Prognostic factors for hormone receptor-positive breast cancer with liver metastasis and establishment of novel nomograms for prediction: a SEER-based study

Introduction

According to the Cancer Statistics 2022 report, breast cancer (BC) is the leading malignancy and primary cause of mortality in women (1). It was estimated that approximately 290,560 American women would be diagnosed with BC in 2022 and 43,000 would die from it (1). Distant metastasis (rather than the primary tumor) is the most frequently lethal to patients with BC (2). Common sites of BC metastasis include the bone, lungs, and liver, among which, the liver is the third most common site (3). Previous studies have shown that more than 30% of patients with BC develop non-lymph node metastasis (4,5). Compared to those with bone and lung metastases, BC patients with liver metastasis have a worse prognosis and a median survival time of only 14 to 16 months, regardless of whether systemic therapy is administered for primary/metastatic sites, which may be related to endocrine therapy resistance and the failure of chemotherapy (6,7). Further, when BC with liver metastasis (BCLM) initially appears, it may result in digestive symptoms, including nausea and anorexia, as well as signs and symptoms of cachexia, including severe hepatomegaly, jaundice, and ascites, which have significant negative effects on the quality of life of patients (8,9).

Thus, a unique predictive diagnostic tool for BCLM urgently needs to be established for clinical practice. The tumor-node-metastasis (TNM) staging system is one of the most widely used conventional prediction tools; however, it has certain drawbacks that prevent it from adequately covering cancer biology and quantifying the prognosis of patients with distant metastatic disease (10). In addition, while some studies have attempted to simulate the prognosis of BCLM patients, they have failed to sufficiently focus on the molecular type (7,11).

Hormone receptor (HR) status, which includes the estrogen receptor (ER) and progesterone receptor, is a vital prognostic factor, and the HR-positive subtype is the most common type of BC (12,13). The prognosis of BC patients varies depending on their HR status; patients with the HR-positive/human epidermal growth factor receptor 2 (HER2)-positive subtype have a better prognosis than those with the HR-negative/HER2-negative subtype (14). To the best of our knowledge, no previous study exists on the prognostic prediction of patients with HR-positive BCLM.

Nomograms can be used to help construct prognostic models to aid in clinical decision making. A nomogram is a solid instrument for quantifying the risk of variables in prognostic models and is broadly used in prognostic analysis in oncology (15). In this study, we identified a representative cohort from the Surveillance, Epidemiology and End Results (SEER) population-based national registration database to evaluate the prognosis of HR-positive de novo liver metastatic BC, and using a large sample size, we established and validated two nomograms to predict overall survival (OS) and breast cancer-specific survival (BCSS). We present this article in accordance with the TRIPOD reporting checklist (available at https://tcr.amegroups.com/article/view/10.21037/tcr-23-874/rc).

Methods

Study population

Using the SEER*Stat program version 8.4.0.1, we obtained HR-positive BCLM data from the SEER database, which comprises one of the largest publicly available cancer data sets. The population included in this analysis was between 2010 and 2015, as the SEER database only started gathering data on the molecular subtypes and locations of distant metastases in 2010.

To be eligible for inclusion in the study, the patients had to meet the following inclusion criteria: (I) be an adult female; (II) have been diagnosed by immunohistochemistry; (III) have liver metastasis; and (IV) be HR-positive. Patients were excluded from the study if they met any of the following exclusion criteria: (I) BC was not the only or was not the first diagnosis of two or more primary cancers; (II) essential details, including TNM stage, surgery, information on metastatic sites, HER2 status or survival months, were unavailable; and/or (III) the follow-up duration was <1 month.

Ultimately, a total of 1,780 eligible patients passed the screening process in the study and were randomly categorized into the following two groups using R software: (I) the training group, which comprised 1,246 (70%) patients and was used to develop a nomogram; and (II) the validation group, which comprised 534 (30%) patients and was used to verify it. Moreover, to establish and additional validation cohort for our research, using a similar method, we identified 312 patients in the SEER database who met the criteria in 2016 (see Figure 1).

Figure 1 Flowchart of the study. SEER, Surveillance, Epidemiology and End Results; HR, hormone receptor; TNM, tumor-node-metastasis.

Data from the SEER research database was used in this study. The National Cancer Institute granted us access to the SEER program’s data files (reference number: 17868-Nov2021). This study was conducted in accordance with the Declaration of Helsinki (as revised in 2013). The local ethics committee waived the requirement of ethical approval for this study, as the patient data extracted from the database were non-identifiable and available for public use for research purposes.

Variable selection

The variables in the analysis included age at diagnosis, race, marital status, T stage, N stage, grade, histopathological type, HER2 status, location, surgery, radiotherapy, chemotherapy, bone metastasis, lung metastasis, and brain metastasis. In addition, several variables were grouped into a specific category based on the SEER database classification. Age was classified into three categories (i.e., <50, 50–69, and ≥70 years). Race was stratified as Black, White or other. The patients were allocated to three subgroups based on their marital status (i.e., married, unmarried, or other). The histological categories included duct carcinoma, lobular carcinoma, or other. The locations were classified as inner, outer, or other. The grades were classified into three categories (grades I–II, grades III–IV, and unknown). Finally, the N stage was categorized into two subcategories (N0–N1 and N2–N3).

Statistical analyses

The key results of the analysis were OS and BCSS. BCSS was defined as the period from presentation to death due to BC. OS was defined as the time from diagnosis to death from any cause or the end of follow-up. All the data analyses were performed using R software (version 4.2.1; https://www.r-project.org/).

The features at the baseline were compared using the chi-square test. A univariate Cox regression analysis was used to investigate the prognostic variables associated with both OS and BCSS in patients with HR-positive BCLM. Variables with P values less than 0.05 were considered statistically significant. In addition, a multivariate Cox regression analysis was used to identify independent risk factors.

Based on the separate prognostic indicators identified in the above analyses, nomograms were developed for OS and BCSS at 1, 3, and 5 years. Further, a unique diagnostic nomogram based on independent risk variables was established using the “rms” package. Receiver operating characteristic (ROC) curves of the nomogram and all the independent variables were created to evaluate discrimination. The associated areas under the curve (AUCs) were also determined. As a general rule, AUC values typically run from 0.5 to 1.0, with 0.5 representing random chance and 1.0 representing perfect discriminative power. A calibration plot was also used to evaluate whether the expected outcomes matched the actual findings. Subsequently, the clinical utility of the nomogram was assessed by contrasting the quantitative net benefits under varying probabilities, which were established by the difference between the expected benefit and expected loss in connection with each proceeding project strategy and therapeutic intervention tactic. This was performed using a decision curve analysis (DCA) (16).

Based on the risk ratings calculated from their nomograms, patients from the training and validation data sets were classified into low- and high-risk groups. The efficacy of the nomogram was further investigated using the Kaplan-Meier (K-M) survival curves and log-rank tests to validate whether there was a statistical difference in the survival times between the two data sets. Finally, we used the data of patients from 2016 in the SEER database for the external validation to further predict the model’s reliability.

Results

Baseline characteristics

A total of 1,780 patients with BCLM who met the selection criteria were included in the study. Of these patients, 1,109 patients died during the follow-up period, of whom, 767 were in the training group and 342 were in the validation group. The R software used a 7:3 ratio to categorize all patients into the training (n=1,246) and validation cohorts (n=534). Patients aged 50–69 years (n=920, 51.7%) accounted for approximately half of all patients, followed by those aged <50 years (n=517, 29.0%) and ≥70 years (n=343, 19.3%). The most common T and N stages were T3–T4 (n=959, 53.9%) and N0–N1 (n=1,298, 72.9%), respectively. In relation to treatment, most patients opted for chemotherapy (n=1,205, 67.7%) rather than surgery (n=477, 26.8%), and 502 (28.2%) patients were selected for radiotherapy. During the diagnosis, 63.6% of patients with BCLM had bone metastasis, while only 7.9% [140] of the patients had brain metastasis. Table 1 summarizes the clinical characteristics and demographics of the training and validation groups.

Cox proportional hazards regression analysis

We conducted a univariate analysis using the data of the 1,780 patients with HR-positive BCLM to exclude the associated key factors. P values, hazard ratios, and 95% confidence intervals (CIs) were calculated for each independent parameter. The results of the univariate and multivariate analyses for OS and BCSS are shown in Tables 2,3. Ultimately, nine factors, including age, marital status, grade, HER2 status, chemotherapy, surgery, bone metastasis, lung metastasis, and brain metastasis were found to be significantly associated with OS. Conversely, 10 variables, including age, marital status, T stage, grade, HER2 status, chemotherapy, surgery, bone metastasis, lung metastasis, and brain metastasis, were found to be independent risk factors for BCSS.

Prognostic nomogram development and validation

A nomogram is a user-friendly prediction tool that enables physicians and patients to calculate the ratings of variables to evaluate the likelihood of survival (17). A nomogram for OS and BCSS was established by incorporating the independent prognostic variables (see Figures 2,3). The ROC curve analysis showed that the AUCs of the nomogram for 1-, 2-, and 3-year OS were 0.724, 0.730, and 0.738 in the training group, 0.731, 0.715, and 0.716 in the validation group and 0.809, 0.785, and 0.758 in the external validation cohort, respectively (see Figure 4). For the BCSS model, the AUCs at 1-, 2-, and 3-year were 0.709, 0.719, and 0.724 in the training cohort, 0.775, 0.777, and 0.781 in the validation cohort and 0.807, 0.794, and 0.763 in the external cohort, respectively (see Figure 5). These results demonstrate that the OS and BCSS nomograms had a good discriminatory ability. Additionally, the calibration curves for OS and BCSS in the training, validation, and external groups at 1, 2, and 3 years were almost parallel to the standard curve, demonstrating a strong connection between the predictions of the model and the actual observational findings (see Figures 6,7), respectively. The results of the DCA showed that the nomograms exhibited a greater net benefit in all groups at 1, 2, and 3 years (see Figures 8,9). Further, according to our models, we calculated the prognostic risk score of each patient with HR-positive BCLM. The patients were the categorized into low- and high-risk groups based on their scores. The K-M survival analysis of the training, validation, and external groups showed significant survival differences in both cohorts (see Figures 10,11).

Figure 2 The nomogram to predict 1-, 2-, and 3-year OS of HR-positive BC patients with liver metastasis. **, P<0.01; ***, P<0.001. HER2, human epidermal growth factor receptor 2; OS, overall survival; HR, hormone receptor; BC, breast cancer.

Figure 3 The nomogram to predict 1-, 2-, and 3-year BCSS of HR-positive BC patients with liver metastasis. *, P<0.05; **, P<0.01; ***, P<0.001. HER2, human epidermal growth factor receptor 2; BCSS, breast cancer-specific survival; HR, hormone receptor; BC, breast cancer.

Figure 4 The ROC curves for the OS model for the training (A), internal validation (B), and external (C) cohorts. ROC, receiver operating characteristic; AUC, area under the curve; OS, overall survival.

Figure 5 The ROC curves for the BCSS model for the training (A), internal validation (B) and external (C) cohorts. ROC, receiver operating characteristic; AUC, area under the curve; BCSS, breast cancer-specific survival.

Figure 6 Calibration curves for predicting patients’ OS at 1-, 2-, and 3-year for the training (A-C), validation (D-F), and external (G-I) cohorts. OS, overall survival.

Figure 7 Calibration curves for predicting patients’ BCSS at 1-, 2-, and 3-year for the training (A-C), validation (D-F), and external (G-I) cohorts. BCSS, breast cancer-specific survival.

Figure 8 The DCA of the nomogram for predicting 1-, 2-, and 3-year OS for the training (A-C), validation (D-F), and external (G-I) cohorts. DCA, decision curve analysis; OS, overall survival.

Figure 9 The DCA of the nomogram for predicting 1-, 2-, and 3-year BCSS for the training (A-C), validation (D-F), and external (G-I) cohorts. DCA, decision curve analysis; BCSS, breast cancer-specific survival.

Figure 10 The K-M survival curves of the risk group stratification for OS in the training (A) and validation (B) cohorts and for BCSS in the cohort (C) and validation (D) cohort. K-M, Kaplan-Meier; OS, overall survival; BCSS, breast cancer-specific survival.

Figure 11 The K-M survival curves of the risk group stratification for OS (A) and BCSS (B) in the external cohort. K-M, Kaplan-Meier; OS, overall survival; BCSS, breast cancer-specific survival.

Discussion

BCLM is a heterogeneous illness with various histopathological and molecular characteristics that are linked to various clinical consequences. HR-positive BC, which is dependent on estrogen for growth and survival, accounts for 70–80% of all BCs (18,19). In addition to traditional treatments, such as surgery and chemotherapy, endocrine therapy has emerged as a targeted treatment method for this population subset due to its specific gene expression (20). However, despite dramatic advances in systemic therapy, the prognosis of patients with BCLM remains poor (14). Thus, to advance the research further, we retrospectively analyzed the OS and BCSS of selected patients with liver metastasis from the SEER database, completed the first detailed retrospective analysis using a multicenter cohort and established two nomograms to assist in making treatment decisions for these patients.

Similar to the findings of previous studies (21,22), we found that age was a major independent prognostic indicator, with younger women surviving longer than older women. Some scholars have emphasized that the risk of death associated with BC increases with age, and this is even more pronounced in later years (i.e., among patients aged >70 years) (23). Advanced disease, which is linked to the delayed time to diagnosis, could be an essential contributor to these age-related prognostic disparities (24). In addition, older women present with larger tumors and have higher rates of lymph node metastasis, which may explain their relatively high mortality rates (25-27).

We also found that marriage was an independent protective factor for OS and BCSS in our adult patients with BCLM, which is consistent with the results of a previous meta-analysis by Yuan et al. (28). When symptoms first appear, intimate partners may notice them and advise their spouses to seek immediate medical attention (7,28). Additionally, single patients have been shown to experience more distress, depression, and anxiety than married patients (7,28). The results of previous studies also support this hypothesis. Notably, Shrout et al. reported that at 6 and 18 months after therapy, married women experienced reductions in their depressive symptoms, tension, and exhaustion, regardless of whether they were content or dissatisfied with their marriages. Conversely, depressive symptoms, stress, exhaustion, and pain did not diminish with time in unmarried (i.e., single, divorced/separated, or widowed) women and remained elevated at 6 and 18 months after the conclusion of therapy (29).

As metastatic BC is incurable in almost all afflicted individuals, surgical therapy is typically only employed for alleviation in metastatic illnesses (30). Thus, performing surgery on such patients remains controversial. In our study, patients who underwent surgery at the main cancer location had a relatively long median OS or BCSS. Rashaan et al. theorized that the resection of the malignant tumor results in the restoration of the immune system and the elimination of cancer stem cells, which are believed to oversee the start of tumor development, growth, metastatic properties, and tumor recurrence (31). Another study by Xiong et al. found that the surgical excision of the initial malignancy enhanced the survival of individuals with de novo stage IV BC (32), which is consistent with the results of our study. According to our study, chemotherapy is another crucial predictor and appears to exert a similar effect by decreasing cancer-related complications by eradicating or suppressing cancer cells, postponing relapse, and lengthening survival time. Unfortunately, because our data lacked specific information on treatment strategies, we were unable to examine how various chemotherapy regimens affected survival rates. Radiotherapy did not affect patient prognosis as expected. Indeed, according to a previous study by the Early Breast Cancer Trialists’ Collaborative Group, patients with early-stage invasive BC who receive radiotherapy live longer (33). In addition, while there is some evidence that breast radiotherapy and the quantity of the dosage are independently correlated with survival in patients with metastatic cancer, the incurable nature of metastatic disease appears to outweigh any marginal survival benefit provided by breast radiotherapy (34-36).

Despite previous studies linking HER2 overexpression to a negative prognosis (37), our study showed that HER2 positivity was strongly associated with a favorable outcome. This is due to the long-term benefits of trastuzumab with paclitaxel or docetaxel. In addition to the above-mentioned points, T stage, grade, and metastases (other than liver metastasis) were also identified as significant prognostic factors. In our analysis, T stage was a separate predictive factor for BCSS but not for OS. However, this result may be related to the small amount of available data.

To the best of our knowledge, this was the first study to develop nomograms to predict the prognosis of HR-positive BCLM patients. We developed nomograms based on the Cox proportional hazards model to predict survival. The calibration plots in our study indicated the standard contractual in the nomograms predicting OS and BCSS to ensure the accuracy of the developed nomograms. Further, ROC curves were constructed and the resulting time-dependent AUCs were used to demonstrate the specific ability of the model. The AUC for each of the study’s nomograms was greater than 0.7, indicating that the models could predict patient prognosis. Our study used other appropriate verification techniques and revisited the SEER database to reassess the performance of the nomograms in the absence of external data. Overall, these nomograms were able to predict the likelihood of survival for patients receiving various types of therapies with a high degree of confidence. Thus, these nomograms could help physicians make better clinical decisions.

It should be noted that distinct histopathological classifications of BC are frequently associated with varying prognoses in clinical practice. For example, patients diagnosed with pure mucinous carcinoma often exhibit more favorable prognostic outcomes, and consequently distinctive treatment approaches have been established to treat this particular subtype (38). However, it should be noted that due primarily to constraints stemming from sample size, the categorization of patients in this study was limited to invasive lobular carcinoma, invasive ductal carcinoma and other unspecified categories. Further, in this study, histopathology did not emerge as an independent risk factor for patients. In addition, our models enabled the personalized prediction of patients with HR-positive BCLM. Conversely, the conventional American Joint Committee on Cancer-TNM staging system adds scores for each variable to screen high- and low-risk patients for various readily available treatments. This is consistent with the growing emphasis on the individualized treatment of patients with cancer. For example, more aggressive care and therapy may be appropriate for high-risk patients. Simultaneously, avoiding unnecessary therapy can lessen the physical and financial strain placed on low-risk patients.

This study had certain limitations. First, due to its retrospective design, this study was inevitably flawed by bias. Second, the SEER database lacks detailed data on therapeutic approaches, including operations, chemotherapy regimens, and endocrine and targeted medicines, which might have affected our findings. Third, since the SEER database only contains data regarding a patient’s condition at the first diagnosis, we were unable to obtain any data on individuals who subsequently developed liver metastasis. Fourth, most of our patients were White (73.3%) or Black (17.7%), but race did not affect the prognosis of patients with BCLM in this study. However, it is currently unclear whether our model applies to other racial groups. Further, it is important to acknowledge that the absence of comprehensive data regarding patient comorbidities and health status in this study represents a significant limitation that could potentially influence our approach to practical treatment strategies. However, it is crucial to note that efforts will be made in subsequent research endeavors to address and rectify this limitation, ensuring a more thorough and informed analysis of these factors that are integral to patient care and decision-making processes.

Conclusions

This study created nomograms to obtain individualized survival estimates for patients with HR-positive BCLM. Additionally, a thorough analysis of the prognosis of these patients was performed using the SEER population-level information. Our nomograms will aid in clinical decision making and can be used to reliably and successfully forecast patient survival data.

Acknowledgments

Funding: This work was funded by the Medical Science and Technology Project of Zhejiang Province (Nos. 2022KY1078 and 2023KY1030)

Footnote

Reporting Checklist: The authors have completed the TRIPOD reporting checklist. Available at https://tcr.amegroups.com/article/view/10.21037/tcr-23-874/rc

Peer Review File: Available at https://tcr.amegroups.com/article/view/10.21037/tcr-23-874/prf

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tcr.amegroups.com/article/view/10.21037/tcr-23-874/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 (as revised in 2013).

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, Miller KD, Fuchs HE, et al. Cancer statistics, 2022. CA Cancer J Clin 2022;72:7-33. [Crossref] [PubMed]
2. Genre L, Roché H, Varela L, et al. External validation of a published nomogram for prediction of brain metastasis in patients with extra-cerebral metastatic breast cancer and risk regression analysis. Eur J Cancer 2017;72:200-9. [Crossref] [PubMed]
3. Bale R, Putzer D, Schullian P. Local Treatment of Breast Cancer Liver Metastasis. Cancers (Basel) 2019;11:1341. [Crossref] [PubMed]
4. Siegel R, Naishadham D, Jemal A. Cancer statistics for Hispanics/Latinos, 2012. CA Cancer J Clin 2012;62:283-98. [Crossref] [PubMed]
5. Tu Q, Hu C, Zhang H, et al. Establishment and Validation of Novel Clinical Prognosis Nomograms for Luminal A Breast Cancer Patients with Bone Metastasis. Biomed Res Int 2020;2020:1972064. [Crossref] [PubMed]
6. Ruiz A, van Hillegersberg R, Siesling S, et al. Surgical resection versus systemic therapy for breast cancer liver metastases: Results of a European case matched comparison. Eur J Cancer 2018;95:1-10. [Crossref] [PubMed]
7. Xiong Y, Shi X, Hu Q, et al. A Nomogram for Predicting Survival in Patients With Breast Cancer Liver Metastasis: A Population-Based Study. Front Oncol 2021;11:600768. [Crossref] [PubMed]
8. Gerratana L, Fanotto V, Bonotto M, et al. Pattern of metastasis and outcome in patients with breast cancer. Clin Exp Metastasis 2015;32:125-33. [Crossref] [PubMed]
9. Shi Y, Wei W, Li L, et al. The global status of research in breast cancer liver metastasis: a bibliometric and visualized analysis. Bioengineered 2021;12:12246-62. [Crossref] [PubMed]
10. Huang Z, Hu C, Liu K, et al. Risk factors, prognostic factors, and nomograms for bone metastasis in patients with newly diagnosed infiltrating duct carcinoma of the breast: a population-based study. BMC Cancer 2020;20:1145. [Crossref] [PubMed]
11. Yang A, Xiao W, Zheng S, et al. Predictive Nomogram of Subsequent Liver Metastasis After Mastectomy or Breast-Conserving Surgery in Patients With Nonmetastatic Breast Cancer. Cancer Control 2021;28:1073274821997418. [Crossref] [PubMed]
12. Schrodi S, Braun M, Andrulat A, et al. Outcome of breast cancer patients with low hormone receptor positivity: analysis of a 15-year population-based cohort. Ann Oncol 2021;32:1410-24. [Crossref] [PubMed]
13. Wei L, Gao H, Yu J, et al. Pharmacological Targeting of Androgen Receptor Elicits Context-Specific Effects in Estrogen Receptor-Positive Breast Cancer. Cancer Res 2023;83:456-70. [Crossref] [PubMed]
14. Ji L, Cheng L, Zhu X, et al. Risk and prognostic factors of breast cancer with liver metastases. BMC Cancer 2021;21:238. [Crossref] [PubMed]
15. Balachandran VP, Gonen M, Smith JJ, et al. Nomograms in oncology: more than meets the eye. Lancet Oncol 2015;16:e173-80. [Crossref] [PubMed]
16. Fitzgerald M, Saville BR, Lewis RJ. Decision curve analysis. JAMA 2015;313:409-10. [Crossref] [PubMed]
17. Wang Z, Cheng Y, Chen S, et al. Novel prognostic nomograms for female patients with breast cancer and bone metastasis at presentation. Ann Transl Med 2020;8:197. [Crossref] [PubMed]
18. Kohler BA, Sherman RL, Howlader N, et al. Annual Report to the Nation on the Status of Cancer, 1975-2011, Featuring Incidence of Breast Cancer Subtypes by Race/Ethnicity, Poverty, and State. J Natl Cancer Inst 2015;107:djv048. [Crossref] [PubMed]
19. Kay C, Martínez-Pérez C, Meehan J, et al. Current trends in the treatment of HR+/HER2+ breast cancer. Future Oncol 2021;17:1665-81. [Crossref] [PubMed]
20. AlFakeeh A, Brezden-Masley C. Overcoming endocrine resistance in hormone receptor-positive breast cancer. Curr Oncol 2018;25:S18-27. [Crossref] [PubMed]
21. Frank S, Carton M, Dubot C, et al. Impact of age at diagnosis of metastatic breast cancer on overall survival in the real-life ESME metastatic breast cancer cohort. Breast 2020;52:50-7. [Crossref] [PubMed]
22. Largillier R, Ferrero JM, Doyen J, et al. Prognostic factors in 1,038 women with metastatic breast cancer. Ann Oncol 2008;19:2012-9. [Crossref] [PubMed]
23. de Boer AZ, Bastiaannet E, Schetelig J, et al. Breast cancer mortality of older patients with and without recurrence analysed by novel multi-state models. Eur J Cancer 2022;174:212-20. [Crossref] [PubMed]
24. Turner N, Zafarana E, Becheri D, et al. Breast cancer in the elderly: which lessons have we learned? Future Oncol 2013;9:1871-81. [Crossref] [PubMed]
25. Botteri E, Bagnardi V, Goldhirsch A, et al. Axillary lymph node involvement in women with breast cancer: does it depend on age? Clin Breast Cancer 2010;10:318-21. [Crossref] [PubMed]
26. Gennari R, Curigliano G, Rotmensz N, et al. Breast carcinoma in elderly women: features of disease presentation, choice of local and systemic treatments compared with younger postmenopasual patients. Cancer 2004;101:1302-10. [Crossref] [PubMed]
27. Wildiers H, Kunkler I, Biganzoli L, et al. Management of breast cancer in elderly individuals: recommendations of the International Society of Geriatric Oncology. Lancet Oncol 2007;8:1101-15. [Crossref] [PubMed]
28. Yuan R, Zhang C, Li Q, et al. The impact of marital status on stage at diagnosis and survival of female patients with breast and gynecologic cancers: A meta-analysis. Gynecol Oncol 2021;162:778-87. [Crossref] [PubMed]
29. Shrout MR, Renna ME, Madison AA, et al. Breast cancer survivors' satisfying marriages predict better psychological and physical health: A longitudinal comparison of satisfied, dissatisfied, and unmarried women. Psychooncology 2021;30:699-707. [Crossref] [PubMed]
30. Waks AG, Winer EP. Breast Cancer Treatment: A Review. JAMA 2019;321:288-300. [Crossref] [PubMed]
31. Rashaan ZM, Bastiaannet E, Portielje JE, et al. Surgery in metastatic breast cancer: patients with a favorable profile seem to have the most benefit from surgery. Eur J Surg Oncol 2012;38:52-6. [Crossref] [PubMed]
32. Xiong Z, Deng G, Wang J, et al. Could local surgery improve survival in de novo stage IV breast cancer? BMC Cancer 2018;18:885. [Crossref] [PubMed]
33. Early Breast Cancer Trialists' Collaborative Group (EBCTCG). Effect of radiotherapy after breast-conserving surgery on 10-year recurrence and 15-year breast cancer death: meta-analysis of individual patient data for 10,801 women in 17 randomised trials. Lancet 2011;378:1707-16. [Crossref] [PubMed]
34. Choi SH, Kim JW, Choi J, et al. Locoregional Treatment of the Primary Tumor in Patients With De Novo Stage IV Breast Cancer: A Radiation Oncologist's Perspective. Clin Breast Cancer 2018;18:e167-78. [Crossref] [PubMed]
35. Gnerlich J, Jeffe DB, Deshpande AD, et al. Surgical removal of the primary tumor increases overall survival in patients with metastatic breast cancer: analysis of the 1988-2003 SEER data. Ann Surg Oncol 2007;14:2187-94. [Crossref] [PubMed]
36. Le Scodan R, Stevens D, Brain E, et al. Breast cancer with synchronous metastases: survival impact of exclusive locoregional radiotherapy. J Clin Oncol 2009;27:1375-81. [Crossref] [PubMed]
37. Dieci MV, Guarneri V. Should triple-positive breast cancer be recognized as a distinct subtype? Expert Rev Anticancer Ther 2020;20:1011-4. [Crossref] [PubMed]
38. Rechsteiner A, Dietrich D, Varga Z. Prognostic relevance of mixed histological subtypes in invasive breast carcinoma: a retrospective analysis. J Cancer Res Clin Oncol 2023;149:4967-78. [Crossref] [PubMed]