Retraction of: Screening of Clinical Factors Related to Prognosis of Breast Cancer Based on the Cox Proportional Risk Model

Abstract

Dr. Degong Mu, one of the co-corresponding authors of the article entitled, “Screening of Clinical Factors Related to Prognosis of Breast Cancer Based on the Cox Proportional Risk Model,” (by Wan Tang, Degong Mu, Ling Han, Xianmin Guo, Bing Han, and Dong Song; J Comput Biol. 2021;28(1):89–98; doi: 10.1089/cmb.2019.0110) has requested a full retraction of the published paper based on the following reasons:

We could not prove that the clinical characteristics listed in Table 1 were independent prognostic factors using a

p value <0.05. Further, we could not prove that patients with younger age and lower stage showed better overall clinical survival according to the column map in Figure 2. We thus cannot ensure the reliability of our conclusions, and we have to retract the article. We apologize to the editors and readers of the journal. [sic]

The Publisher secured all coauthors' agreements to officially retract the article.

1. Introduction

In recent years, breast cancer has been the first in female malignant tumors, the disease is one of the sex hormone receptor-dependent tumors, and the presence of estrogen can promote the growth and appreciation of the cancer cells (Müller et al., 2001; Veronesi et al., 2005; Elston and Ellis, 2010; Higa and Fell, 2013). The etiology of breast cancer is not very clear, the pathogenesis is very complex, and the risk factors of breast cancer are hormones (DrPH, 1984; Lu et al., 2000), genetics (Nathanson et al., 2001), immunization (Gilewski et al., 2001; Longenecker et al., 2010), physiology (Shao et al., 2011; Budde et al., 2012), and so on. Although the domestic and foreign scholars have carried out a lot of research, there are still not sure which one of the risk factors is the main pathogenic factor of breast cancer. However, it is agreed that the overlap of multiple risk factors is bound to increase the risk of breast cancer (Sestak et al., 2018).

With the continuous maturation of radiotherapy technology, its role in the field of breast cancer is becoming more and more prominent, especially the reduction of recurrence and metastasis rate of breast cancer showed obvious effect (Hickey et al., 2013; Efird et al., 2018). Screening and identifying the clinically significant factors associated with breast cancer can help predict the risk of disease and, if not treated, the risk of dying from breast cancer, to identify and anticipate the outcome of a specific treatment regimen for the patient.

Proper evaluation of the relevant clinical factors for the prognosis of breast cancer is particularly important in the selection of appropriate therapeutic strategies. The follow-up data of intact breast cancer patients were downloaded from METABRIC (Molecular Taxonomy of Breast Cancer International Consortium) database, and the prognostic factors correlated with radiotherapy factors were screened using the Cox risk regression model analysis of prognostic factors. The response of different clinical features to radiotherapy was also evaluated by survival prognosis analysis and prediction.

2. Methods

2.1. Patients

A total of 1980 breast cancer patients were enrolled in this study, including 1173 patients who received the radiotherapy treatment (with radiotherapy group) and 807 patients without radiotherapy treatment (without radiotherapy group). The clinical data of the breast cancer patients were downloaded from the METABRIC database (http://molonc.bccrc.ca) (Milioli et al., 2015). The median age of patients was 61.09 years (21.93–96.29 years), and the median follow-up time was 125.13 months (0.1–355 months). In all patients, the survival rate was 42.27%. In addition, the claudin subtype, chemotherapy, hormotherapy, grade, stage, estrogen receptor (ER), progesterone receptor (PR), and overall survival time of all patients were collected for the subsequent analysis.

2.2. Clinical prognostic factors screening

The clinical prognostic factors were screened using the single factor and multivariate Cox regression analysis in the survival package version of the R3.4.0 language (https://cran.r-project.org/web/packages/survival/index.html) (Wang et al., 2016), and log-rank p-value <0.05 was considered as the significant threshold for screening. Then, the independent prognostic factors that are significantly correlated with the overall survival prognosis (overall survival) were selected, and the Kaplan–Meier curve (Lacny et al., 2015) was analyzed for the clinical prognostic factors of the screening.

2.3. The correlation between clinical factors and survival prognosis

To further study the correlation between the clinical prognostic factors and the overall survival, the RMS package Version 5.1–2 (https://cran.r-project.org/web/packages/rms/index.html) in the R3.4.0 language was used to build the column map (nomogram) (Anderson et al., 1989; Eng et al., 2015). The graph is essentially the visualization of the regression equation, according to the size of all the self-variable regression coefficients to set the scoring criteria, give each of the independent variables of the level of a score, each sample can be calculated to obtain a total score, and then the probability of the outcome of each sample by the conversion function to calculate the probability of the occurrence of each specimen. We predict the survival rate of different clinical indicators based on the prognostic significance of independent prognostic clinical factors using the Nomogram method.

3. Results

3.1. Characteristics analysis

A total of 1980 breast cancer patients were enrolled in this study, including 1173 patients who received the radiotherapy treatment and 807 patients without radiotherapy treatment. The demographic and baseline characteristics of patients are shown in Table 1. The age was 59.71 ± 12.87 years in the with radiotherapy group and 63.09 ± 12.82 years in the without radiotherapy group. A total of 857 patients who received the radiotherapy treatment were ER positive and 316 were ER negative, and a total of 649 patients without radiotherapy treatment were ER positive and 158 were ER negative. A total of 150 patients who received the radiotherapy treatment were human epidermal growth factor receptor 2 (HER2) positive and 1023 were HER2 negative, and a total of 97 patients without radiotherapy treatment were HER2 positive and 710 were HER2 negative. A total of 592 patients who received the radiotherapy treatment were PR positive and 581 were PR negative, and a total of 448 patients without radiotherapy treatment were PR positive and 359 were PR negative. Overall survival time of the patients who received the radiotherapy treatment was 124.83 ± 75.06 months, and the overall survival time of the patients without radiotherapy treatment was 125.56 ± 77.69 months. For the characteristic data of patients, the age, radiotherapy, chemotherapy, hormotherapy, grade, stage, ER, HER2, and PR were independent prognostic factors that were significantly correlated with the overall survival (p < 0.05, Table 1).

Table 1.

Clinical Characteristics of All Breast Cancer Patients

Clinical characteristic	With radiotherapy (n = 1173)	Without radiotherapy (n = 807)	p
Age (years, mean ± SD)	59.71 ± 12.87	63.09 ± 12.82	1.039E-08^a
Claudin subtype (basal/claudin-low/Her2/LumA/LumB/other)	141/145/134/378/299/76	68/73/90/322/176/78	4.428E-05^a
Chemotherapy (yes/no)	335/838	77/730	2.2E-16^a
Hormotherapy (yes/no)	768/405	448/359	9.591E-06^a
Grade (1/2/3/—)	84/418/643/28	85/353/309/60	1.419E-09^a
Stage (0/1/2/3/4/—)	3/289/560/100/5/216	9/212/265/18/5/298	4.914E-09^a
ER status (positive/negative/—)	857/316	649/158	1.723E-04^a
HER2 status (positive/negative/—)	150/1023	97/710	0.661^a
PR status (positive/negative/—)	592/581	448/359	0.0305^a
Death (dead/alive)	629/544	514/293	8.715E-06^a
Overall survival time (months, mean ± SD)	124.83 ± 75.06	125.56 ± 77.69	0.835^b

Group t test.

Fisher's exact test.

ER, estrogen receptor; HER2, human epidermal growth factor receptor 2; PR, progesterone receptor; SD, standard deviation.

3.2. Clinical prognostic factors analysis

To further study the correlation between the clinical prognostic factors and the overall survival, the single factor and multivariate Cox regression analysis were performed using the survival package version of the R3.4.0 language. For univariable Cox regression analysis, age (hazard ratio [95% confidence interval]: 1.036 [1.031–1.041], p = 2E-16), radiotherapy (0.851 [0.758–0.957], p = 0.006757), grade (1.283 [1.166–1.411], p = 3.128E-07), stage (1.812 [1.626–2.019], p = 2E-16), and HER2 (1.465 [1.237–1.736], p = 1.02E-05) were significantly correlated with the overall survival (p < 0.05, Table 2). For multivariable Cox regression, age (1.043 [1.035–1.05], p = 2E-16), radiotherapy (0.825 [0.709–0.961], p = 0.0134), grade (1.172 [1.034–1.327], p = 0.0129), stage (1.599 [1.410–1.813], p = 2.31E-13), and HER2 (1.481 [1.195–1.835], p = 3.39E-04) were significantly correlated with the overall survival (p < 0.05, Table 2). These results further suggested that age, radiotherapy, grade, stage, and HER2 were independent prognostic factors that were significantly correlated with the overall survival (p < 0.05, Table 2). In addition, the Kaplan–Meier curve analysis for the clinical prognostic factors implied that the patients with age <60 years, receiving radiotherapy, grade 1, stage 0–1, or HER2 negative had a better overall clinical survival (Fig. 1). The column map also implied that the patients with younger age, receiving radiotherapy, grade lower, stage lower, or HER2 negative showed a better overall clinical survival (Fig. 2).

FIG. 1.

The Kaplan–Meier curve analysis for the clinical prognostic factors. (A) The Kaplan–Meier curve analysis for age: black and light grey curves represent the patients aged ≤60 years and the patients aged >60 years, respectively. (B) The Kaplan–Meier curve analysis for radiotherapy: black and light grey curves indicate the patients without radiotherapy treatment and patients with radiotherapy treatment, respectively. (C) The Kaplan–Meier curve analysis for grade: black, light grey, and dark grey curves represent grade 1, 2, and 3, respectively. (D) The Kaplan–Meier curve analysis for stage and prognosis-related km curve: black and light grey curves represent stage 0–1 and 2–4, respectively. (E) The Kaplan–Meier curve analysis for HER2 status: black and light grey curves indicate HER2 negative and HER2 positive, respectively. HER2, human epidermal growth factor receptor 2.

FIG. 2.

The column map of 3-year survival probability and 5-year survival probability for five clinical prognostic factors, including age, radiotherapy, grade, stage, and HER2.

Table 2.

Clinical Prognostic Factors Analysis of All Breast Cancer Patients

Clinical characteristic	Univariable Cox regression		Multivariable Cox regression
Clinical characteristic	HR [95% CI]	p	HR [95% CI]	p
Age (years, mean ± SD)	1.036 [1.031–1.041]	2E-16	1.043 [1.035–1.05]	2E-16
Claudin subtype (basal/claudin-low/Her2/LumA/LumB/other)	1.024 [0.981–1.069]	0.282	—	—
Radiotherapy (yes/no)	0.851 [0.758–0.957]	0.006757	0.825 [0.709–0.961]	0.0134
Chemotherapy (yes/no)	1.262 [1.089–1.461]	0.001871	1.524 [1.204–1.928]	0.455
Hormotherapy (yes/no)	1.267 [1.121–1.431]	0.0001387	0.871 [0.737–1.029]	0.103
Grade (1/2/3/—)	1.283 [1.166–1.411]	3.128E-07	1.172 [1.034–1.327	0.0129
Stage (0/1/2/3/4/—)	1.812 [1.626–2.019]	2E-16	1.599 [1.410–1.813]	2.31E-13
ER status (positive/negative/—)	0.863 [0.751–0.991]	0.0355	0.995 [0.792–1.249]	0.963
HER2 status (positive/negative/—)	1.465 [1.237–1.736]	1.02E-05	1.481 [1.195–1.835]	3.39E-04
PR status (positive/negative/—)	0.792 [0.705–0.889]	8.10E-05	0.938 [0.793–1.109]	0.451

HR, hazard ratio; CI, confidence interval.

3.3. Analysis of clinical prognostic factors correlated with radiotherapy

To further study the correlation between the radiotherapy treatment and the clinical prognostic factors, the single factor and multivariate Cox regression analysis of the patients with radiotherapy treatment group and the patients without radiotherapy treatment group was performed using the survival package version of the R3.4.0 language. For the patients without radiotherapy treatment group, age (1.845 [1.521–2.239], p = 3.013E-10) and stage (2.365 [1.865–3.00], p = 2.74E-13) in univariable Cox regression analysis were significantly correlated with the overall survival (p < 0.05, Table 3). In addition, age (2.282 [1.705–3.053], p = 2.85E-08) and stage (2.216 [1.698–2.892], p = 4.79E-09) in multivariate Cox regression analysis were significantly correlated with the overall survival (p < 0.05, Table 3).

Table 3.

Analysis of Clinical Prognostic Factors Correlated with Radiotherapy

Clinical characteristic	Univariable Cox regression		Multivariable Cox regression
Clinical characteristic	HR [95% CI]	p	HR [95% CI]	p
Without radiotherapy
Age (≤60 years/>60 years)	1.845 [1.521–2.239]	3.013E-10	2.282 [1.705–3.053]	2.85E-08
Claudin subtype (basal/claudin-low/Her2/LumA/LumB/other)	1.003 [0.939–1.072]	0.926	—	—
Chemotherapy (yes/no)	1.454 [1.085–1.948]	0.0116	1.676 [1.041–2.698]	0.335
Hormotherapy (yes/no)	1.305 [1.092–1.558]	0.00326	0.846 [0.661–1.084]	0.187
Grade (1/2/3)	1.274 [1.107–1.466]	7.29E-04	1.106 [0.921–1.327]	0.281
Stage (0/1/2/3/4)	2.365 [1.865–3.00]	2.74E-13	2.216 [1.698–2.892]	4.79E-09
ER status (positive/negative)	0.924 [0.734–1.162]	0.498	—	—
HER2 status (positive/negative)	1.297 [0.991–1.699]	0.0578	—	—
PR status (positive/negative)	0.869 [0.730–1.035]	0.116	—	—
With radiotherapy
Age (≤60 years/>60 years)	1.984 [1.684–2.338]	1.11E-16	2.331 [1.881–2.889]	1.12E-14
Claudin subtype (basal/claudin-low/Her2/LumA/LumB/other)	1.031 [0.974–1.091]	0.294	—	—
Chemotherapy (yes/no)	1.301 [1.092–1.552]	0.00323	1.429 [1.094–1.869]	0.891
Hormotherapy (yes/no)	1.278 [1.079–1.513]	0.00439	1.012 [0.805–1.271]	0.919
Grade (1/2/3)	1.359 [1.188–1.555]	7.47E-06	1.109 [0.939–1.310]	0.219
Stage (0/1/2/3/4)	1.786 [1.461–2.184]	9.97E-09	1.518 [1.206–1.910]	0.000373
ER status (positive/negative)	0.802 [0.673–0.955]	0.0132	0.960 [0.722–1.277]	0.78
HER2 status (positive/negative)	1.603 [1.288–1.995]	1.97E-05	1.506 [1.164–1.949]	0.00186
PR status (positive/negative)	0.726 [0.620–0.849]	5.81E-05	0.879 [0.709–1.089]	0.238

For the patients with radiotherapy treatment group, age (1.984 [1.684–2.338], p = 1.11E-16), stage (1.786 [1.461–2.184], p = 9.97E-09), and HER2 status (1.603 [1.288–1.995], p = 1.97E-05) in univariable Cox regression analysis were significantly correlated with the overall survival (p < 0.05, Table 3). In addition, age (2.331 [1.881–2.889], p = 1.12E-14) stage (1.518 [1.206–1.910], p = 0.000373), and HER2 status (1.506 [1.164–1.949], p = 0.00186) in multivariate Cox regression analysis were significantly correlated with the overall survival (p < 0.05, Table 3). The Kaplan–Meier curve analysis for the clinical prognostic factors implied that the patients with age <60 years, receiving radiotherapy, grade 1, stage 0–1, or HER2 negative had a better overall clinical survival (Fig. 3). The column map also implied that the patients with younger age, stage lower, or HER2 negative showed a better overall clinical survival, and the patients who received radiotherapy had a better 3-year survival probability and 5-year survival probability (Fig. 4).

FIG. 3.

The Kaplan–Meier curve analysis for the without radiotherapy group (A) and the with radiotherapy group (B). The Kaplan–Meier curve analysis for age: black and light grey curves represent the patients aged ≤60 years and the patients aged >60 years, respectively. The Kaplan–Meier curve analysis for stage: black, light grey, and dark grey curves represent stage 0–1 and stage 2–4, respectively. The Kaplan–Meier curve analysis for HER2 status: black and light grey curves indicate HER2 negative and HER2 positive, respectively.

FIG. 4.

The column map of 3-year survival probability and 5-year survival probability for five clinical prognostic factors, including age, stage, and HER2.

4. Discussion

Proper evaluation of the relevant clinical factors for the prognosis of breast cancer is particularly important in the selection of appropriate therapeutic strategies. Screening and identifying the clinically significant factors associated with breast cancer can help predict the risk of disease. To further screen and identify the clinically significant factors associated with breast cancer, the prognostic factors correlated with radiotherapy factors were screened using the Cox risk regression model analysis of prognostic factors. The follow-up data of breast cancer patients were downloaded from METABRIC database, a total of 1980 breast cancer patients were enrolled in this study, including 1173 patients who received the radiotherapy treatment and 807 patients without radiotherapy treatment. The correlation implied that the patients with younger age receiving radiotherapy, grade lower, stage lower, or HER2 negative showed a better clinical survival. Radiotherapy for breast cancer reduces disease recurrence and breast cancer mortality and is an essential part of the multimodality treatment of breast cancer (Recht et al., 1988; Bese et al., 2010; Gonzalez et al., 2010). The response of different clinical features to radiotherapy was also evaluated by survival prognosis analysis and prediction.

The association analysis of the radiotherapy treatment and the clinical prognostic factors implied that the patients with younger age, stage lower, or HER2 negative showed a better overall clinical survival, and the patients who received radiotherapy had a better 3-year survival probability and 5-year survival probability. HER2/erbB-2 belongs to a family of four transmembrane receptors involved in signal transduction pathways that regulate cell growth and differentiation. Overexpression/amplification of HER2 is associated with malignancy and a poor prognosis in breast cancer. HER2 acts as a networking receptor that mediates signaling to cancer cells, causing them to proliferate. HER receptors exist as monomers but dimerize on ligand binding (Harari and Yarden, 2000; Yarden, 2001). Trastuzumab, which was a humanized monoclonal antibody against HER2, has been shown to improve disease-free survival after chemotherapy in women with HER2-positive breast cancer (Piccart-Gebhart et al., 2005; Smith et al., 2007; Arteaga et al., 2011; Cameron et al., 2017). Bevacizumab, a monoclonal antibody against vascular endothelial growth factor A, has shown clinical efficacy in patients with HER2-negative metastatic breast cancer (Von et al., 2012; JPOG, 2014).

In conclusion, screening and identifying the clinically significant factors associated with breast cancer can help predict the risk of disease. Age, stage, or HER2 status was prognostic factors correlated with radiotherapy treatment.

Footnotes

ETHICAL APPROVAL AND CONSENT TO PARTICIPATE

This study was approved by Ethics Committee of The First Hospital of Jilin University.

Authors' Contributions

Conception and design and article writing: W.T. and D.M. Administrative support: L.H. Provision of study materials or patients: X.G. Collection and assembly of data: B.H. Data analysis and interpretation: D.S. Final approval of article: all authors.

Author Disclosure Statement

The authors declare they have no competing financial interests.

Funding Information

No funding was received for this article.

References

Anderson

W.I.

, Schlafer

D.H.

, and Vesely

K.R.

1989. Thyroid follicular carcinoma with pulmonary metastases in a beaver (Castor canadensis). J. Wildl. Dis. 25, 599–600.

Arteaga

C.L.

, Sliwkowski

M.X.

, Osborne

C.K.

, et al. 2011. Treatment of HER2-positive breast cancer: Current status and future perspectives. Nat. Rev. Clin. Oncol. 9, 16.

Bese

N.S.

, Kiel

, Belk

E.G

et al. 2010. Radiotherapy for breast cancer in countries with limited resources: Program implementation and evidence-based recommendations. Breast J. 12, S96–S102.

Budde

M.D.

, Gold

, Jordan

E.K.

, et al. 2012. Differential microstructure and physiology of brain and bone metastases in a rat breast cancer model by diffusion and dynamic contrast enhanced MRI. Clin. Exp. Metastas. 29, 51–62.

Cameron

, Piccartgebhart

M.J.

, Gelber

R.D.

, et al. 2017. 11 years' follow-up of trastuzumab after adjuvant chemotherapy in HER2-positive early breast cancer: Final analysis of the HERceptin Adjuvant (HERA) trial. Lancet, 389, 1195–1205.

DrPH

D.B.T

.M. 1984. Do hormones cause breast cancer?. Cancer, 53, 595–604.

Efird

J.T.

, Hunter

, Chan

, et al. 2018. The association between age, comorbidities and use of radiotherapy in women with breast cancer: Implications for survival. Medicines (Basel), 5, 62.

Elston

C.W.

, and Ellis

I.O.

2010. 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, 19, 403–410.

Eng

K.H.

, Schiller

, and Morrell

2015. On representing the prognostic value of continuous gene expression biomarkers with the restricted mean survival curve. Oncotarget, 6, 36308–36318.

10.

Gilewski

, Ragupathi

, Bhuta

, et al. 2001. Immunization of metastatic breast cancer patients with a fully synthetic globo H conjugate: A phase I trial. Proc. Natl. Acad. Sci. U. S. A. 98, 3270–3275.

11.

Gonzalez

A.B.D.

, Curtis

R.E.

, Gilbert

, et al. 2010. Second solid cancers after radiotherapy for breast cancer in SEER cancer registries. Br. J. Cancer, 102, 220–226.

12.

Harari

, and Yarden

2000. Molecular mechanisms underlying ErbB2/HER2 action in breast cancer. Oncogene, 19, 6102.

13.

Hickey

B.E.

, Francis

D.P.

, and Lehman

2013. Sequencing of chemotherapy and radiotherapy for early breast cancer. Cochrane Database Syst. Rev. 4, CD005212.

14.

Higa

G.M.

, and Fell

R.G.

2013. Sex hormone receptor repertoire in breast cancer. Int. J. Breast Cancer, 2013:284036.

15.

JPOG

.c.p.b.f.b.c. 2014. Neoadjuvant chemotherapy plus bevacizumab for breast cancer. Lancet Oncol. 15, 131–132.

16.

Lacny

, Wilson

, Clement

, et al. 2015. Kaplan-Meier survival analysis overestimates the risk of revision arthroplasty: A meta-analysis. Clin. Orthop. Relat. Res. 473, 3431–3442.

17.

Longenecker

B.M.

, Reddish

, Koganty

, et al. 2010. Immune responses of mice and human breast cancer patients following immunization with synthetic sialyl-Tn conjugated to KLH plus detox adjuvant. Ann. NY Acad. Sci. 690, 276–291.

18.

L.J.W.

, Anderson

K.E.

, Grady

J.J.

, et al. 2000. Decreased ovarian hormones during a soya diet: Implications for breast cancer prevention. Cancer Res. 60, 4112–4121.

19.

Milioli

H.H.

, Vimieiro

, Riveros

, et al. 2015. The discovery of novel biomarkers improves breast cancer intrinsic subtype prediction and reconciles the labels in the METABRIC data set. PLoS One, 10, e0129711.

20.

Müller

, Homey

, Soto

, et al. 2001. Involvement of chemokine receptors in breast cancer metastasis. Nature, 410, 50–56.

21.

Nathanson

K.N.

, Wooster

, and Weber

B.L.

2001. Breast cancer genetics: What we know and what we need. Nat Med. 7, 552–556.

22.

Piccart-Gebhart

M. J.

, Procter

, Leyland-Jones

, et al. 2005. Trastuzumab after adjuvant chemotherapy in HER2-positive breast cancer. N. Engl. J. Med. 353, 1659.

23.

Recht

, Silen

, Schnitt

S.J.

, et al. 1988. Time-course of local recurrence following conservative surgery and radiotherapy for early stage breast cancer. Int. J. Radiat. Oncol. Biol. Phys. 15, 255–261.

24.

Sestak

, Buus

, Cuzick

, et al. 2018. Comparison of the performance of 6 prognostic signatures for estrogen receptor-positive breast cancer: A secondary analysis of a randomized clinical trial. JAMA Oncol, 4:545–553.

25.

Shao

, Grossbard

M.L.

, and Klein

2011. Breast cancer in female-to-male transsexuals: Two cases with a review of physiology and management. Clin Breast Cancer, 11, 417–419.

26.

Smith

, Procter

, Gelber

R.D.

, et al. 2007. 2-year follow-up of trastuzumab after adjuvant chemotherapy in HER2-positive breast cancer: A randomised controlled trial. Lancet, 369, 29–36.

27.

Veronesi

, Boyle

, Goldhirsch

, et al. 2005. Breast cancer. Lancet, 365, 1727–1741.

28.

Von

M. G.

, Eidtmann

, Rezai

, et al. 2012. Neoadjuvant chemotherapy and bevacizumab for HER2-negative breast cancer. New Engl J Med, 366, 299.

29.

Wang

, Wang

, Hang

, et al. 2016. A novel gene expression-based prognostic scoring system to predict survival in gastric cancer. Oncotarget, 7, 55343–55351.

30.

Yarden

2001. Biology of HER2 and its importance in breast cancer. Oncology, 61, 1–13.