonography and clinicopathological features of breast cancer in predicting axillary lymph node metastases. BMC Cancer 2022;22:1155. [Crossref] [PubMed]
22. Guo Q, Dong Z, Zhang L, Ning C, Li Z, Wang D, Liu C, Zhao M, Tian J. Ultrasound Features of Breast Cancer for Predicting Axillary Lymph Node Metastasis. J Ultrasound Med 2018;37:1345-53.
23. Zheng X, Yao Z, Huang Y, Yu Y, Wang Y, Liu Y, Mao R, Li F, Xiao Y, Wang Y, Hu Y, Yu J, Zhou J. Deep learning radiomics can predict axillary lymph node status in early-stage breast cancer. Nat Commun 2020;11:1236. [Crossref] [PubMed]
24. Zhou LQ, Wu XL, Huang SY, Wu GG, Ye HR, Wei Q, Bao LY, Deng YB, Li XR, Cui XW, Dietrich CF. Lymph Node Metastasis Prediction from Primary Breast Cancer US Images Using Deep Learning. Radiology 2020;294:19-28. [Crossref] [PubMed]
25. van Breest Smallenburg V, Duijm LE, Voogd AC, Groenewoud JH, Jansen FH, van Beek M, Louwman MW. Lower sensitivity of screening mammography after previous benign breast surgery. Int J Cancer 2012;130:122-8. [Crossref] [PubMed]
26. Nakamura M, Ishizuka Y, Horimoto Y, Shiraishi A, Arakawa A, Yanagisawa N, Iijima K, Saito M. Clinicopathological features of breast cancer without mammographic findings suggesting malignancy. Breast 2020;54:335-42. [Crossref] [PubMed]
27. Di Grezia G, Nazzaro A, Schiavone L, Elisa C, Galiano A, Gianluca G, Vincenzo C, Scaglione M. From Variability to Standardization: The Impact of Breast Density on Background Parenchymal Enhancement in Contrast-Enhanced Mammography and the Need for a Structured Reporting System. Cancers (Basel) 2025;17:2523. [Crossref] [PubMed]
28. Zhang J, Cai L, Pan X, Chen L, Chen M, Yan D, Liu J, Luo L. Comparison and risk factors analysis of multiple breast cancer screening methods in the evaluation of breast non-mass-like lesions. BMC Med Imaging 2022;22:202. [Crossref] [PubMed]
29. Lee JM, Arao RF, Sprague BL, Kerlikowske K, Lehman CD, Smith RA, Henderson LM, Rauscher GH, Miglioretti DL. Performance of Screening Ultrasonography as an Adjunct to Screening Mammography in Women Across the Spectrum of Breast Cancer Risk. JAMA Intern Med 2019;179:658-67. [Crossref] [PubMed]
30. Xu Z, Lin Y, Huo J, Gao Y, Lu J, Liang Y, Li L, Jiang Z, Du L, Lang T, Wen G, Li Y. A bimodal nomogram as an adjunct tool to reduce unnecessary breast biopsy following discordant ultrasonic and mammographic BI-RADS assessment. Eur Radiol 2024;34:2608-18. [Crossref] [PubMed]
31. Goh RSJ, Chong B, Yeo S, Neo SY, Ng QX, Goh SSN. Current Strategies to Reducing Interval Breast Cancers: A Systematic Review. Breast Cancer (Dove Med Press) 2025;17:531-44. [Crossref] [PubMed]
32. Du Y, Zha HL, Wang H, Liu XP, Pan JZ, Du LW, Cai MJ, Zong M, Li CY. Ultrasound-based radiomics nomogram for differentiation of triple-negative breast cancer from fibroadenoma. Br J Radiol 2022;95:20210598. [Crossref] [PubMed]
33. Yoon GY, Cha JH, Kim HH, Shin HJ, Chae EY, Choi WJ. Sonographic features that can be used to differentiate between small triple-negative breast cancer and fibroadenoma. Ultrasonography 2018;37:149-56. [Crossref] [PubMed]
34. Sheng DL, Shen XG, Shi ZT, Chang C, Li JW. Survival outcome assessment for triple-negative breast cancer: a nomogram analysis based on integrated clinicopathological, sonographic, and mammographic characteristics. Eur Radiol 2022;32:6575-87. [Crossref] [PubMed]
35. Bianchini G, De Angelis C, Licata L, Gianni L. Treatment landscape of triple-negative breast cancer - expanded options, evolving needs. Nat Rev Clin Oncol 2022;19:91-113. [Crossref] [PubMed]
36. Liu X, Zhuang C, Liu L, Xiong L, Xie X, He P, et al. Exploratory phase II trial of an anti-PD-1 antibody camrelizumab combined with a VEGFR-2 inhibitor apatinib and chemotherapy as a neoadjuvant therapy for triple-negative breast cancer (NeoPanDa03): efficacy, safety and biomarker analysis. Signal Transduct Target Ther 2025;10:237. [Crossref] [PubMed]
37. Glück S, Ross JS, Royce M, McKenna EF Jr, Perou CM, Avisar E, Wu L. TP53 genomics predict higher clinical and pathologic tumor response in operable early-stage breast cancer treated with docetaxel-capecitabine ± trastuzumab. Breast Cancer Res Treat 2012;132:781-91. [Crossref] [PubMed]
38. Tong YY, Sun PX, Zhou J, Shi ZT, Chang C, Li JW. The Association Between Ultrasound Features and Biological Properties of Invasive Breast Carcinoma Is Modified by Age, Tumor Size, and the Preoperative Axilla Status. J Ultrasound Med 2020;39:1125-34. [Crossref] [PubMed]
39. Li JW, Tong YY, Zhou J, Shi ZT, Sun PX, Chang C. Tumor Proliferation and Invasiveness Derived From Ultrasound Appearances of Invasive Breast Cancers: Moving Beyond the Routine Differential Diagnosis. J Ultrasound Med 2020;39:1589-99. [Crossref] [PubMed]