Introduction 1 2 3 1 4 4 Fig. 1 Proportion of breast cancer patients (3-year moving average) diagnosed between 1973 and 1993 who survived 10 years or longer in Southeastern Netherlands  Methods We initially searched PubMed, using the search MESH term for ‘breast neoplasms’ AND ‘prognoses’ AND ‘long-term’. Only papers published in English between 1995 and 2006 (September) which researched female adults (19+ years) were included. We retrieved 528 articles and studied the abstracts (sometimes also the methods section). We selected only articles that assess or show the results for those surviving 10 years or longer with cohorts having a mean/median follow-up of 10 years or longer. If mean/median follow-up time was not reported, we examined the proportion of patients who survived 10 years after diagnosis, and this ought to be larger than 50%. If, for a specific topic of interest, no relevant studies with a follow-up of at least 10 years were found (such as BRCA mutation or gene profiling, which have been studied only during the last decade), then studies with the longest available follow-up were chosen. Furthermore, the following inclusion criteria were used: overall and/or BC-specific survival was reported; relative risk or hazard rate and statistical probability values were given; at least 250 BC patients included at the beginning of study. We also searched the reference lists collected by this search strategy and selected those that were relevant to both our study question and inclusion criteria. Reviews and books that gave general overviews were also included in the reference list. 5 Throughout the text the term long-term and/or survival will frequently be mentioned; this corresponds to at least 10-year survival unless otherwise indicated. Results and discussions Determinants of survival BC 10 years or longer Patient characteristics Age at diagnosis 6 7 2 8 9 6 10 11 6 6 12 13 Fig. 2 Relative survival of breast cancer patients (n: 13,279) diagnosed in 1990–2002 and followed until 2004, according to age at diagnosis in southeastern Netherlands  Comorbidity 13 12 13 13 12 14 Period of diagnosis 15 18 15 17 19 20 17 Time after diagnosis 16 21 16 16 3 5 23 Fig. 3 22 Dashed line solid line  Socioeconomic status (SES) and race 24 25 26 27 25 28 14 29 Tumour-related characteristics Tumour size 4 7 30 8 31 32 33 Fig. 4 x  Histological type 34 35 36 37 9 38 34 36 39 40 Histological grade 41 42 7 8 31 43 45 46 47 37 48 Regional lymph node involvement 5 8 32 8 9 49 32 7 31 50 Fig. 5 x  Lymphovascular invasion (LVI) and molecular markers of tumours angiogenesis 51 52 53 46 50 44 54 55 56 Grouped prognostic factors 41 16 57 Fig. 6 Cumulative survival of breast cancer patients diagnosed in southern Netherlands in 1970–1994 and followed-up until 2004, according to second cancer. Follow-up for patients with second cancer begins at the date of second cancer diagnosis. ■ no second cancer (n: 8137) • second breast cancer (n: 744) ▴ second non-breast cancer (n: 819)  11 46 49 11 Recurrence, metastasis and second cancer 9 21 58 61 21 62 21 63 59 9 64 6 65 59 Other tumour markers Hormone receptors 66 30 43 37 67 HER-2 expression 17 68 17 17 69 70 17 37 69 68 Mitotic Activity Index (MAI) 48 71 Gene expression profile 72 73 74 Various molecular markers 75 P 76 77 10 69 78 79 69 80 81 Miscellaneous Lifestyle 82 85 83 84 82 84 83 84 86 87 85 88 89 Modification of BC’s prognostic factors 90 93 92 93 91 Changing importance of prognostic factors over time after diagnosis 31 94 31 94 95 31 95 94 8 8 67 71 96 The role of early detection 19 97 98 99 100 100 The role of treatment 101 16 102 51 66 103 105 106 107 108 109 110 Conclusion 16 111 111 1 16 2 Table 1 Overview of studies reporting long-term prognostic factors for breast cancer (BC) patients No. Author, year No. of patients Mean/median follow-up (yrs) Univariate (UV) analysis significant Multivariate (UV) analysis significant Not significant Remarks 1 70 a 490 10.6    Her2 Overall survival was measured. P53 was related to absence of tubular formation, high G, ER-negative, high PCNA (proliferating cell nuclear antigen) score 2 10 392 11.1 P53, N, T, Htyp, tubular formation, intraductal growth, margin formation, necrosis, DNA ploidy, S-phase fraction    Overall survival was measured. P53 is related to younger age, MI, AI, G, nuclear pleomorphism 3 79 a 490 11 PCNA T, G, PR Her2 & PCNA (Proliferating cell nuclear antigen) Overall survival was measured. PR was only an independent factor in N-positive pts. Her2 & PCNA were related to more positive N, higher G, ER-/PR-negative 4 39 163,808 NR. Range 1mos-19 yrs Histological type by stage (localized & regional BC)   Breast cancer specific survival was measured 5 12 1196 NR. Diagnosed: 1973–1986. End FU: 1994. Comorbidity Level of comorbidity Adjusted for age, race, stage, N, therapy. Values for these factors were not shown Overall survival was measured. Charlson comorbidity index was used. There is no difference in the significance of comorbidity on survival of Caucasians and African American (AA) 6 112 a 487 >10 Ki-67, PCNA T, G, PR Ki-67 & PCNA. Overall survival was measured. PR only an independent factor for N+ patients. Ki-67 was related to T, MI, G 7 36 Mucinous BC: 4082. Infiltrating duct BC: 139,154 NR. Diagnosed: 1973–1990  HTyp, period of diagnosis, Stage, G Age Breast cancer specific survival was measured 8 11 a 2879 age ≤70, T <5cm >10 Age <35, NPI T, G, N Age Overall survival was measured. Younger than 35 yrs had higher grade, more LVI and worse NPI group. After 10yrs NPI did not change OS 9 95 8802 age <70 Diagnosed: 1965–74. End FU: 1991 Survival categorized by age, stage and follow-up time   Excess hazard from breast cancer was measured. After 8 yrs being younger than 35 does not influence survival. Stage was an important prognosticator up to 20 yrs 10 47 465 12 N by grade, Treatment by grade   Overall survival was measured. The study aimed to confirm value of Nottingham grading system for survival. N+G3 patients benefited from prolonged chemotherapy 11 113        BRCA1 vs. BRCA2 vs. control  12 28   At 10 yrs 59–67% patients were alive Race, G, N, T, stage, waiting time, smoking, being a widow, having other family as dependent Race, age, stage UV: alcohol, family history Overall survival was measured. AA is more likely to be younger at diagnosis, have larger tumour, higher stage and more lymph nodes 13 114 350 12.5 Bone marrow micrometastases N, T Bone marrow micrometastases, LVI Overall and breast cancer-specific survival was measured. Bone marrow metastases may be useful as prognostic factor for BC pts without information on T and N 14 46 a 319 T ≤1cm >10 G, N, LVI, NPI G, N LVI Overall survival was measured 15 45 2468 NR. Diagnosed: 1977–85. End FU: 1996 T, N, G, detection mode, HTyp TXN, age*N, Htyp*N, T*N*G  Overall survival was measured. Screening arrests disease progression. Tumour progression is more rapid in BC patients <50 yrs. OS of T1a(1–5 mm) vs. T1b(6–10 mm) NS. 16 85 1982 13.1 BMI ≥30 3rd  to 5th quintile of protein protein intake after diagnosis, N, T, G BMI, protein intake prior diagnosis, alcohol intake  17 58 1857 <80 yrs stage I-III 12 Second cancer and recurrence Age, ER, N, recurrence, second cancer   18 37 613 T1-2N0 15.5 Age >50, T, G G, T, treatment UV: treatment, ER, PR, Her2, P53 Overall survival was measured. Her2 was related to PR-, ER-negative, P53, G. P53 was related to PR-. Treatment was ovarian & locoregional irradiation that had lower mortality rate 19 7 3700 pre- & perimenopausal 12 Age <35 vs. ≥35 N, T, G, age <35*ER+ Age, ER Overall survival was measured. Younger patients with ER+ who were not amenorrhoea had a significantly shorter survival 20 78 297 N- 11 T, ER, P53 T, ER Age, PR, G Breast cancer-specific survival was measured. P53 was related to grade, T, ER-negative. P53 was continuous variable 21 6 10,356 age <50 NR. Diagnosed: 1978–96.  Age, T, N, G Period of treatment and surgery Relative survival was measured for excess mortality due to BC. When chemotherapy was given BC at young age does have worse prognosis 22 69 488 10 ER, uPA, G, N, PR, P53 by Her2    Overall survival was measured. For patients who received chemotherapy uPA, T & N determined OS. For patients who received hormonal therapy uPA, Her2 & N determined OS 23 44 a 377 10 T, N, G AMC, T, N, G Necrosis Overall survival was measured. AMC is a good prognostic factor for N- and T2-3 patients 24 96 791 16.3 T, N, G, ER, Her2, p53, MIB-1, MAI, AI   UV: age. MV All patients: AI, MI, MIB-1, ER, G MV in N- & N+: AI, MI, ER, G. Breast cancer-specific mortality was measured. When patient FU was truncated at 5 yrs, MI was prognostic factor for N+ and N- 25 115 311 no adjuvant therapy. 11.6 High risk group (ER- or T≥3cm) vs. low risk (ER+ and T <2 cm) T, risk group (high vs. low) G, MI Overall survival was measured. MI was only significant when FU was truncated at 5 yrs. Grade was significant prognostic factor for short- and long-term survival 26 48 270 T1N0M0 12.5 G, Tubular score, MI Tubular score and MI  Breast cancer-specific survival was measured. This study confirmed the use of Nottingham grading system in their cohort 27 49 402 ≥16 T, N, Htyp, G, LVI, NPI T, N, G  Overall survival was measured. NPI gives similar survival prognosis as T, N, G 28 26 23786 At 10 yrs about 50% patients were alive Age stratified by SES Intermediate vs. high SES group corrected for age, ER, N, T, stage Deprived vs. high SES group corrected for age, ER, N, T, stage Deprived women have more ER- tumours. ER distribution and treatment method accounted for 20% of disparities in survival 29 30 269 stage II 12 NR T, N, age, ER/PR MS, therapy Overall survival was measured 30 43 685 T≤3 cm 10.8 G, N, ER, necrosis N, necrosis, G UV: Vascular density, LVI, age, PR Overall survival was measured. Intratumoral vascular density was related to larger tumour size and higher grade 31 77    Stage, age, period of BC diagnosis, FU time (after 2 and 3 yrs of diagnosis) BRCA1, BRCA2 5-year relative survival was measured for excess mortality due to BC 32 35 9520 12  Tubular BC type vs. other type, by nodal status and chemotherapy  Overall survival was measured. Tubular BC type had better prognosis than other type. This type was more likely to have low G & ER+ 33 50 a 422 10 P53, MI, necrosis, T, N, LVI MI, T, N UV: AI  34 54 a 398 10 BVI, T, N, G, chemotherapy BVI, T, N, G, chemotherapy UV: necrosis Overall survival was measured 35 67 670 11.4 N, T, age, ER/PR N, T, age MS, ER/PR Breast cancer-specific survival was measured. After 5 yrs of FU ER and PR were not independent prognostic factors 36 17 1928 Diagnosed in 1968–69 and 1978–79 Her2, N, T, MS, lymphoid infiltration, PR- G, T, N, lymphoid infiltration  Overall survival was measured. HER-2 was related to large tumours, higher G, lymphoid infiltration, higher mitotic index, PR- 37 74 295 age <53, stage I-II 6.7 Gene profile (Good vs. bad prognosis) for all patients, N+, N− Gene profile, T, N, chemotherapy VI, G, age, hormonal therapy Overall survival was measured 38 72 117 age <55 NR    Better classification of patients with high risk of metastasis and in need of chemotherapy 39 21 3180 15.8  OS <5 year: N, G, recurrence or metastasis OS≥5 yr: G and recurrence or metastasis Age, T If patient remains without recurrence or metastasis, effect of prognostic factors decreases over time. With metastases, this effect increases 40 73 99 6.1 Gene profile (luminal 1–3 vs. basal 1–2 & Her2 type)   Luminal-like 1–3 was predominantly ER+. Basal-like 1–2 and Her2 was predominantly ER- 41 83 3385 N−, ER+ 13.8  BMI <18.5 and BMI ≥30 higher total mortality and other deaths. BMI on deaths after BC events. Total mortality, death after BC events and other deaths as well as recurrence rate and occurrence of a second cancer was were measured. MV was adjusted for treatment, age, MS, race, T, ER and PR. Reference group was BMI 18.5–24.9. 42 57 620 stage IIIB-M1 >20 Supraclavicular BC, Stage IIIB and M1   Overall and breast cancer specific survival were measured. Patients with supraclavicular metastases had significantly better survival than patients with M1. Survival of these patients resembles that of BC stage IIB. (FU for living patients 20 yrs, for all patients 4.5 yrs) 43 32 905 N+ Chemotherapy+ 22.6 N+ (N1–3 vs. N4–9 vs. N >10), also by treatment and follow-up time N, T, MS MV: NXT, MSXT, additional vincristine and prednison Overall survival was measured. N was related to T. MS was related to receptor status 44 16 54,228 At 10 yrs 65% patients were alive Period of diagnosis, stage by age, FU time by stage    45 56 905 aged 25–81 11.7 In N- : CD105+ vessels. In all pts: CD31, Tie-2/Tek  In all pts: T, Htyp, CD31, PR, age. In N−: T, CD31 vessels, ER, age Overall survival was measured. MV: Tie-2/Tek showed significant role for predicting OS in all patients and N- patients 46 15 18,578 age <50 NR. Diagnosed: 1953–1999. Period of diagnosis, stage, stage*period, time after diagnosis   Improvement of prognosis for BC patients younger than 50 over the past decades. Relative survival remains lowered even 40 yrs after diagnosis 47 76 584 Ashkenazi Jewish 116 BRCA1, T, N, ER, age, chemotherapy BRCA1, T, N, Age Tamoksifen, BRCA2 Breast cancer-specific survival was measured. No effect of BRCA on non-BC death. BRCA1 only predicted BC death in patients without chemotherapy 48 33 1187 LVI-, N-, Adjuvant systemic therapy- 10.4 T, G TXG  Overall and breast cancer specific survival were measured. Patients with higher grade and size have greater chance to die from other & those with low risk disease greater chance of death from BC 49 18 15,416 NR. Diagnosed 1946–2001. Period of diagnosis   Over the decades, there were less extensive surgery and lymph node examination, less radiotherapy, more chemo- and hormonal therapy 50 116 527 age ≥40 4.7 Comorbidity, N, Therapy, age≥70, comorbidity*N In age <70: comorbidity, N In age ≥70: comorbidity, age   51 53 374 22.4 LVI, G, N, Therapy LVI, G, N LMVD (Lymphatic Microvessel Density), T, Htyp, ER, age, MS Overall survival was measured. LVI is related to young premenopausal BC, lower G, N+ 52 31 2299 >10 G, N, T, Metastases G, N, T, Metastases  Breast cancer specific survival was measured. All studied factors predicted long-term survival, but their value decreased over time 53 82 6792 14 BMI 25–29, BMI ≥30 lower overall survival BMI ≥30 BMI 25–29 Overall survival and also disease free survival were measured. Reference group was BMI ≤ 24.9. MV adjusted for ER, T, N, MS, treatment, chemotherapy, hormonal- in combination with chemotherapy 54 84 4077 N-, ER- NR  BMI ≥35 and AA had higher overall mortality and non- BC death BMI and race on death after BC events  55 87 2987 96 months  Physical activity after diagnosis (MET ≥9) on BC and total mortality  Breast cancer and total mortality were measured. MV corrected for age, interval between diagnosis and physical activity assessment, smoking, BMI, MS, hormone therapy use, age at first birth, parity, energy intake, stage and treatment. MET: metabolic equivalent task hours per week. Patients with BMI ≥25 and more physical activity before diagnosis there was a significant trend for less breast cancer death 56 27 5042 NR. Diagnosed: 1964–92. End FU: 1992 NR Location of home, age at first child, physical activity at work MV corrected for: age, period of diagnosis, birth cohort, educational level Breast cancer-specific survival was measured. Incidence of BC increases with higher educational level, and case fatality decreases by increasing education level 57 38 1490 received breast- conserving treatment 13.9   Medullary BC vs. other BC type Overall survival was measured. Medullary BC type had better prognosis than other type. This type was more likely to have ER+, PR+ & less BRCA1/2 mutation. Medullary type was only a prognostic factor for the first 5 yrs 58 63 2102 stage I-II, 314 with local recurrence (LR) 13.1 NR No LR treatment, Invasive LR, time (yrs) to local recurrence, age at initial BC diagnosis T, detection method, number of nodes sampled, ER/PR, histological type, G, LVI, margins  59 62 266 BC with LR 11.2 after LR for living pts NR Location of LR, size of LR, skin involvement of LR, N+ for primary tumour  Overall survival was measured. Early detection of local recurrence may improve the treatment outcome 60 13 8966 Diagnosed 1995–2001. End FU: 2004 2 or more comorbidities, diabetes mellitus and previous cancer Previous cancer, CVD, DM, cerebrovascular disease, dementia, 2 or more comorbidities, stage, treatment (RT, ST, age)   61 14 906 10 Number of severe comorbidities, race, type of comorbidity All patients: 3 or more comorbidities adjusted for stage, age, ER, surgery, chemotherapy, radiotherapy  Overall survival was measured. AA had more diabetes and hypertension. After adjustment for these 2 comorbidities disparity disappeared 62 55 909/918 age: 25–81 11.3 Tie2 – UV: VEGFR-2, VEGFR-2 Overall survival was measured. VEGFR-1 and Tie2 were reported as independent prognostic factors corrected for T, G, Htyp, in all patients and N- 63 40 6184 Inflammatory BC NR. Diagnosed 1973–1995. End FU: 2000 Period of diagnosis   Breast cancer-specific survival was measured. Prognosis has improved over the decades due to more aggressive therapy 64 71 492 T1–2 N0 >10 yrs MAI  OS: HTyp, therapy, period of diagnosis. BCS: therapy, period of diagnosis, age, T, Htyp  65 8 2410 T ≤7 cm N1–2 19 T, skin fixation, muscle fixation, G, N, age     Overall survival was measured. Long-term effect of prognostic factors vanishing 66 29 90,124. White American: 76,111. AA: 14,013   Age, stage, SES  Meta-analysis. African American is an independent predictor of poor outcome for overall survival and breast cancer specific mortality 67 24 407,435 women followed for BC death (N:1408) Women who died of BC in 1968–96 Level of education by period of diagnosis   Breast cancer death among women with the highest education compared to women with the lowest education in 1968–74 was 0.43; and in 1990–96: 1.17 (NS) 68 25 3920 age <70 NR. Diagnosed in 1980–2000 SES SES corrected for age, period of diagnosis, marital status, country of birth, Htyp, ER, detection method, stage, sector of care, therapy  Overall survival was measured. Lowest SES had less frequently screen-detected cancers, less stage I, less lobular BC, less BCT, less lymph node dissection 69 117 992, age ≥70 6.9 Being in wheelchair, renal insufficiency, dementia, CHF, cardiac arrhythmia, DM, IHD, osteoporosis, PVD, cerebrovascular disease, COPD, Parkinson’s disease, valvular heart disease  Systemic therapy Overall survival was measured. CHF: Cardiac Heart Failure. DM: Diabetes Mellitus. IHD: Ischemic Heart Disease. PVD: Peripheral Heart Disease. COPD: Chronic Obstructive Pulmonary Disease. Role of comorbidity varies by age 70 68 639 premenopausal N+ 10 Her2 amplification Her2 corrected for age, N, ER, type of surgery  Overall survival was measured. Those with amplified Her2 have improved survival with CEF 71 52 (A) Adjuvant therapy − : 990. (B) Adjuvant treatment +: 1765  LVI  B: ER, age, Htyp  a Table 2 Selected prognostic factors for long-term overall mortality of breast cancer (BC) patients Patient groups based Hazard ratio (HR) for overall follow-up or survival probability (S) 10 years after diagnosis Morphology based Hazard ratio (HR) for overall follow-up or survival probability (S) 10 years after diagnosis Molecular based Hazard ratio (HR) for overall follow-up or survival probability (S) 10 years after diagnosis 8 HR: 31 HR: 69 HR:        <35 vs. 35–44 P   N≥1 vs. N0 2.4 (1.9–2.9)   >500 vs. ≤500 1.82 (1.1–2.9)       45–54 vs. 35–44 1.1 (ns) Metastases vs. N0 22.73 (16.1–32.2)  Only in node-positive        55–64 vs. 35–44 P           65–75 vs. 35–44 P     16 b 31 HR: Cell proliferation index HR:       1972–1976 59%   T10–14 vs. T1–9 1.2 (0.8–1.9) (MAI)        1977–1986 64%   T15–19 vs. T1–9 1.7 (1.1–2.6) 10 1.02 (1.00–1.03)       1987–1991 70%   T20–29 vs. T1–9 2.5 (1.6–3.9)  Only in node-positive patients   T30–49 vs. T1–9 3.8 (2.4–6.0)     T≥50 vs. T1–9 4.6 (2.9–7.6)   16 Relative survival: 44 HR: 74 S:     0 vs. 5 yrs after diagnosis    II vs. I 2.5 (1.0–6.1)  55% vs. 95%        Regional BC 79% vs. 84%   III vs. I 5.7(2.6–12.4) f         Locally advanced BC 53% vs. 68%     26 HR: 34 S: 30 HR:        Intermediate vs. affluent 1.2 (1.0–1.4) d,e <50% vs. >80%   Positive vs. negative 0.38 (0.02–1.06)        Deprived vs. affluent 1.2 (0.99–1.53)     Lifestyle HR:            Body Mass Index      2 85 c            Physical activity      a 87     a b c d 8 e f Healthier lifestyle generally increases long-term survival. Modifiable risk factors (such as alcohol consumption and obesity) not only affect incidence but also tumour’ clinical behaviour and thus survival. Although a lot is known about the prognosis for BC patients, effect of traditional prognostic factors appears to attenuate over time, leaving room for studies on the role of other and newer factors for long-term survival.