Development of a prognostic nomogram for lymph node positive HR⁺/HER2^– breast cancer patients: a study of SEER database and a Chinese cohort

Introduction

Breast cancer (BC) has become the most commonly diagnosed cancer and the leading cause of cancer-related deaths in women. Since 2020, it has surpassed lung cancer as the world’s leading diagnosed cancer, with 2.26 million new cases worldwide (1). BC manifests a high degree of heterogeneity at both clinical and molecular levels, and stratifying BC into different subtypes has proved to be an effective strategy to overcome heterogeneity (2). Estrogen receptor, progesterone receptor, human epidermal growth factor receptor (HER)2, and Ki-67 are used to define molecular subtypes (3). The molecular subtypes are closely related to BC prognosis. Patients with hormone receptor positive (HR⁺) tumors are the most common and generally have a better prognosis owing to their sensitivity to endocrine therapy. However, patients with HR⁺ BC are still at high risk of long-range recurrence and cancer-specific death compared with other subtypes (4,5). To reduce the rate of late recurrence, it is necessary to improve the management of patients in this group.

Given the biological heterogeneity of BC, current tumor node metastasis (TNM) staging system, even with its ability to incorporate molecular subtype information, is still inadequate to fully meet the corresponding clinical needs. Moreover, other clinical factors such as age, race, adjuvant therapy, and genetic background, may influence the prognosis. In recent years, genetic testing has been gradually selected as a tool to assess the risk of BC recurrence and make chemotherapy decisions, which mainly includes Oncotype DX recurrence score, Prediction Analysis of Microarray 50 (PAM50) recurrence score, MammaPrint (70 gene test), and breast cancer index (BCI). Axillary lymph node metastasis is an important prognostic factor that affects the recurrence and survival of invasive BC (6). However, most genetic tests are performed in populations with negative lymph nodes or a small number of lymph nodes with metastasis. Studies have found that these test results are often inconsistent, and that the correlation between these scores and the lymph node status is not strong (7-9). Therefore, there is a current demand to identify more specific prognostic factors to predict the prognosis of BC patients individually and guide clinical practice.

In the current study, we evaluated the prognostic value of clinical factors for HR⁺/HER2^– patients with lymph node metastasis. We selected the nomogram as the presentation form of the prediction model, which has been proved to be superior to the traditional TNM staging system, and has obvious advantages in predicting tumor recurrence, prognosis and outcome (10). On the basis of multivariate Cox proportional hazards regression, the nomogram assigns different scores of each influencing factor according to the contribution degree to the outcome variable. Each score is added to get the total score. Finally, the predicted value of the outcome event of the individual is calculated through the function conversion relationship between the total score and the probability of the outcome event. No previous nomogram has been established and validated to show the survival of patients with HR⁺/HER2^– and lymph node-positive BC. We extracted data from the Surveillance, Epidemiology and End Results (SEER) database, which is the most comprehensive cancer database including particular patient information of the United States population. We also collected a patient cohort to externally validate our nomogram, thus helping us confirm whether the nomogram was generalisable to another cohort. We present this article in accordance with the TRIPOD reporting checklist (available at https://gs.amegroups.com/article/view/10.21037/gs-23-177/rc).

Methods

Patient selection and data acquisition

SEER is a cancer database that collects disease information and survival outcomes for patients with cancer, and was established by the National Cancer Institute. In this study, we used SEER*Stat 8.3.5 to download and extract data on BC patients diagnosed between 2010 and 2015.

The inclusion criteria were as follows: (I) diagnosis confirmed only by positive histology; (II) patients actively followed up with complete survival information and survival time of at least one day; (III) patient’s only or first diagnosed primary cancer being BC; (IV) subtype of BC being HR⁺ and HER2^–; and (V) histological type being invasive ductal carcinoma (IDC), invasive lobular carcinoma (ILC), or mixed.

The following patients were excluded: (I) men; (II) age of diagnosis <18; (III) neoadjuvant therapy received; (IV) missing-information on race, marital status at diagnosis, tumor location, histological type, grade, American Joint Committee on Cancer (AJCC) T, N and M stage, metastasis site, radiotherapy, chemotherapy, living status, subtype or number of positive lymph nodes. A total of 25,338 patients were included in this study (Figure 1).

Figure 1 The flowchart of patients identified in this study. HR, hormone receptor; HER2, human epidermal growth factor receptor 2; SEER, Surveillance, Epidemiology and End Results; BC, breast cancer; IDC, invasive ductal carcinoma; ILC, invasive lobular carcinoma.

The enrolled patients were randomly divided into the training cohort and internal validation cohort with a ratio of 7:3. The training cohort was used to determine the independent prognostic factors of patients and establish a prognostic nomogram prediction model. The internal validation set was used to verify the nomogram-prediction model.

To verify our nomogram, we included 202 patients diagnosed with BC in 2010–2015 from the First Hospital of China Medical University as the external validation cohort. Patients were enrolled according to the inclusion and exclusion criteria used in the training cohort. The last follow-up was in January 2021. The experimental protocol was approved by the Research Ethics Committee of the First Hospital, China Medical University (No. 2020-203) in accordance with the principles of the Declaration of Helsinki (as revised in 2013) and the written informed consent was obtained from all subjects.

Variable collection

The variables selected in this study included age at diagnosis, race, marital status, tumor location, histological type, histological grade, AJCC T status, AJCC N status, AJCC M status, metastatic site (lung, liver, brain, bone), treatment information (radiotherapy, chemotherapy), and ER, PR and HER2 expression. Breast cancer-specific survival (BCSS), defined as the interval from diagnosis to BC-specific death, was used as the primary endpoint.

Statistical analysis

First, the significant influencing factors were screened out through univariate Cox analysis (P<0.05), and subsequently included in the multivariate Cox analysis. The nomogram was constructed based on the results of the multivariate Cox proportional hazards regression. (P<0.05).

The nomogram prediction model was constructed using R software. Receiver operating characteristic (ROC) curves and area under the curve (AUC)/concordance (C) index (C-index) were used to quantitatively evaluate the model efficacy and predictive ability. A calibration curve was used to verify the prediction model and evaluate the performance of the nomogram. We calculated the total scores of patients using the “survival”, “rms”, “nomogram Ex”, and “nomogram Formula” package of R studio and established a risk stratification system according to X-Tile software. The patients were divided into low-, medium-, and high-risk subgroups with significantly different BCSS. Kaplan-Meier curves were used to show the survival for different risk groups.

SPSS 23.0 and R software 4.0.3 were used for statistical analysis, and R software 4.0.3 was used for model visualization and verification. Statistical significance was set at P<0.05 in a two-tailed test.

Results

Clinicopathologic characteristics

A total of 25,338 patients that met the inclusion and exclusion criteria were finally enrolled from the SEER database, and were randomly divided into a training cohort (n=17,738) and an internal validation cohort (n=7,600) (Figure 1). Table 1 shows the detailed demographic and clinicopathological characteristics of all patients.

The median age of all cases was 57, with the largest proportion of patients aged 50–59 accounting for 27.5% of all patients. The median follow-up period was 45 months. The 3- and 5-year BCSS rates were 94.8% and 92.2%, respectively. The most common histological type was IDC, with the highest proportion of patients belonging to the grade II category. In the AJCC T staging system, the largest number of patients belonged to the T2 staging category. In the AJCC N staging system, the largest proportion of patients belonged to the N1 category (75%). Bone metastasis was the most common (3.09%, n=784), lung metastasis rate was 0.84% (n=212), liver metastasis rate was 0.78% (n=197), with the brain metastasis rate being the lowest at 0.075% (n=19).

Univariate and multivariable Cox analysis and the construction of nomogram

The hazard ratio (HR) results of the univariate and multivariable Cox proportional hazards regressions are shown in Table 2. Univariate Cox regression analysis showed that three general conditions: age (P<0.001), race (P<0.001), marital status (P<0.001) and nine clinical conditions: histological grade (P<0.001), AJCC T stage (P<0.001), AJCC N stage (P<0.001), radiotherapy (P<0.001), chemotherapy (P<0.001), bone metastasis (P<0.001), liver metastasis (P<0.001), lung metastasis (P<0.001), brain metastasis (P<0.001) were related variables of BCSS. Location (P=0.653) and histological type (P=0.344) were not included in the multivariable Cox regression analysis.

In multivariate Cox regression analysis, the older age [age ≥80; HR =1.592; 95% confidence interval (CI): 0.953–2.659; P=0.076], black race (P<0.001), unmarried status (HR =1.171; 95% CI: 1.053–1.302; P=0.003), histology grade III (HR =3.292; 95% CI: 2.684–4.038; P<0.001), AJCC T4 stage (HR =5.255; 95% CI: 4.330–6.376; P<0.001), AJCC N3 stage (HR =2.614; 95% CI: 2.295–2.977; P<0.001), AJCC M1 stage (HR =3.096; 95% CI: 2.468–3.883; P<0.001), not receiving chemotherapy and radiotherapy (P<0.001), bone metastasis (HR =3.504; 95% CI: 2.980–4.119; P<0.001), brain metastasis (HR =4.682; 95% CI: 2.576–8.510; P<0.001), liver metastasis (HR =3.146; 95% CI: 2.429–4.074; P<0.001) and lung metastasis (HR =1.501; 95% CI: 1.156–1.950; P=0.002) were independently associated with BCSS rate. These variables were used to construct the nomogram (Figure 2).

Figure 2 The nomogram for predicting the breast cancer-specific survival for HR⁺/HER2^– patients with lymph node metastasis. HR, hormone receptor; HER2, human epidermal growth factor receptor 2.

Development and validation of the prognostic nomogram

According to the results of multivariable Cox regression analysis, we used the R software to construct nomogram models of patients’ 3- and 5-year BCSS.

The ROC curve showed that the AUC values of the 3- and 5-year BCSS in the training cohort were 0.800 and 0.800 (Figure 3A,3B). The AUC values at 3- and 5-year BCSS were 0.831 and 0.819 for the internal validation set (Figure 3C,3D) and those for the external validation set were 0.765 and 0.735 respectively (Figure 3E,3F), demonstrating that our model had satisfactory discrimination. The calibration curves also showed that the actual observed results were in good agreement with the predicted results in the training and validation cohorts (Figure 4).

Figure 3 The ROC curves for the 3- and 5-year BCSS. (A) Prediction of 3-year BCSS in the training cohort; (B) prediction of 5-year BCSS in the training cohort; (C) prediction of 3-year BCSS in the internal validation cohort; (D) prediction of 5-year BCSS in the internal validation cohort; (E) prediction of 3-year BCSS in the external validation cohort; (F) prediction of 5-year BCSS in the external validation cohort. AUC, area under the curve; ROC, receiver operating characteristic; BCSS, breast cancer-specific survival.

Figure 4 The calibration curves for predicting BCSS. (A) Prediction of 3-year BCSS in the training cohort; (B) prediction of 3-year BCSS in the internal validation cohort; (C) prediction of 3-year BCSS in the external validation cohort; (D) prediction of 5-year BCSS in the training cohort; (E) prediction of 5-year BCSS in the internal validation cohort; (F) prediction of 5-year BCSS in the external validation cohort. BCSS, breast cancer-specific survival.

Risk stratification system

The corresponding score for this variable can be obtained by referring to the standard scale above each variable. The total score was obtained by adding the scores of all factors, and the corresponding value of the total score was the 3/5-year BCSS rate of the target population. The total points for all patients in both cohorts were calculated based on the predicted nomogram score. The cutoff value of the survival data was determined using the X-Tile software. According to the cutoff value, patients were divided into low- (total scores ≤243.27), medium- (243.27 < total scores <301.52), and high-risk (total scores ≥301.52) groups. However, in the external validation cohorts, the total number of scores in this cohort did not cover all three risk groups, which was probably due to the low number of patients. Therefore, we validated our risk stratification system using only the internal validation cohort.

In both the training and internal validation cohorts, the 3-year BCSS of the low-, medium- and high-risk groups were 98.4%, 92.8% and 72.8%, respectively, and the 5-year BCSS were 96.6%,87.3% and 62.4%, respectively. The results of the Kaplan-Meier survival analysis are shown in Figure 5 (P<0.0001).

Figure 5 Survival curves of nomogram-based risk stratification. (A) Training cohort; (B) internal validation cohort.

Discussion

Lymph node status is a key indicator of patient prognosis (11,12). Patients with HR⁺ BC have a better prognosis than those with the other two subtypes [HER2 positive and triple-negative breast cancer (TNBC)] (2). The HR⁺/HER2^– subtype patients account for the highest proportion among all molecular subtypes, and its annual incidence is predicted to increase by 2% over the next 3 years (13,14). Nevertheless, HR⁺/HER2^– patients with lymph node metastasis have not drawn much attention in clinical treatment. Consequently, this leads to incorrect judgment of the prognosis and under- or over-treatment in this group.

To resolve this issue, we developed a nomogram to predict the survival of HR⁺/HER2^– patients with lymph node metastasis. A nomogram was applied as an integrated statistical model based on multivariable analysis, which demonstrated good accuracy. Studies have shown that nomograms have advantages in terms of assessing the cancer risk, selecting therapies and medicines, and predicting survival outcomes in various types of cancer (10,15).

We constructed the first nomogram based on the Cox proportional hazards model of HR⁺/HER2^– and lymph node-positive patients. Due to the large number of this subtype of patients, the nomogram may be widely used in clinical practice in the future. Through the patient information easily obtained in clinical work, the 3-year and 5-year survival rates of patients can be easily and quickly estimated. Beyond that, a certain degree of support can be provided for identifying high-risk patients and deciding whether to prolong the endocrine therapy of patients.

The nomogram was internally validated, and its performance was evaluated using the C-index and calibration curve. According to previous studies, the C-index was between 0.6 and 0.8, indicating that our nomogram had a high predictive value on prognosis. In this study, the calibration curve showed the best consistency in predicting BCSS, thereby ensuring the reliability of the established nomogram. In conclusion, our nomogram provides an accurate estimate of patient outcome. This was an original study in which a visual predictive model was constructed to improve survival in HR⁺/HER2^– patients and provide useful information.

Compelling evidence have shown that young age is an independent negative predictor of BC survival (16,17) and is associated with the risk of the occurrence of contralateral BC and local recurrence (18). Previous studies have also shown that elderly patients with BC have higher mortality rates, which may be because elderly patients appear to be at greater risk of developing chemotherapy-related cardiotoxicity than their younger counterparts and could not tolerate intensive standard chemotherapy (19). In addition, elderly patients are more likely to suffer from other chronic diseases (e.g., diabetes), which could also affect the survival. Similar to prior findings (20), patients aged ≥70 and ≤30 were at a higher risk of poor survival in our nomogram.

Radiotherapy is a vital adjuvant treatment for patients with lymph node metastasis. Different subtypes of BC possess differential radiosensitivity, aggressiveness, and malignancy, consequently different outcomes after radiotherapy. It has been shown that luminal A BC has the most favorable clinical benefits after receiving radiotherapy compared to the subtypes of HER2⁺ and TNBC (21,22). In addition, when considering the cause of the radiosensitivity disparity, TNBC and HER2⁺ BC are more aggressive compared with luminal BC because of the tumor itself. In our study, radiotherapy did not have a strong effect on the BCSS rate.

Chemotherapy is also an important and common treatment option for BC. In this Chinese cohort, 202 patients received chemotherapy but it was not an important prognostic factor in the present study. Moreover, chemotherapy had the least effect on the current nomogram. Whether chemotherapy is required for HR⁺/HER2^– BC patients in the early stage has always been a controversial issue. Most current guidelines recommend that this requirement is determined through gene testing, e.g., Oncotype DX or MammaPrint. However, gene testing has its limitations and defects. In the past few years, several retrospective analyses have shown that the results varied when chemotherapy was considered necessary for these BC patients (23-25). The DBCG77B trial reported that not all patients could benefit from adjuvant chemotherapy, even among high-risk populations (26). In the TAILORx study, the risk grouping criteria were redefined (27), which indicated that it was essential to categorize patients more accurately to implement precise treatment. Although the patients included in this study had a high risk for lymph node metastasis, our results could not prove the degree of benefit of chemotherapy in HR⁺/HER2^– BC patients with lymph node metastasis. Our study focused on survival after treatment. For HR⁺ patients, endocrine therapy was critical, but the information of endocrine therapy was not included in SEER database. However, our nomogram revealed that the included patients benefitted from endocrine therapy to a large extent.

In our study, distant metastases of bone, brain and liver, T stage, and histological grade had a great impact on prognosis, while lymph node grade did not have great weight, which is somewhat contradictory to the current clinical consensus. According to the current consensus, higher degree of lymph nodes metastasis would indicate greater risk of recurrence. It is worth considering whether the degree of lymph node metastasis is equivalent to the degree of recurrence risk when other factors are different. However, this question warrants further research.

Using Cox regression analysis, tumor location and histological type were excluded from our model, and the same conclusion was reached as in the previous model (20). Our model included specific distant metastasis of brain, bone, liver and lung metastasis, while the AJCC M stage contained the similar information of the condition of distant metastasis. Because our nomogram needed to calculate the survival rate by combining the corresponding scores of different influencing factors, the AJCC M stage was not included to avoid duplicate calculation. Our nomogram showed that patients with brain metastasis had the highest risk of death compared to those with bone, liver, and lung metastases.

However, several limitations of this study merit attention. First, as this was a retrospective study, selection bias was inevitable. Moreover, because of the lack of data, the effects of endocrine therapy, specific radiotherapy and chemotherapy regimens, and family history could not be fully studied. In addition, because of the small number of patients included in the external validation set, the application of our nomogram to the risk group in the Chinese patient cohort could not be verified. We will attempt to collect prospective data and add more prospective validation in future studies to improve the accuracy and integrity of the nomogram.

Conclusions

We constructed a novel nomogram for BCSS and risk stratification in HR⁺/HER2^– BC patients with axillary lymph node metastasis. Our analysis identified 12 factors including age, race, marital status, histological grade, T stage, N stage, radiotherapy, chemotherapy, metastasis of bone, liver, lung and brain, to predict BCSS. Despite some limitations, our population-based study provides several influential survival factors and guidance for further prospective studies on HR⁺/HER2^– and lymph node-positive BC. The manifestation of the nomogram provides a more intuitive approach to assist clinicians to identify high-risk populations and a reference for postoperative treatments to some extent.

Acknowledgments

We are grateful for the Surveillance, Epidemiology, and End Results database and all our colleagues in Department of Breast Surgery, The First Hospital of China Medical University.

Funding: This work was supported by grants from the Natural Science Foundation of Liaoning Province (grant No. 20180551215), Key R & D Plan Guidance Project of Liaoning Province (grant No. 2019JH8/10300020), and the Key Project of China Health Promotion Foundation (grant No. CHPF-RXO180301).

Footnote

Reporting Checklist: The authors have completed the TRIPOD reporting checklist. Available at https://gs.amegroups.com/article/view/10.21037/gs-23-177/rc

Data Sharing Statement: Available at https://gs.amegroups.com/article/view/10.21037/gs-23-177/dss

Peer Review File: Available at https://gs.amegroups.com/article/view/10.21037/gs-23-177/prf

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://gs.amegroups.com/article/view/10.21037/gs-23-177/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 experimental protocol was approved by the Research Ethics Committee of the First Hospital, China Medical University (No. 2020-203) in accordance with the principles of the Declaration of Helsinki (as revised in 2013) and the written informed consent was obtained from all subjects.

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. Sung H, Ferlay J, Siegel RL, et al. Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. CA Cancer J Clin 2021;71:209-49. [Crossref] [PubMed]
2. Taherian-Fard A, Srihari S, Ragan MA. Breast cancer classification: linking molecular mechanisms to disease prognosis. Brief Bioinform 2015;16:461-74. [Crossref] [PubMed]
3. Goldhirsch A, Wood WC, Coates AS, et al. Strategies for subtypes--dealing with the diversity of breast cancer: highlights of the St. Gallen International Expert Consensus on the Primary Therapy of Early Breast Cancer 2011. Ann Oncol 2011;22:1736-47. [Crossref] [PubMed]
4. Effects of chemotherapy and hormonal therapy for early breast cancer on recurrence and 15-year survival: an overview of the randomised trials. Lancet 2005;365:1687-717. [Crossref] [PubMed]
5. Saphner T, Tormey DC, Gray R. Annual hazard rates of recurrence for breast cancer after primary therapy. J Clin Oncol 1996;14:2738-46. [Crossref] [PubMed]
6. Recht A, Houlihan MJ. Axillary lymph nodes and breast cancer: a review. Cancer 1995;76:1491-512. [Crossref] [PubMed]
7. 7Sparano JA, Gray RJ, Makower DF, et al. Adjuvant Chemotherapy Guided by a 21-Gene Expression Assay in Breast Cancer. N Engl J Med 2018;379:111-21.
8. Lænkholm AV, Jensen MB, Eriksen JO, et al. PAM50 Risk of Recurrence Score Predicts 10-Year Distant Recurrence in a Comprehensive Danish Cohort of Postmenopausal Women Allocated to 5 Years of Endocrine Therapy for Hormone Receptor-Positive Early Breast Cancer. J Clin Oncol 2018;36:735-40. [Crossref] [PubMed]
9. Bartlett JMS, Sgroi DC, Treuner K, et al. Breast Cancer Index and prediction of benefit from extended endocrine therapy in breast cancer patients treated in the Adjuvant Tamoxifen-To Offer More? (aTTom) trial. Ann Oncol 2019;30:1776-83. [Crossref] [PubMed]
10. Balachandran VP, Gonen M, Smith JJ, et al. Nomograms in oncology: more than meets the eye. Lancet Oncol 2015;16:e173-80. [Crossref] [PubMed]
11. Zhang H, Zhang N, Moran MS, et al. Special subtypes with favorable prognosis in breast cancer: A registry-based cohort study and network meta-analysis. Cancer Treat Rev 2020;91:102108. [Crossref] [PubMed]
12. Kehl KL, Giordano SH. BRCA1 and BRCA2 Testing Among Young Breast Cancer Survivors. JAMA Oncol 2016;2:688-9. [Crossref] [PubMed]
13. Burstein HJ. Systemic Therapy for Estrogen Receptor-Positive, HER2-Negative Breast Cancer. N Engl J Med 2020;383:2557-70.
14. Zuo T, Zeng H, Li H, et al. The influence of stage at diagnosis and molecular subtype on breast cancer patient survival: a hospital-based multi-center study. Chin J Cancer 2017;36:84. [Crossref] [PubMed]
15. Iasonos A, Schrag D, Raj GV, et al. How to build and interpret a nomogram for cancer prognosis. J Clin Oncol 2008;26:1364-70. [Crossref] [PubMed]
16. Anders CK, Hsu DS, Broadwater G, et al. Young age at diagnosis correlates with worse prognosis and defines a subset of breast cancers with shared patterns of gene expression. J Clin Oncol 2008;26:3324-30. [Crossref] [PubMed]
17. Narod SA. Breast cancer in young women. Nat Rev Clin Oncol 2012;9:460-70. [Crossref] [PubMed]
18. Reiner AS, Watt GP, John EM, et al. Smoking, Radiation Therapy, and Contralateral Breast Cancer Risk in Young Women. J Natl Cancer Inst 2022;114:631-4. [Crossref] [PubMed]
19. Hershman DL, Till C, Shen S, et al. Association of Cardiovascular Risk Factors With Cardiac Events and Survival Outcomes Among Patients With Breast Cancer Enrolled in SWOG Clinical Trials. J Clin Oncol 2018;36:2710-17. [Crossref] [PubMed]
20. Li Y, Ma L. Nomograms predict survival of patients with lymph node-positive, luminal a breast cancer. BMC Cancer 2021;21:965. [Crossref] [PubMed]
21. He L, Lv Y, Song Y, et al. The prognosis comparison of different molecular subtypes of breast tumors after radiotherapy and the intrinsic reasons for their distinct radiosensitivity. Cancer Manag Res 2019;11:5765-75. [Crossref] [PubMed]
22. Sjöström M, Lundstedt D, Hartman L, et al. Response to Radiotherapy After Breast-Conserving Surgery in Different Breast Cancer Subtypes in the Swedish Breast Cancer Group 91 Radiotherapy Randomized Clinical Trial. J Clin Oncol 2017;35:3222-9. [Crossref] [PubMed]
23. Ravdin PM, Siminoff LA, Davis GJ, et al. Computer program to assist in making decisions about adjuvant therapy for women with early breast cancer. J Clin Oncol 2001;19:980-91. [Crossref] [PubMed]
24. Goldhirsch A, Wood WC, Gelber RD, et al. Progress and promise: highlights of the international expert consensus on the primary therapy of early breast cancer 2007. Ann Oncol 2007;18:1133-44. [Crossref] [PubMed]
25. Chia SK, Speers CH, Bryce CJ, et al. Ten-year outcomes in a population-based cohort of node-negative, lymphatic, and vascular invasion-negative early breast cancers without adjuvant systemic therapies. J Clin Oncol 2004;22:1630-7. [Crossref] [PubMed]
26. Nielsen TO, Jensen MB, Burugu S, et al. High-Risk Premenopausal Luminal A Breast Cancer Patients Derive no Benefit from Adjuvant Cyclophosphamide-based Chemotherapy: Results from the DBCG77B Clinical Trial. Clin Cancer Res 2017;23:946-53. [Crossref] [PubMed]
27. Sparano JA, Gray RJ, Makower DF, et al. Prospective Validation of a 21-Gene Expression Assay in Breast Cancer. N Engl J Med 2015;373:2005-14. [Crossref] [PubMed]