OL Oncology Letters 1792-1074 1792-1082 D.A. Spandidos 10.3892/ol.2018.9772 OL-0-0-9772 Articles Use of clinical nomograms for predicting survival outcomes in young women with breast cancer LinHui 1 * ZhangFan 2 * WangLuanhong 3 ZengDe 1 Department of Breast Medical Oncology, Cancer Hospital of Shantou University Medical College, Shantou, Guangdong 515041, P.R. China Guangdong Provincial Key Laboratory for Breast Cancer Diagnosis and Treatment, Cancer Hospital of Shantou University Medical College, Shantou, Guangdong 515041, P.R. China Department of Gynecology, Cancer Hospital of Shantou University Medical College, Shantou, Guangdong 515041, P.R. China Correspondence to: Dr Hui Lin, Department of Breast Medical Oncology, Cancer Hospital of Shantou University Medical College, 7 Raoping Road, Shantou, Guangdong 515041, P.R. China, E-mail: hlin3@stu.edu.cn

Contributed equally

 02 2019 28 11 2018 17 2 1505 1516 04122017 12102018 Copyright: © Lin et al. 2019 This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made.

Early-onset breast cancer (BC) has been recognized to be more aggressive compared with its later counterparts. Survival models of BC in young patients have rarely been reported in previous studies. The current study aimed to establish and validate prediction models with clinicopathological variables for visceral metastasis-free survival (VFS), disease-free-survival (DFS) and overall survival (OS) time in young patients with BC. Clinicopathological data were obtained for 351 patients with primary breast tumors who were ≤40 years old. Univariate and multivariate analyses were performed and nomograms were established to screen and illustrate the prognostic factors. Risk scores were calculated based on coefficients from the Cox regression analysis. Internal validation of the prediction models was conducted by predicting the prognosis of cases randomly sampled from the cohort used in the current study. Multivariate analysis demonstrated that N stage (P=0.004), molecular subtype (P=0.007) and age (P=0.005) were significant independent prognostic factors for VFS. Similarly, N stage (P=0.002) and molecular subtype (P=0.001) were significantly associated with DFS. In addition, N stage (P=0.006), molecular subtype (P=0.006) and the presence of an initially inoperable tumor (P=0.005) were significant independent prognostic factors for OS. According to the Cox regression analysis, nomograms were generated to illustrate the effect of independent prognostic factors on VFS, DFS and OS. Risk scores were calculated and internal validation demonstrated the reliability of the prediction models. In conclusion, N stage and molecular subtype were identified as predictors for VFS, DFS and OS in early-onset BC. Furthermore, an age of <35 years at diagnosis was revealed to be unfavorable for VFS and the presence of an initially inoperable tumor was identified to reduce OS time.

 young patient with breast cancer survival model nomogram visceral metastasis-free survival disease-free survival overall survival Introduction

Breast cancer (BC) is the leading cause of cancer-associated mortality among women worldwide (1). In the past decade, the mortality rate has decreased in the majority of high-income countries; however, the incidence and mortality rates have increased in China (1). This may be due to a number of factors, including the one-child policy, lower cancer screening rates and delays in cancer diagnosis (2). In addition, the median age at diagnosis of BC is 48–50 years in China and 62.9% of patients are premenopausal at that time (2).

BC in younger women has been recognized to be more aggressive and exhibits a worse prognosis compared with BC in older women (3,4). Previous studies have identified that, compared with older patients, younger women with BC present with a larger tumor size, a higher incidence of lymph node involvement (4,5) and an increased 5-year risk of developing metastasis (3,6). Compared with older women, young women exhibit higher proportions of hormone receptor (HR)+/human epidermal growth factor receptor 2 (HER-2)+, HR−/HER2+ and triple-negative BC (5,7). Diverse molecular subtypes usually have distinct disease-free survival (DFS) and overall survival (OS) rates (6,8), and age has been identified to serve different roles (9,10). Clinicians use certain risk scores in clinical practice, including the commonly used St Gallen risk factor grading system (11). In this grading system, age is one of the most valuable factors, which suggests that similar to estrogen receptor (ER) status and lymph node status, age is fundamental in predicting BC prognosis. Previous studies have predominantly focused on the clinicopathological features of BC in young patients (3,12). However, to the best of our knowledge, a survival model remains to be established. The current study investigated a number of factors, including T stage, N stage, pathological type, grade, surgical type, neoadjuvant chemotherapy, age and molecular subtype, for predicting survival in young patients with BC. The study aimed to assess an array of clinicopathological variables that are potentially associated with visceral metastasis-free survival (VFS), DFS and OS. In addition, the ultimate aim of the study was to establish and validate prediction models for survival outcomes in young patients with BC.

 Patients and methods <sec> <title>Definition of a young patient with BC

The definition of a young patient with BC varies among previous studies. Previously, the upper age limit has ranged from 35 (13) to 40 years (14,15). The current study defined young BC as patients ≤40 years old at preliminary diagnosis.

 Study population

A total of 351 females with primary BC who were diagnosed at ≤40 years old and treated at the Cancer Hospital of Shantou University Medical College (Guangdong, China) between April 2009 and May 2014 were included in the current study. The inclusion criteria were: i) female; ii) breast cancer confirmed by pathological diagnosis; and iii) age ≤40 years old. Patients with distant metastasis at primary diagnosis and patients with a follow-up time <6 months were excluded. The mean age of the patients was 35.74 years with a range of 19 to 40 years. Every patient had undergone mammographic and/or ultrasound radiological imaging, a chest radiograph or computed tomography scan of the chest, Doppler ultrasound examination or a computed tomography scan of the abdomen, a complete blood count test and blood biochemistry assays to evaluate the primary tumor stage and the appropriate treatment. Bone scans and brain magnetic resonance imaging were performed if patients experienced bone pain, central nervous symptoms or exhibited a locally advanced stage of BC. Patients with primary resectable tumors received a mastectomy or breast-conserving surgery with axillary lymph node dissection or sentinel lymph node biopsy. A core needle biopsy was performed in a standardized manner when the surgeon identified that a tumor was inoperable. Neoadjuvant chemotherapy was administered to patients with initially inoperable tumors, the majority of which were stages T3/T4 and/or N2/N3 according to the 7th edition of the American Joint Committee on Cancer staging system (16), to increase the possibility of radical surgeries later on. The requirement of adjuvant chemotherapy and the protocol of the chemotherapy treatment were guided by the St. Gallen BC guidelines (11).

Clinical and pathological data were collected from patient records. Histopathological features of surgical resection specimens included tumor type and size, histological grade, evidence of lymphovascular invasion and axillary nodal status. ER, progesterone receptor (PR), HER-2, Ki-67 and other markers were stained in the majority of the biopsy and resection specimens. Adjuvant radiotherapy, chemotherapy, endocrine treatment and targeted treatment were recorded. In addition, other basic information, including age of menarche, fertility status, hepatitis B virus (HBV) infection and family history were recorded. Follow-up information was obtained from patient records. The median follow-up time was 38.3 months (range, 6.0–106.6 months).

Written informed consent was obtained from all participants for the use of clinicopathological data. The current study was approved by the Ethics Committee of the Cancer Hospital of Shantou University Medical College.

 Classification of survival and molecular subtypes

VFS was defined as the time from radical surgery to visceral metastasis, excluding local relapse and metastasis of the lymph nodes and bones. DFS was defined as the time from radical surgery to disease relapse or metastasis, including visceral metastasis. OS was defined as the time from diagnosis to mortality from any cause. Molecular subtypes were differentiated according to the status of ER, PR and HER-2, as determined by immunohistochemistry (IHC). As the cut-off value of Ki-67 has not previously been determined (17) and since testing for Ki-67 was not routinely performed in the study period, the current study did not use Ki-67 for the classification of molecular subtypes. The molecular subtypes were defined as follows: The luminal A subtype, which was HER-2−, ER+ and/or PR+; the luminal B subtype, which was HER-2+, ER+ and/or PR+; the HER-2+ subtype, which was HER-2+, ER− and PR−; and the triple-negative subtype, which was HER-2, ER and PR. HER-2 positivity was defined as HER-2 gene amplification in a fluorescence in situ hybridization test or HER-2 protein stained as ‘+++’ in IHC, as described previously (18).

 Statistical analysis

All statistical analyses were performed using SPSS software (version 13.0; SPSS Inc., Chicago, IL, USA) and R software (version 3.3.0; www.r-project.org). The univariate analysis for assessing the prognostic factors was performed using the Kaplan-Meier method with a log-rank test. Variables associated with survival (P<0.05) were selected for multivariate Cox regression analysis using forward stepwise selection. Nomograms were then generated to illustrate the effect of the prognostic factors on DFS, VFS and OS. Risk scores were created based on Cox regression coefficients. Each patient was assigned a risk score that was a linear combination of the values of the independent prognostic factors weighted by their respective Cox regression coefficients (19). Internal validation of the prediction models was performed by evaluating the accuracy of the risk score on the prognosis of 200, 250 and 300 patients who were randomly selected from the total 351 patients. P<0.05 was considered to indicate a statistically significant difference.

 Results <sec> <title>Univariate survival analysis for predicting DFS, VFS and OS in young patients with BC

To preliminarily determine the potential prognostic factors, univariate survival analysis was performed for VFS, DFS and OS. The median follow-up time was 38.3 months and the median values for VFS, DFS and OS were 38.0, 33.5 and 38.2 months, respectively. The variables included in the analysis were age, T stage, N stage, M stage, site of involvement, pathological type, differentiation grade, molecular subtype, surgical type, neoadjuvant chemotherapy, adjuvant radiation, age of menarche, fertility status, HBV infection and family history.

The 1-, 3- and 5-year VFS rates were 94.5, 87.6 and 80.6%, respectively. The 1-, 3- and 5-year DFS rates were 89.8, 76.2 and 64.6%, respectively. The 1-, 3- and 5-year OS rates were 98.2, 87.4 and 73.3%, respectively. Survival rates for different clinicopathological features were analyzed and tested with Kaplan-Meier analysis and a log-rank test (Tables I–III). This analysis identified that for VFS, N stage (P=0.004), molecular subtype (P=0.007), age (P=0.005), T stage (P=0.014), pathological type (P=0.029) and neoadjuvant chemotherapy (P=0.020) were statistically significant variables. For DFS, N stage (P=0.002) and molecular subtype (P=0.001) were statistically significant. For OS, T stage (P=0.029), N stage (P=0.006), M stage (P=0.002), molecular subtype (P=0.006), surgical type (P<0.001) and neoadjuvant chemotherapy (P=0.005) were statistically significant variables.

 Multivariate survival analysis for predicting VFS, DFS and OS in young patients with BC

To further analyze the prognostic factors for VFS, DFS and OS, multivariate survival analysis was performed. Variables revealed as statistically significant by Kaplan-Meier analysis (P<0.05) were selected for Cox regression analysis to identify independent factors. As presented in Table IV, the variables analyzed for VFS were as follows: N stage (P<0.001); molecular subtype (P=0.027); and age (P<0.001). As presented in Table V, the variables analyzed for DFS included: N stage (P=0.004) and molecular subtype (P=0.002). As presented in Table VI, the variables analyzed for OS were as follows: N stage (P=0.029), molecular subtype (P=0.006) and neoadjuvant chemotherapy (P=0.006). Nomograms were created to illustrate the effect of the prognostic factors on VFS, DFS and OS using multivariate Cox regression coefficients (Figs. 1–3).

 Risk scores for predicting survival outcomes in young patients with BC

Based on the regression analysis, prediction models for VFS, DFS and OS were generated through the calculations of risk scores, previously established by Shukla et al (19). Each patient was assigned a risk score; a linear combination of the values of the independent prognostic factors weighted by their respective Cox regression coefficients. Risk scores for VFS were calculated as follows: Risk score = 1.091 × N stage (N1/N0) + 1.499 × N stage (N2/N0) + 2.163 × N stage (N3/N0) + 0.355 × molecular subtype (luminal B/luminal A) + 1.087 × molecular subtype (HER-2/luminal A) + 1.016 × molecular subtype (triple-negative/luminal A) + 1.319 × age (<35/≥35). Risk scores for DFS were calculated as follows: Risk score = 0.555 × N stage (N1/N0) + 0.831 × N stage (N2/N0) + 1.112 × N stage (N3/N0) + 0.613 × molecular subtype (luminal B/luminal A) + 1.109 × molecular subtype (HER-2/luminal A) + 0.665 × molecular subtype (triple-negative/luminal A). Risk scores for OS were calculated as follows: Risk score = 0.050 × N stage (N1/N0) + 0.636 × N stage (N2/N0) + 1.166 × N stage (N3/N0) - 0.033 × molecular subtype luminal B/luminal A) + 1.033 × molecular subtype (HER-2/luminal A) + 1.182 × molecular subtype (triple-negative/luminal A) + 1.001 × neoadjuvant chemotherapy (yes/no).

Internal validation of the prediction models was conducted by evaluating the effect of the risk score on the prognosis of patients. A total of 200, 250 and 300 cases were randomly selected 10 times from the total 351 cases and univariate Cox proportional hazard regression analysis was performed. As presented in Tables VII–IX, the range of the HR was 1.692–2.239 for VFS with P≤0.005, 1.910–2.879 for DFS with P≤0.003 and 1.938–2.652 for OS with P≤0.003. Therefore, risk scores and nonograms were demonstrated to be reliable for predicting VFS, DFS and OS time in young patients with BC.

 Discussion

China has a high prevalence of young patients with BC, who exhibit a poor prognosis (2). A number of studies have demonstrated that age (3,6,8,9,20) and molecular subtype (4,7) are associated with survival in these patients, in addition to a larger tumor size, higher incidence of lymph node involvement (4,5) and higher incidence of poorly differentiated tumors (4,5). However, to the best of our knowledge, a prediction model for these patients has not been established. Nomograms are widely used to present prediction models for a number of cancer types (21–23). Due to their distinctness and clarity, nomograms are useful for patients to understand the prognosis of their disease and for doctors to decide the most appropriate treatment protocol. Nomograms have been generated for BC to predict the outcome of patients who have undergone neoadjuvant chemotherapy (24) and of patients with advanced tumors (21). In addition, nomograms have been established to predict axillary lymph node status (25) and loco-regional recurrence (26), thus assisting surgeons with the decision of surgical type. The current study created and displayed survival models as nomograms to predict the outcome of young patients with BC.

In the current study, the prediction model for DFS included two independent variables, N stage and molecular subtype, which was consistent with a previous study (8). N stage represented the tumor burden and the capacity of metastasis, while the molecular subtype represented the biological characteristics of the tumor. Patients with the luminal A subtype exhibited the longest DFS time, while patients with the HER-2+ subtype exhibited the worst prognosis. A significant difference was identified in the DFS between these molecular subtypes, as demonstrated in previous studies (8,27).

Notably, to the best of our knowledge, the current study is the first to introduce the concept of VFS for breast cancer, which is defined as the time from radical surgery to the first visceral metastasis or mortality. Previous studies have typically used the concept of distant recurrence-free survival (DDFS) (28), which is defined as the time from radical surgery to the first distant metastasis or mortality. The difference between DDFS and VFS is the metastatic sites. Bone metastasis and distant lymph node metastasis are included in DDFS, but not in VFS. Savci-Heijink et al (29) reported that BC cases without visceral metastasis exhibited improved survival rates compared with those with visceral metastasis. It was identified that patients with local relapse, lymph node metastasis and bone metastasis exhibited improved survival rates compared with patients with visceral metastasis. Therefore, the current study assumed that VFS was a valuable measurement for prognostic prediction. The current study identified that VFS was associated with molecular subtype, N stage and age, but not local relapse, bone metastasis and lymph node metastasis. This result differed from the prediction model for DFS time, as age at diagnosis was identified as an independent predictor for VFS time. Previous studies revealed that a younger age is associated with a more aggressive cancer that is more likely to metastasize to visceral organs (3–6) Additionally, a previous study demonstrated that age is an independent predictor of DFS and OS time (30). The current study also demonstrated that age (<35 years) was negatively associated with VFS.

Furthermore, molecular subtype has previously been associated with patterns of metastasis (29,31). Patients with certain molecular subtypes, including ER− and HER-2+ subtypes, have been associated with visceral metastasis, while patients with an ER+ subtype have been associated with bone metastasis (29,31,32). The current study revealed that patients with the luminal A subtype experienced the longest VFS time, while patients with the HER-2+ subtype experienced the shortest VFS time and the highest frequency of visceral metastasis. The unfavorable outcome of patients with the HER-2+ subtype may partially be due to the low percentage of patients in this group who experienced targeted treatment. However, by July 2017 >75,000 patients with HER-2+ breast cancer in China benefited from the Herceptin Patient Assistance Program and received targeted treatment (unpublished data), which may increase their survival rates.

The current study identified that N stage, molecular subtype and neoadjuvant chemotherapy were associated with OS. N stage and molecular subtype have been associated with OS in previous studies (8,28,29,31). However, a significant association between OS and neoadjuvant chemotherapy was also identified in the current study. To the best of our knowledge, this result has not previously been reported. In the current study, only 1 patient received neoadjuvant chemotherapy prior to breast conservation surgery. The remaining 45 cases received neoadjuvant chemotherapy due to the presence of initially inoperable tumors. The prediction model demonstrated that patients with a HER-2+ subtype, an advanced N stage or an initially inoperable tumor exhibited unfavorable OS.

According to the survival analysis, nomograms were created and risk scores (19) were calculated based on the Cox regression coefficients for VFS, DFS and OS time. Internal validation was performed in patients randomly sampled from the total population. This validation demonstrated that the risk scores were associated with VFS, DFS and OS time. This suggests that the nomograms constructed following Cox regression analysis were reliable. However, the lack of a validation cohort is a limitation of the current study. Future studies should collect a larger number of cases to further validate the nomograms.

In conclusion, the current study constructed and validated survival models displayed as nomograms to predict VFS, DFS and OS time in young patients with BC using retrospective data from patients <40 years old at diagnosis. In addition, the concept of VFS was introduced. Molecular subtype and N stage were identified as independent predictors for VFS, DFS and OS time. Age at diagnosis was revealed to independently predict VFS and neoadjuvant chemotherapy was identified as an unfavorable factor for OS. Risk scores based on these survival models were established for young patients with BC. These survival models were validated and the current study recommends their use in the survival analysis of young patients with BC in the future.

 Acknowledgements

The authors would like to thank Professor William Au from Shantou University Medical College (Shantou, China) for providing assistance in editing the original manuscript.

 Funding

The current study was supported by the Shantou Health and Technology Program (grant no. 123).

 Availability of data and materials

The datasets used and analyzed during the current study are available from the corresponding author on reasonable request.

 Authors' contributions

HL and FZ designed the study. FZ conducted the statistical analysis. HL and FZ analyzed and interpreted the data. HL, FZ, DZ and LW were involved in the data acquisition. HL, FZ, DZ and LW wrote the manuscript. All authors have read and approved the final submitted manuscript. HL takes final responsibility.

 Ethics approval and consent to participate

Written informed consent was obtained from all participants for the use of clinicopathological information. The current study was approved by the Ethics Committee of the Cancer Hospital of Shantou University Medical College (Guangdong, China).

 Patient consent for publication

Not applicable.

 Competing interests

The authors declare that they have no competing interests.

 Abbreviations BC

breast cancer

 VFS

visceral metastasis-free survival

 DFS

disease-free survival

 OS

overall survival

 ER

estrogen receptor

 PR

progesterone receptor

 HBV

hepatitis B virus

 IHC

immunohistochemistry

 References DeSantisCEBrayFFerlayJLortet-TieulentJAndersonBOJemalAInternational variation in female breast cancer incidence and mortality ratesCancer Epidemiol Biomarkers Prev2414951506201510.1158/1055-9965.EPI-15-053526359465 FanLStrasser-WeipplKLiJJSt LouisJFinkelsteinDMYuKDChenWQShaoZMGossPEBreast cancer in ChinaLancet Oncol15E279E289201410.1016/S1470-2045(13)70567-924872111 TangLCYinWJDiGHShenZZShaoZMUnfavourable clinicopathologic features and low response rate to systemic adjuvant therapy: Results with regard to poor survival in young Chinese breast cancer patientsBreast Cancer Res Treat12295104201010.1007/s10549-009-0537-419760032 ZabickiKColbertJADominguezFJGaddMAHughesKSJonesJLSpechtMCMichaelsonJSSmithBLBreast cancer diagnosis in women <or=40 versus 50 to 60 years: Increasing size and stage disparity compared with older women over timeAnn Surg Oncol1310721077200610.1245/ASO.2006.03.05516865599 GoksuSSTastekinDArslanDGunduzSTatliAMUnalDSalimDGulerTCoskunHSClinicopathologic features and molecular subtypes of breast cancer in young women (age </=35)Asian Pac J Cancer prev1566656668201410.7314/APJCP.2014.15.16.666525169505 TjokrowidjajaALeeCKHoussamiNLordSMetastatic breast cancer in young women: A population-based cohort study to describe risk and prognosisIntern Med J44764770201410.1111/imj.1248124863750 KeeganTHDeRouenMCPressDJKurianAWClarkeCAOccurrence of breast cancer subtypes in adolescent and young adult womenBreast Cancer Res14R55201210.1186/bcr315622452927 ChenHLDingAWangFWPrognostic effect analysis of molecular subtype on young breast cancer patientsChin J Cancer Res27428436201526361413 LiedtkeCRodyAGluzOBaumannKBeyerDKohlsEBLausenKHankerLHoltrichUBeckerSKarnTThe prognostic impact of age in different molecular subtypes of breast cancerBreast Cancer Res Treat152667673201510.1007/s10549-015-3491-326195120 AzimHAJrNguyenBBroheeSZoppoliGSotiriouCGenomic aberrations in young and elderly breast cancer patientsBMC Med13266201510.1186/s12916-015-0504-326467651 GoldhirschAWoodWCGelberRDCoatesASThürlimannBSennHJ10th StGallen conference: Progress and promise: Highlights of the international expert consensus on the primary therapy of early breast cancer 2007Ann Oncol1811331144200710.1093/annonc/mdm27117675394 PlichtaJKRaiUTangRCoopeySBBuckleyJMGaddMASpechtMCHughesKSTaghianAGSmithBLFactors associated with recurrence rates and long-term survival in women diagnosed with breast cancer ages 40 and youngerAnn Surg Oncol2332123220201610.1245/s10434-016-5404-z27406095 FredholmHEakerSFrisellJHolmbergLFredrikssonILindmanHBreast cancer in young women: Poor survival despite intensive treatmentPLoS One4e7695200910.1371/journal.pone.000769519907646 HearneBJTeareMDButtMDonaldsonLComparison of nottingham prognostic index and adjuvant online prognostic tools in young women with breast cancer: Review of a single-institution experienceBMJ open5e005576201510.1136/bmjopen-2014-00557625628047 KarihtalaPWinqvistRBloiguRJukkola-VuorinenALong-term observational follow-up study of breast cancer diagnosed in women </=40 years oldBreast19456461201010.1016/j.breast.2010.05.00120570150 EdgeSBComptonCCThe American joint committee on cancer: The 7th edition of the AJCC cancer staging manual and the future of TNMAnn Surg Oncol1714711474201010.1245/s10434-010-0985-420180029 AlcoGBozdoganASelamogluDPilanciKNTuzlaliSOrduCIgdemSOkkanSDincerMDemirGOzmenVClinical and histopathological factors associated with Ki-67 expression in breast cancer patientsOncol Lett910461054201510.3892/ol.2015.285225663855 WangSSaboorianMHFrenkelEHynanLGokaslanSTAshfaqRLaboratory assessment of the status of Her-2/neu protein and oncogene in breast cancer specimens: Comparison of immunohistochemistry assay with fluorescence in situ hybridisation assaysJ Clin Pathol53374381200010.1136/jcp.53.5.37410889820 ShuklaSPia PatricIRThinagararjanSSrinivasanSMondalBHegdeASChandramouliBASantoshVArivazhaganASomasundaramKA DNA methylation prognostic signature of glioblastoma: Identification of NPTX2-PTEN-NF-kappaB nexusCancer Res7365636573201310.1158/0008-5472.CAN-13-029824078801 TangLCJinXYangHYHeMChangHShaoZMDiGHLuminal B subtype: A key factor for the worse prognosis of young breast cancer patients in ChinaBMC Cancer15201201510.1186/s12885-015-1207-z25885213 LeeCKHudsonMStocklerMCoatesASAcklandSGebskiVLordSFriedlanderMBoyleFSimesRJA nomogram to predict survival time in women starting first-line chemotherapy for advanced breast cancerBreast cancer Res Treat129467476201110.1007/s10549-011-1471-921445568 VernereyDHuguetFVienotAGoldsteinDPaget-BaillySVan LaethemJLGlimeliusBArtruPMooreMJAndréTPrognostic nomogram and score to predict overall survival in locally advanced untreated pancreatic cancer (PROLAP)Br J Cancer115281289201610.1038/bjc.2016.21227404456 LiangWZhangLJiangGWangQLiuLLiuDWangZZhuZDengQXiongXDevelopment and validation of a nomogram for predicting survival in patients with resected non-small-cell lung cancerJ Clin Oncol33861869201510.1200/JCO.2014.56.666125624438 KeamBImSAParkSNamBHHanSWOhDYKimJHLeeSHHanWKimDWNomogram predicting clinical outcomes in breast cancer patients treated with neoadjuvant chemotherapyJ Cancer Res Clin Oncol13713011308201110.1007/s00432-011-0991-321717200 QiuSQZengHCZhangFChenCHuangWHPleijhuisRGWuJDvan DamGMZhangGJA nomogram to predict the probability of axillary lymph node metastasis in early breast cancer patients with positive axillary ultrasoundSci Rep621196201610.1038/srep2119626875677 WitteveenAVliegenIMSonkeGSKlaaseJMMJIJSieslingSPersonalisation of breast cancer follow-up: A time-dependent prognostic nomogram for the estimation of annual risk of locoregional recurrence in early breast cancer patientsBreast Cancer Res Treat152627636201510.1007/s10549-015-3490-426162567 SongNChoiJYSungHJeonSChungSParkSKHanWLeeJWKimMKLeeJYPrediction of breast cancer survival using clinical and genetic markers by tumor subtypesPLoS One10e0122413201510.1371/journal.pone.012241325867717 HennigsARiedelFGondosASinnPSchirmacherPMarméFJägerDKauczorHUStieberALindelKPrognosis of breast cancer molecular subtypes in routine clinical care: A large prospective cohort studyBMC Cancer16734201610.1186/s12885-016-2766-327634735 Savci-HeijinkCDHalfwerkHHooijerGKHorlingsHMWesselingJvan de VijverMJRetrospective analysis of metastatic behaviour of breast cancer subtypesBreast Cancer Res Treat150547557201510.1007/s10549-015-3352-025820592 ZhaoYDongXLiRSongJZhangDCorrelation between clinical-pathologic factors and long-term follow-up in young breast cancer patientsTransl Oncol8265272201510.1016/j.tranon.2015.05.00126310372 KastKLinkTFriedrichKPetzoldANiedostatekASchofferOWernerCKlugSJWernerAGatzweilerAImpact of breast cancer subtypes and patterns of metastasis on outcomeBreast Cancer Res Treat150621629201510.1007/s10549-015-3341-325783184 BartmannCDiessnerJBlettnerMHäuslerSJanniWKreienbergRKrockenbergerMSchwentnerLSteinRStüberTFactors influencing the development of visceral metastasis of breast cancer: A retrospective multi-center studyBreast316675201710.1016/j.breast.2016.10.01627816834

Nomogram for predicting VFS in young patients with breast cancer. VFS, visceral metastasis-free survival; HER-2, human epidermal growth factor receptor 2.

Nomogram for predicting DFS in young patients with breast cancer. DFS, disease-free survival; HER-2, human epidermal growth factor receptor 2.

Nomogram for predicting OS in young patients with breast cancer. OS, overall survival; HER-2, human epidermal growth factor receptor 2.

Clinicopathological characteristics of patients and the associated 1-, 3- and 5-year VFS rates.

VFS, %
Characteristic Cases, n (%) 1-year 3-year 5-year P-value
Age, years 0.005
  <35 108 (30.8) 93.4 82.8 67.0
  ≥35 243 (69.2) 95.0 89.6 85.3
T stage 0.014
  Tis 1 (0.3)
  T1 64 (18.2) 96.4 89.0 81.8
  T2 163 (46.4)
  T3 52 (14.8) 88.2 79.5 74.2
  T4 25 (7.1)
  Unknown 46 (13.1)
N stage 0.004
  N0 144 (41.0) 97.2 93.7 92.2
  N1 80 (22.8) 96.2 87.1 74.2
  N2 57 (16.2) 94.6 88.3 71.9
  N3 56 (16.0) 83.5 73.3 73.3
  Unknown 14 (4.0)
M stage 0.544
  M0 339 (96.6) 94.6 87.5 80.4
  M1a 6 (1.7) 83.3 83.3 0.0
  Unknown 6 (1.7)
Site of involvement 0.596
  Left 177 (50.4) 94.8 89.4 81.2
  Right 166 (47.3) 93.8 86.2 80.8
  Bilateral 8 (2.3) 100.0 75.0 75.0
Pathological type 0.029
  IDC 290 (82.6) 93.3 85.3 76.7
  ILC 11 (3.1) 100.0 100.0 100.0
  DCIS 22 (6.3) 100.0 100.0 100.0
  Other 27 (7.7) 100.0 96.0 96.0
  Unknown 1 (0.3)
Grade 0.063
  I 15 (4.3) 100.0 100.0 100.0
  II 103 (29.3) 92.0 88.5 86.2
  III 96 (27.4) 93.7 78.2 70.6
  Unknown 137 (39.0)
Molecular subtype 0.007
  Luminal A 161 (45.9) 98.1 90.9 85.9
  Luminal B 40 (11.4) 97.4 92.0 78.7
  HER-2+ 47 (13.4) 80.6 70.8 66.6
  Triple-negative 65 (18.5) 90.7 83.0 75.4
  Unknown 38 (10.8)
Surgical type 0.120
  Modified radical mastectomy 276 (78.6) 93.0 86.3 78.2
  Breast-conserving surgery 58 (16.5) 100.0 92.6 92.6
  Mastectomy and SLNB 12 (3.4) 100.0 100.0 100.0
  Simple resectionb 5 (1.4) 100.0 75.0 −c
Neoadjuvant chemotherapy 0.020
  Yes 46 (13.1) 86.9 76.5 69.6
  No 305 (86.9) 95.6 89.2 82.1
Adjuvant radiation 0.399
  Yes 190 (54.1) 92.9 84.5 80.3
  No 161 (45.9) 96.3 91.1 81.3
Age of menarche, years 0.934
  ≤15 252 (71.8) 95.1 88.3 79.2
  >15 50 (14.2) 94.0 86.6 86.6
  Unknown 49 (14.0)
Fertility status 0.566
  Yes 323 (92.0) 94.3 87.6 80.5
  No 27 (7.7) 96.3 87.4 80.1
  Unknown 1 (0.3)
HBV infection 0.477
  Yes 9 (2.6) 88.9 74.1 74.1
  No 342 (97.4) 94.6 87.9 80.7
Family history 0.143
  BC 8 (2.3) 100.0 100.0 50.0
  Other cancer types 13 (3.7) 100.0 64.1 64.1
  No 330 (94.0) 94.1 88.2 82.2

Patients with bone metastasis at diagnosis

patients received simple resection in another hospital prior to administration

censored data. VFS, visceral metastasis-free survival; Tis, tumor in situ; IDC, invasive ductal carcinoma; ILC, invasive lobular carcinoma; DCIS, ductal carcinoma in situ; HER-2, human epidermal growth factor receptor 2; SLNB, sentinel lymph node biopsy; BC, breast cancer; HBV, hepatitis B virus.

Clinicopathological characteristics of patients and the associated 1-, 3- and 5-year DFS rates.

DFS, %
Characteristic Cases, n (%) 1-year 3-year 5-year P-value
Age, years 0.241
  <35 108 (30.8) 86.7 74.1 61.1
  ≥35 243 (69.2) 91.1 77.2 65.5
T stage 0.053
  Tis 1 (0.3)
  T1 64 (18.2) 91.9 79.2 68.8
  T2 163 (46.4)
  T3 52 (14.8) 83.9 68.7 59.0
  T4 25 (7.1)
  Unknown 46 (13.1)
N stage 0.002
  N0 144 (41.0) 93.0 86.5 79.8
  N1 80 (22.8) 93.6 76.5 60.4
  N2 57 (16.2) 91.0 72.2 46.9
  N3 56 (16.0) 75.3 61.2 61.2
  Unknown 14 (4.0)
M stage 0.102
  M0 339 (96.6) 90.4 76.8 66.1
  M1a 6 (1.7)
  Unknown 6 (1.7)
Site of involvement 0.178
  Left 177 (50.4) 90.2 78.2 65.3
  Right 166 (47.3) 89.4 75.4 64.7
  Bilateral 8 (2.3) 87.5 50.0 50.0
Pathological type 0.078
  IDC 290 (82.6) 89.8 73.9 61.0
  ILC 11 (3.1) 100.0 100.0 100.0
  DCIS 22 (6.3) 95.0 95.0 76.0
  Other 27 (7.7) 80.7 76.7 76.7
  Unknown 1 (0.3)
Grade 0.241
  I 15 (4.3) 86.2 79.0 79.0
  II 103 (29.3) 86.1 80.1 72.1
  III 96 (27.4) 88.2 65.2 59.3
  Unknown 137 (39.0)
Molecular subtype 0.001
  Luminal A 161 (45.9) 96.1 81.4 72.0
  Luminal B 40 (11.4) 89.9 76.6 46.6
  HER-2+ 47 (13.4) 69.2 55.0 50.8
  Triple-negative 65 (18.5) 86.0 71.4 64.9
  Unknown 38 (10.8)
Surgical type 0.077
  Modified radical mastectomy 276 (78.6) 88.9 74.3 62.1
  Breast-conserving surgery 58 (16.5) 93.0 84.1 77.1
  Mastectomy and SLNB 12 (3.4) 100.0 100.0 100.0
  Simple resectionb 5 (1.4) 80.0 53.3 −c
Neoadjuvant chemotherapy 0.051
  Yes 46 (13.1) 81.7 63.0 56.7
  No 305 (86.9) 90.9 78.2 65.7
Adjuvant radiation 0.081
  Yes 190 (54.1) 91.8 77.6 71.6
  No 161 (45.9) 87.4 74.5 57.1
Age of menarche, years 0.494
  ≤15 252 (71.8) 91.0 78.1 63.6
  >15 50 (14.2) 83.7 71.5 67.8
  Unknown 49 (14.0)
Fertility status 0.966
  Yes 323 (92.0) 89.2 76.1 63.8
  No 27 (7.7) 96.3 76.0 69.1
  Unknown 1 (0.3)
HBV infection 0.584
  Yes 9 (2.6) 88.9 74.1 49.4
  No 342 (97.4) 89.8 76.3 65.1
Family history 0.139
  BC 8 (2.3) 100.0 85.7 57.1
  Other cancer types 13 (3.7) 83.3 53.3 26.7
  No 330 (94.0) 89.7 76.9 66.3

Patients with bone metastasis at diagnosis

patients received simple resection in another hospital prior to administration

censored data. DFS, disease-free survival; Tis, tumor in situ; IDC, invasive ductal carcinoma; ILC, invasive lobular carcinoma; DCIS, ductal carcinoma in situ; HER-2, human epidermal growth factor receptor 2; SLNB, sentinel lymph node biopsy; BC, breast cancer; HBV, hepatitis B virus.

Clinicopathological characteristics of patients and the associated 1-, 3- and 5-year OS rates.

OS, %
Characteristic Cases, n (%) 1-year 3-year 5-year P-value
Age, years 0.387
  <35 108 (30.8) 96.2 84.6 71.5
  ≥35 243 (69.2) 99.6 88.4 75.9
T stage 0.029
  Tis 1 (0.3)
  T1 64 (18.2) 99.5 90.5 80.2
  T2 163 (46.4)
  T3 52 (14.8) 94.7 80.8 73.7
  T4 25 (7.1)
  Unknown 46 (13.1)
N stage 0.006
  N0 144 (41.0) 98.6 93.2 84.2
  N1 80 (22.8) 98.7 89.7 81.5
  N2 57 (16.2) 98.2 85.9 64.6
  N3 56 (16.0) 98.1 78.2 64.9
  Unknown 14 (4.0)
M stage 0.002
  M0 339 (96.6) 98.5 88.9 75.9
  M1a 6 (1.7) 100.0 50.0 −c
  Unknown 6 (1.7)
Site of involvement 0.439
  Left 177 (50.4) 98.2 90.6 77.3
  Right 166 (47.3) 98.8 83.9 71.4
  Bilateral 8 (2.3) 100.0 75.0 75.0
Pathological type 0.289
  IDC 290 (82.6) 98.2 86.2 71.8
  ILC 11 (3.1) 100.0 100.0 75.0
  DCIS 22 (6.3) 100.0 95.0 95.0
  Other 27 (7.7) 100.0 86.6 86.6
  Unknown 1 (0.3)
Grade 0.103
  I 15 (4.3) 100.0 100.0 100.0
  II 103 (29.3) 98.0 86.3 82.3
  III 96 (27.4) 97.9 81.3 69.1
  Unknown 137 (39.0)
Molecular subtype 0.006
  Luminal A 161 (45.9) 100.0 90.8 78.9
  Luminal B 40 (11.4) 97.4 94.9 81.9
  HER-2+ 47 (13.4) 97.8 71.8 57.1
  Triple-negative 65 (18.5) 95.2 81.5 65.3
  Unknown 38 (10.8)
Surgical type <0.001
  Modified radical mastectomy 276 (78.6) 98.1 86.8 72.9
  Breast-conserving surgery 58 (16.5) 100.0 95.3 90.8
  Mastectomy and SLNB 12 (3.4) 100.0 88.9 −c
  Simple resectionb 5 (1.4) 100.0 0.0 0.0
Neoadjuvant chemotherapy 0.005
  Yes 46 (13.1) 93.3 70.9 63.0
  No 305 (86.9) 99.3 89.8 76.3
Adjuvant radiation 0.559
  Yes 190 (54.1) 98.9 90.0 74.9
  No 161 (45.9) 98.1 84.2 74.3
Age of menarche, years 0.193
  ≤15 252 (71.8) 98.3 88.6 76.6
  >15 50 (14.2) 98.0 83.8 69.8
  Unknown 49 (14.0)
Fertility status 0.849
  Yes 323 (92.0) 98.4 87.4 73.4
  No 27 (7.7) 96.2 86.5 71.9
  Unknown 1 (0.3)
HBV infection 0.592
  Yes 9 (2.6) 100.0 70.0 70.0
  No 342 (97.4) 98.2 87.8 73.4
Family history 0.986
  BC 8 (2.3) 100.0 100.0 66.7
  Other cancer types 13 (3.7) 100.0 88.9 74.1
  No 330 (94.0) 98.1 87.0 73.5

Patients with bone metastasis at diagnosis

patients received simple resection in another hospital prior to administration

censored data. OS, overall survival; Tis, tumor in situ; IDC, invasive ductal carcinoma; ILC, invasive lobular carcinoma; DCIS, ductal carcinoma in situ; HER-2, human epidermal growth factor receptor 2; SLNB, sentinel lymph node biopsy; BC, breast cancer; HBV, hepatitis B virus.

Cox regression analysis for predicting visceral metastasis-free survival.

95% CI
Characteristic HR P-value Lower Upper
N stage <0.001
  N1/N0 2.977 0.025 1.148 7.722
  N2/N0 4.477 0.003 1.641 12.211
  N3/N0 8.695 <0.001 0.296 22.937
Molecular subtype 0.027
  Luminal B/luminal A 1.426 0.536 0.463 4.390
  HER-2/luminal A 2.965 0.007 1.342 6.552
  Triple-negative/luminal A 2.763 0.017 1.201 6.353
Age, years
  <35/≥35 3.739 <0.001 1.905 7.338

HR, hazard ratio; CI, confidence interval; HER-2, human epidermal growth factor receptor 2.

Cox regression analysis for predicting disease-free survival.

95% CI
Characteristic HR P-value Lower Upper
N stage 0.004
  N1/N0 1.742 0.082 0.933 3.253
  N2/N0 2.295 0.009 1.230 4.284
  N3/N0 3.041 0.001 1.621 5.704
Molecular subtype 0.002
  Luminal B/luminal A 1.846 0.090 0.908 3.751
  HER-2/luminal A 3.030 <0.001 1.707 5.379
Triple-negative/luminal A 1.944 0.029 1.071 3.528

HR, hazard ratio; CI, confidence interval; HER-2, human epidermal growth factor receptor 2.

Cox regression analysis for predicting overall survival.

95% CI
Characteristic HR P-value Lower Upper
N stage 0.029
  N1/N0 1.052 0.920 0.392 2.822
  N2/N0 1.888 0.193 0.725 4.921
  N3/N0 3.210 0.009 1.347 7.653
Molecular subtype 0.006
  Luminal B/luminal A 0.968 0.959 0.276 3.399
  HER-2/luminal A 2.809 0.009 1.290 6.119
  Triple-negative/luminal A 3.262 0.003 1.504 7.075
Neoadjuvant chemotherapy
  Yes/no 2.722 0.006 1.336 5.543

HR, hazard ratio; CI, confidence interval; HER-2, human epidermal growth factor receptor 2.

Internal validation of risk scores for predicting visceral metastasis-free survival in randomly sampled patients by Cox regression analysis.

200 cases 250 cases 300 cases
Subset no. P-value HR P-value HR P-value HR
1 0.002 1.836 <0.001 1.861 <0.001 2.113
2 <0.001 1.759 <0.001 2.388 <0.001 2.239
3 <0.001 2.025 <0.001 1.834 <0.001 1.942
4 <0.001 2.157 <0.001 1.812 <0.001 1.846
5 <0.001 2.041 0.002 1.666 <0.001 2.109
6 0.005 1.692 <0.001 1.915 <0.001 1.942
7 <0.001 1.860 <0.001 2.371 <0.001 1.968
8 <0.001 2.003 <0.001 2.146 <0.001 1.990
9 <0.001 1.902 0.001 1.656 <0.001 1.823
10 0.001 1.764 <0.001 2.177 <0.001 1.786

HR, hazard ratio.

Internal validation of risk scores for predicting disease-free survival in randomly sampled patients by Cox regression analysis.

200 cases 250 cases 300 cases
Subset no. P-value HR P-value HR P-value HR
  1 <0.001 2.228 <0.001 2.141 <0.001 2.482
  2 <0.001 2.296 <0.001 2.646 <0.001 2.489
  3 <0.001 2.422 0.001 2.023 <0.001 2.585
  4 <0.001 2.352 0.001 2.059 <0.001 2.300
  5 <0.001 2.799 <0.001 2.131 <0.001 2.879
  6 0.001 2.354 <0.001 2.244 <0.001 2.601
  7 0.001 2.072 <0.001 2.473 <0.001 2.492
  8 0.003 1.910 <0.001 2.724 <0.001 2.493
  9 <0.001 2.212 0.001 1.992 <0.001 2.019
10 <0.001 2.646 <0.001 2.447 <0.001 2.204

HR, hazard ratio.

Internal validation of risk scores for predicting overall survival in randomly sampled patients by Cox regression analysis.

200 cases 250 cases 300 cases
Subset no. P-value HR P-value HR P-value HR
1 0.003 1.938 <0.001 2.208 <0.001 2.253
2 <0.001 2.442 <0.001 2.986 <0.001 2.141
3 <0.001 2.323 <0.001 2.071 <0.001 2.338
4 <0.001 2.652 <0.001 2.243 <0.001 2.052
5 <0.001 2.496 <0.001 2.007 <0.001 2.269
6 <0.001 2.090 <0.001 2.181 <0.001 2.369
7 <0.001 2.243 <0.001 2.396 <0.001 2.175
8 <0.001 2.273 <0.001 2.516 <0.001 2.353
9 0.001 2.106 <0.001 2.365 <0.001 2.03
10 <0.001 2.241 <0.001 2.249 <0.001 2.039

HR, hazard ratio.