Purpose: There is limited knowledge of risk factors for breast cancer recurrence within 2 years. This study aimed to predict early failure and identify high-risk patients for prognostic and therapeutic purposes.

Experimental Design: We studied 739 patients from a randomized trial who were <56 years of age and had ≥4 or more positive lymph nodes, no distant metastases, and no previous other malignancies. After complete surgical treatment, patients received conventional-dose anthracycline-based chemotherapy or a high-dose scheme of anthracycline-based plus alkylating chemotherapy. We assessed clinical and (immuno)histological parameters to predict recurrence within 2 years.

Results: Early failure occurred in 19% (n = 137). Median survival after early failure was limited to 0.7 year. Estrogen and progesterone receptor negativity and visceral relapse predicted poor prognosis. Early failure was associated with young age, large tumors, high histological grade, angio-invasion, apical node metastasis, and ≥10 involved nodes. Estrogen receptor, progesterone receptor, and p27 negativity; HER2 overexpression; and p53 positivity also predicted early failure. The surgical or chemotherapy regimen and histological type did not. The same parameters except tumor size were associated with early death. Grade III, ≥10 involved nodes, and estrogen receptor negativity were independently associated with early failure and together identified a subset of patients (7%) with 3-fold increased early failure and 5-fold increased early death.

Conclusions: Early failure is associated with poor survival. The combination of three commonly determined parameters constitutes a strong predictive model for early failure and death.

Axillary lymph node status is the strongest prognostic factor in patients with primary operable breast cancer. Compared with patients without axillary lymph node involvement, 10-year survival of lymph node-positive breast cancer patients is decreased from ∼65–80% to ∼25–48%. The prognosis worsens with increasing numbers of involved lymph nodes. Patients with four or more positive nodes are regarded as a particularly unfavorable subgroup. Further subdivision into categories (e.g., 4–10 versus ≥10 positive lymph nodes) can also be useful for prognostic purposes (1, 2).

Meta-analyses of randomized clinical trials have shown that adjuvant chemotherapy improves the prognosis of all breast cancer patients, irrespective of the risk of relapse. In absolute numbers, however, patients with lymph node metastases benefit most from adjuvant treatment (3).

Conventional-dose adjuvant chemotherapy improves patient outcome mainly by reducing the incidence of first local or regional and distant soft-tissue relapses, whereas first recurrences in bone or viscera are less influenced (4). Clinical trials in metastatic breast cancer, in which dose intensity of chemotherapy was the most important variable, have shown that the response rate to chemotherapy rises with increasing dose intensity (5, 6, 7, 8). More intensive treatments might also reduce the incidence of bone and visceral relapses and have a greater influence on survival. We recently published a multicenter randomized clinical trial investigating whether increasing the dosage of adjuvant chemotherapy improves the outcome of operable breast cancer patients with four or more positive lymph nodes (9). Within the framework of this trial, we evaluated the association between several tumor characteristics and short-term disease progression. Recurrence of breast cancer within 2 years of initial treatment predicted a dismal outcome (10, 11). Identifying the patient subgroup that will have such short-term progression can be clinically relevant. The main aim of this study was therefore to define factors that predict early failure in lymph node-positive breast cancer.

High-Dose versus Conventional-Dose Chemotherapy Trial.

We conducted a multicenter prospective randomized Phase III trial investigating the effectiveness of high-dose adjuvant chemotherapy in patients with operable breast cancer (9). Patients had undergone modified radical mastectomy or breast-conserving surgery with complete axillary clearance. Major eligibility criteria were histologically confirmed stage IIA, IIB, or IIIA carcinoma of the breast with involvement of four or more axillary lymph-nodes, age under 56 years, no evidence of distant metastases, and no previous other malignancies. The trial design consisted of two treatment arms to which patients were randomly assigned:

(a) The “conventional-dose” chemotherapy arm, comprising five courses of 5-fluorouracil (500 mg/m2), epirubicin (90 mg/m2), and cyclophosphamide (500 mg/m2); and

(b) The “high-dose” chemotherapy arm, comprising four courses of 5-fluorouracil-epirubicin-cyclophosphamide and a high-dose trialkylator regimen [cyclophosphamide (6 g/m2), thiotepa (480 mg/m2), and carboplatin (1.6 g/m2)]. During the third 5-fluorouracil-epirubicin-cyclophosphamide course, peripheral-blood stem cells were mobilized with granulocyte-colony-stimulating factor and harvested. During high-dose chemotherapy patients received granulocyte-colony-stimulating factor and peripheral blood stem cell infusion.

Within 6 weeks after chemotherapy, all patients underwent conventional external beam radiotherapy to the breast or chest wall and to the regional lymph node areas (axilla and internal mammary chain), according to local protocols. Patients with hormone receptor-positive tumors received 5 years of adjuvant tamoxifen.

Informed consent was obtained from all patients, and the study was approved by the institutional review committees at each of the participating centers.

Patients.

Between 1993 and 2000, 885 patients were entered in nine institutes. For 745 patients, H&E-stained and paraffin-embedded unstained tumor slides were available for this study. Six patients died of toxic complications and without evidence of a breast cancer recurrence. We analyzed the remaining 739 patients. Histological features were centrally revised by one of the authors (J. L. P.). The following tumor characteristics were recorded: largest diameter, WHO type, histological grade according to Elston and Ellis (12), presence of angio-invasion, total number of retrieved lymph nodes, number of lymph nodes with metastatic tumor cells, and status of apical node.

Immunohistochemistry.

Slides from formalin-fixed, paraffin-embedded tumors were stained with antibodies against HER2 [3B5 (13); 1:10,000 dilution], P53 (DO7; 1:8,000 dilution; mouse IgG2b; DAKO, Glostrup, Denmark), estrogen receptor (1D5; 1:150; mouse IgG1; DAKO), progesterone receptor (PR1; 1:3000 dilution; mouse IgG1; ImmunoVision, Brisbane, CA), p27 (Clone DCS-72.F6; 1:200 dilution; mouse IgG1; NeoMarkers, Fremont, CA). Staining was performed according to an earlier published protocol (14). Replacement of the primary antibody with 1% bovine albumin solution in PBS served as negative control. Archival breast cancer samples with known strong expression were used as positive controls.

Staining patterns recognized as positive were circumferential membrane-bound staining (HER2) and nuclear staining [estrogen receptor (ER), progesterone receptor (PR), p27, and p53]. HER2 was scored with the system that has recently come into use for clinical testing (0 = negative; 1+ ≥10% cells weakly positive and not circumferential; 2+ ≥10% of cells with moderate circumferential staining; 3+ ≥10% of cells with strong circumferential staining). Other immunohistochemical results were scored semiquantitatively on a three-point scale for percentage of positively staining tumor cells (negative = 0%; focal staining = up to 50%; homogeneous staining = 50–100%).

Definition of Early Failure and Statistical Analyses.

Early failure was defined as any recurrence of breast cancer (local, regional, or distant) <2 years from the time from randomization. Early death was defined as death within the same period. Randomization took place after surgery and definitive pathological diagnosis and before the start of chemotherapy.

The association of all parameters with early death and early failure was tested for statistical significance with the two-sided Pearson’s χ2 test in cross-tabulations. Multivariate analyses were made in a binary logistic regression model. Survival analysis was performed according to the Kaplan-Meier method and with the log-rank test for comparing groups. P < 0.05 was considered statistically significant. Additionally, we applied Bonferroni α-level adjustment for multiple testing in the single-variable tests. All data were analyzed with the Statistical Product and Service Solutions software (SPSS 10.0.7 for Windows; SPSS, Chicago, IL).

Pooled Data from Both Treatment Arms.

This trial was originally designed to assess the possible effects of high-dose chemotherapy on relapse and survival. As has recently been published, no statistically significant survival advantage was found for dose intensification per se(9). We therefore analyzed the data from both treatment arms pooled in one study population.

Patients.

A total of 739 patients were analyzed for this study. Mean patient age at randomization was 44.5 years (range, 25.6–55.6; median, 45.5 years). Mastectomy was performed in 572 patients (77%). All other patients underwent breast-conserving surgery with complete axillary dissection. Randomization assigned 369 patients to the conventional-dose chemotherapy treatment group and 370 patients to the experimental high-dose treatment arm.

Tumor Characteristics.

The mean tumor diameter was 3.6 cm (range 0.5–14.5 cm; median, 3.0 cm). Most tumors were invasive ductal carcinomas (61%), including seven cases of extensive in situ carcinoma with microinvasive foci. Forty-five percent of tumors were histological grade III, reflecting the unfavorable nature of the disease in the selected population.

Angio-invasion was present in 70% of cases, consistent with the lymph node-positive status of the patients. We investigated an average of 15.5 axillary lymph nodes/patient (range, 4–41; median, 15). The mean number of tumor-positive lymph nodes was 8.8 (range, 4–33; median, 8). The apical node was tumor positive in 67% of cases. This corresponds with stage IIIC disease according to the 6th edition of the American Joint Committee on Cancer staging system for breast cancer (15).

Follow-Up.

At the time of analysis, the mean duration of follow-up was 4.3 years (range, 0.6–9.1 years; median, 4.1 years). Early recurrence of breast cancer occurred in 137 patients (19%; Table 1). Early death occurred in 69 patients (9%). Early death did not occur as a first event.

Patients with early locoregional failure (n = 40) had recurrences in the remaining breast(s) and scar, chest wall skin (including lymphangitis), and contralateral and ipsilateral regional lymph nodes. Patients with early distant failure (n = 119) had metastases in various—often multiple—sites, mainly bone (n = 54), liver (n = 46), lung (n = 22), and brain (n = 16) as well as in cervical, retroperitoneal, mediastinal, and hilar lymph nodes; they also developed pleuritis, meningitis, and peritonitis.

Survival after Early Failure.

Median survival after early failure was 0.7 years (Table 2; Fig. 1, A and B), and 124 patients (91%) died during follow-up. Parallel to the study by Goldhirsch et al.(10), we compared three groups on the basis of first relapse site:

(i) Locoregional relapse: patients with only locoregional involvement;

(ii) Nonvisceral metastasis: patients with skeletal or extraregional skin or lymph node metastases but no organ involvement; and

(iii) Visceral metastasis: patients with any involved organ (e.g., brain, liver, or lung), bone marrow involvement, or pleural, peritoneal, or meningeal effusion.

The site of the recurrences was predictive of post-early failure survival (P = 0.04). The patients with only nonvisceral relapse (combination of groups i and ii) had significant survival advantage over those with visceral involvement (group iii; P = 0.02).

Immunohistochemical Markers.

ER staining was positive in 71% of tumors, and 58% had homogeneous staining. PR staining was positive in 57% and homogeneous in 27% of tumors. P53 staining was positive in 47% of tumors, and 20% had homogeneous staining. Seventy percent of tumors had P27 staining, and 13% had homogeneous staining. HER2 was positive (3+) in 23% of assessed tumors.

Associations with Early Failure and Early Death.

Both early failure and early death occurred significantly more often with young age, larger tumors, high histological grade, angio-invasion, with ≥10 axillary lymph node metastases, and with metastasis to the apical node (Table 3). Negative ER, PR, and p27 status; HER2 overexpression; and positive p53 status were also associated with early failure and with early death.

Of the parameters associated with early failure, only ER (P = 0.001) and PR (P = 0.02) were predictive of survival after early failure; negative hormone receptor status was associated with worse outcome (Table 2; Fig. 1, C and D).

Early failure and death were not associated with the randomly assigned chemotherapy regimen (high- versus conventional-dose), with the type of surgical treatment used (mastectomy versus breast conserving surgery), or with the primary tumor type (WHO classification).

Multivariate Analysis and Predictive Model.

The parameters that were predictive of early treatment failure and early death were entered in a binary logistic multivariate analysis, with early failure and early death as the dependent variables. An independent association with early failure was thus established for age <40 years (P = 0.03), histological grade III (P = 0.009), ≥10 positive lymph nodes (P = 0.002), and negative ER status (P = 0.004). Significant independent association with early death was established for histological grade III (P = 0.03) and ER status (P < 0.001).

Three factors—grade III, ER-negative tumors, and ≥10 positive lymph nodes—were used as a model for the identification of high-risk patients, which was tested for its predictive value (Table 4). Among 726 assessable cases with early failure and early death prevalences of 134 (18%) and 68 (9%), respectively, 52 patients qualified as high-risk (7%). The sensitivity of the model was 21% for early failure and 27% for early death. The positive predictive values were 54% for early failure and 35% for early death. The relative risks for high-risk versus low-risk patients were 6.3 for early failure and 6.6 for early death.

Breast cancer survival statistics show that a subset of patients develop distant metastases and die early. For example, the slope of the survival curve for premenopausal node-positive breast cancer patients distinctly decreases after 5 years of follow-up: mortality decreases from 30–40% in the first 5 years to 15–20% in the second 5 years (3). A disease-free interval <2 years is one of the most accurate predictors of a dismal outcome (10): the risk of breast cancer death is highest among women who have a recurrence within 2 years (11). A reliable identification of the women who are at high-risk of such early failure is of prognostic significance. It could be useful for selectively offering intensified follow-up controls or additional (experimental) forms of treatment.

We have analyzed 739 breast cancer patients under 56 years of age, with extensive lymph node involvement. They were entered in a randomized clinical trial of adjuvant chemotherapy with an anthracycline-based regimen or with an additional high-dose alkylating regimen. We compared patients who had disease progression within 2 years (early failure) versus those who did not.

On the basis of the post-early failure survival data we found that the prognosis for patients with early failure is poor, particularly for patients with early visceral metastases: not many more than one-third of the early failure patients survived longer than 3 years after initial treatment. This validates the use of parameters associated with early failure as prognostic factors.

Single-variable analyses of the relationships between the tested factors and early failure and early death were mostly consistent with previously reported data.

Large tumor size, poor histological grade, extensive axillary node involvement, apical node involvement, and ER and PR negativity are all commonly used in surgical treatment planning or in the choice of an adjuvant treatment strategy, because they are established prognostic factors associated with poor outcome. Many parameters are also interrelated: negative ER status is associated with negative PR status and with high histological grade, and angio-invasion is associated with lymph node metastases.

Young age—variably defined from study to study as younger than 25 to younger than 40 years of age—has previously been established as a negative prognostic factor, and histological differences are described from breast carcinoma in older patients, e.g., increased histological grade and angio-invasion (16, 17, 18, 19, 20). These associations were also found in our population. In accordance with earlier published studies, we further found HER2 overexpression in 23% of tumors, mostly of high histological grade; HER2 overexpression has been associated with increased risk of early failure and death (21, 22, 23, 24, 25). Increased p53 expression is generally associated with poor survival and high histological grade (26, 27, 28, 29). This was essentially confirmed in our series: p53 positivity was associated with high histological grade and with early failure. Also in line with the published literature [reviewed by Cariou et al.(30) and by Chiarle et al.(31)] is the inverse correlation between p27 expression and histological tumor grade and significantly more frequent early failures in focally staining and negative tumors (32). We have no explanation for the much lower proportion of tumors that stained homogeneously for p27 in our study population: 13% compared with ≥30% in other studies (33, 34).

In the multivariate analysis, most of the tested parameters were not independent prognostic factors. Probably this is the result of the strong association of these factors with histological tumor grade and ER status, which may have masked other independent effects.

The combined independent predictors histological grade III, negative ER status, and involvement of ≥10 axillary lymph nodes identified women with a 3-fold increased risk of early failure and 5-fold increased risk of early death. Such patients likely are an interesting group for prognostic reasons and are possible candidates for primary treatment with alternative regimens. With any single one of these factors, a more significant subgroup of patients with an increased risk of poor prognosis is identified, but none is by itself associated with even a 2-fold increased risk of early failure. These parameters are already routinely determined in most, if not all, breast cancer patients and are therefore applicable without further additional effort or cost, making the model practical for clinical purposes. Certainly the limited number of patients thus identified as high-risk does limit the clinical relevance of this predictive model. In this group 7% of the patients met the criteria for being considered “high risk.” Our analyses were done in premenopausal, lymph node-positive breast cancer patients. The model is likely to be applicable for lymph node-positive postmenopausal patients as well. Whether it is applicable to lymph node-negative patients is not known.

In conclusion, our findings in this large, well-defined set of lymph node-positive breast cancer patients confirm that early failure is a strong negative prognosticator. The presented data underline the strength of several known prognosticators and add their clinical use in a combination model that is easily applicable.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Requests for reprints: Marc. J. van de Vijver, Department of Pathology, The Netherlands Cancer Institute/Antoni van Leeuwenhoek Hospital, Plesmanlaan 121, 1066 CX Amsterdam, the Netherlands. Phone: 31 20 5122750; Fax: 31 20 5122758; E-mail: m.vd.vijver@nki.nl

Fig. 1.

Kaplan-Meier survival curves for 137 patients after early failure (percentage of surviving patients against time in years). A, all 137 patients; B, post-early failure survival by site of recurrence (P = 0.04). Solid gray line, patients with only nonvisceral metastases (n = 31); dashed black line, patients with only locoregional recurrence (n = 18); solid black line, patients with visceral metastases (n = 88). C, post-early failure survival by estrogen receptor (ER) status (P = 0.007). Solid gray line, patients with homogeneous ER staining (n = 42); dashed black line, patients with focal ER staining (n = 14); solid black line, patients with negative ER staining (n = 79). D, post-early failure survival by progesterone receptor (PR) status (P = 0.02). Solid gray line, patients with homogeneous PR staining (n = 13); dashed black line, patients with focal PR staining (n = 24); solid black line, patients with negative PR staining (n = 98).

Fig. 1.

Kaplan-Meier survival curves for 137 patients after early failure (percentage of surviving patients against time in years). A, all 137 patients; B, post-early failure survival by site of recurrence (P = 0.04). Solid gray line, patients with only nonvisceral metastases (n = 31); dashed black line, patients with only locoregional recurrence (n = 18); solid black line, patients with visceral metastases (n = 88). C, post-early failure survival by estrogen receptor (ER) status (P = 0.007). Solid gray line, patients with homogeneous ER staining (n = 42); dashed black line, patients with focal ER staining (n = 14); solid black line, patients with negative ER staining (n = 79). D, post-early failure survival by progesterone receptor (PR) status (P = 0.02). Solid gray line, patients with homogeneous PR staining (n = 13); dashed black line, patients with focal PR staining (n = 24); solid black line, patients with negative PR staining (n = 98).

Close modal
Table 1

Early failure in 739 patients

Type of early eventnPercentage of total (n = 739)
Any breast cancer relapse 137 19 
Locoregional recurrence 18 
Distant metastasis 97 13 
Locoregional and distant recurrence (simultaneous) 22 
Early death (<2 years) 69 
Type of early eventnPercentage of total (n = 739)
Any breast cancer relapse 137 19 
Locoregional recurrence 18 
Distant metastasis 97 13 
Locoregional and distant recurrence (simultaneous) 22 
Early death (<2 years) 69 
Table 2

Survival after early failure in 137 patients

Patients with EFa (n)Patients with EF who died during follow-up (n)Patients alive 1 year after EF (%)Median survival in years (95% confidence interval)P (log-rank)
All 137 124 36 0.7 (0.6–0.8)  
Site of first relapse     0.04 
 Nonvisceral only 31 27 53 1.0 (0.3–1.7)  
 Locoregional only 18 17 33 0.9 (0.6–1.1)  
 Visceral 88 80 30 0.6 (0.4–0.8)  
ER (n = 135)     0.001 
 Homogeneous 42 35 64 1.5 (1.1–1.8)  
 Focal 14 13 29 0.5 (0.2–0.8)  
 Negative 79 75 22 0.6 (0.5–0.8)  
PR (n = 135)     0.02 
 Homogeneous 13 10 54 1.0 (0.0–2.2)  
 Focal 24 22 54 1.0 (0.5–1.6)  
 Negative 98 91 28 0.6 (0.5–0.8)  
Patients with EFa (n)Patients with EF who died during follow-up (n)Patients alive 1 year after EF (%)Median survival in years (95% confidence interval)P (log-rank)
All 137 124 36 0.7 (0.6–0.8)  
Site of first relapse     0.04 
 Nonvisceral only 31 27 53 1.0 (0.3–1.7)  
 Locoregional only 18 17 33 0.9 (0.6–1.1)  
 Visceral 88 80 30 0.6 (0.4–0.8)  
ER (n = 135)     0.001 
 Homogeneous 42 35 64 1.5 (1.1–1.8)  
 Focal 14 13 29 0.5 (0.2–0.8)  
 Negative 79 75 22 0.6 (0.5–0.8)  
PR (n = 135)     0.02 
 Homogeneous 13 10 54 1.0 (0.0–2.2)  
 Focal 24 22 54 1.0 (0.5–1.6)  
 Negative 98 91 28 0.6 (0.5–0.8)  
a

EF, early failure; ER, estrogen receptor; PR, progesterone receptor.

Table 3

Statistical associations of patient and tumor parameters with early failure and early death

ParameterTotal (n)EFa+EF−P2)ED+ED− (n)P2)
n%n%
All  739 137 19 602  69 670  
Chemotherapy dose Conventional dose 369 78 21 291 0.07 39 11 330 0.25 
 High dose 370 59 16 311  30 340  
Surgery Mastectomy 572 109 19 463 0.50 55 10 517 0.63 
 Breast conserving 167 28 17 139  14 153  
Age <40 years 193 51 26 142 0.001b 25 13 168 0.045 
 ≥40 years 546 86 16 460  44 502  
Tumor classification (n = 690) pT1 (<2 cm) 113 17 15 96 0.014 105 0.11 
 pT2 (2–5 cm) 406 68 17 338  34 372  
 pT3 (≥5 cm) 171 45 26 126  23 14 148  
WHO tumor type (n = 739) IDC 452 86 19 366 0.64 41 411 0.98 
 ILC 126 21 17 105  12 10 114  
 Mixed invasive type 85 13 15 72  11 76  
 Other 76 17 22 59  69  
Histological grade (n = 724) Grade I 125 119 <0.001b 123 <0.001b 
 Grade II 273 37 14 236  16 257  
 Grade III 326 93 29 233  51 16 275  
Angio-invasion (n = 732) None 223 28 13 195 <0.001b 14 209 <0.001b 
 Focal 316 54 17 262  23 293  
 Extensive 193 55 29 138  332 17 161  
No. of lymph nodes with tumor (n = 711) 4–10 462 65 14 397 <0.001b 35 427 0.022 
 ≥10 249 66 27 183  32 13 217  
Apical node status (n = 688) No tumor 225 30 13 195 0.013 14 211 0.044 
 With metastasis 463 98 21 365  51 11 412  
ER (n = 718) Negative (no staining) 209 79 38 130 <0.001b 53 25 156 <0.001b 
 Focal (<50%) 94 14 15 80  87  
 Homogeneous (≥50%) 415 42 10 373  407  
PR (n = 719) Negative (no staining) 306 98 32 208 <0.001b 59 19 247 <0.001b 
 Focal (<50%) 221 24 11 197  216  
 Homogeneous (≥50%) 192 13 179  187  
HER2 (n = 721) Negative (0, 1+, 2+) 558 93 17 465 0.009 45 513 0.020 
 Positive (3+) 163 42 26 121  23 14 140  
p53 (n = 715) Negative (no staining) 382 62 16 320 0.002b 28 354 0.003b 
 Focal (<50%) 193 31 16 162  16 177  
 Homogeneous (≥50%) 140 41 29 99  24 17 116  
p27 (n = 624) Negative (no staining) 184 53 29 131 <0.001b 28 15 156 0.008 
 Focal (<50%) 358 62 17 296  29 329  
 Homogeneous (≥50%) 82 76  78  
ParameterTotal (n)EFa+EF−P2)ED+ED− (n)P2)
n%n%
All  739 137 19 602  69 670  
Chemotherapy dose Conventional dose 369 78 21 291 0.07 39 11 330 0.25 
 High dose 370 59 16 311  30 340  
Surgery Mastectomy 572 109 19 463 0.50 55 10 517 0.63 
 Breast conserving 167 28 17 139  14 153  
Age <40 years 193 51 26 142 0.001b 25 13 168 0.045 
 ≥40 years 546 86 16 460  44 502  
Tumor classification (n = 690) pT1 (<2 cm) 113 17 15 96 0.014 105 0.11 
 pT2 (2–5 cm) 406 68 17 338  34 372  
 pT3 (≥5 cm) 171 45 26 126  23 14 148  
WHO tumor type (n = 739) IDC 452 86 19 366 0.64 41 411 0.98 
 ILC 126 21 17 105  12 10 114  
 Mixed invasive type 85 13 15 72  11 76  
 Other 76 17 22 59  69  
Histological grade (n = 724) Grade I 125 119 <0.001b 123 <0.001b 
 Grade II 273 37 14 236  16 257  
 Grade III 326 93 29 233  51 16 275  
Angio-invasion (n = 732) None 223 28 13 195 <0.001b 14 209 <0.001b 
 Focal 316 54 17 262  23 293  
 Extensive 193 55 29 138  332 17 161  
No. of lymph nodes with tumor (n = 711) 4–10 462 65 14 397 <0.001b 35 427 0.022 
 ≥10 249 66 27 183  32 13 217  
Apical node status (n = 688) No tumor 225 30 13 195 0.013 14 211 0.044 
 With metastasis 463 98 21 365  51 11 412  
ER (n = 718) Negative (no staining) 209 79 38 130 <0.001b 53 25 156 <0.001b 
 Focal (<50%) 94 14 15 80  87  
 Homogeneous (≥50%) 415 42 10 373  407  
PR (n = 719) Negative (no staining) 306 98 32 208 <0.001b 59 19 247 <0.001b 
 Focal (<50%) 221 24 11 197  216  
 Homogeneous (≥50%) 192 13 179  187  
HER2 (n = 721) Negative (0, 1+, 2+) 558 93 17 465 0.009 45 513 0.020 
 Positive (3+) 163 42 26 121  23 14 140  
p53 (n = 715) Negative (no staining) 382 62 16 320 0.002b 28 354 0.003b 
 Focal (<50%) 193 31 16 162  16 177  
 Homogeneous (≥50%) 140 41 29 99  24 17 116  
p27 (n = 624) Negative (no staining) 184 53 29 131 <0.001b 28 15 156 0.008 
 Focal (<50%) 358 62 17 296  29 329  
 Homogeneous (≥50%) 82 76  78  
a

EF, early failure; ED, early death; IDC, invasive ductal carcinoma; ILC, invasive lobular carcinoma; ER, estrogen receptor; PR, progesterone receptor.

b

P < 0.004 (0.05/14), indicating that the parameter is statistically significant after Bonferroni α-level adjustment for multiple testing.

Table 4

Predictive model for early failure with three unfavorable tumor characteristics

TotalEFa+%EF−P2)ED+%ED−P2)
Total 726 134 19 592 <0.001 68 658 <0.001 
Unfavorable combination (ER−, grade III, and ≥10 positive lymph nodes) 52 28 54 24  18 35 34  
Other 674 106 16 568  50 624  
TotalEFa+%EF−P2)ED+%ED−P2)
Total 726 134 19 592 <0.001 68 658 <0.001 
Unfavorable combination (ER−, grade III, and ≥10 positive lymph nodes) 52 28 54 24  18 35 34  
Other 674 106 16 568  50 624  

Note. With versus without unfavorable characteristics: Odds ratio for EF = 6.3 (95% confidence interval: 3.5–11.2). Odds ratio for ED = 6.6 (95% confidence interval: 3.5–12.6).

a

EF, early failure; ED, early death.

1
Fisher B, Bauer M, Wickerham DL, et al Relation of number of positive axillary nodes to the prognosis of patients with primary breast cancer. An NSABP update.
Cancer (Phila)
,
52
:
1551
-7,  
1983
.
2
Rodenhuis S, Richel DJ, van der Wall E, et al Randomised trial of high-dose chemotherapy and haemopoietic progenitor-cell support in operable breast cancer with extensive axillary lymph-node involvement.
Lancet
,
352
:
515
-21,  
1998
.
3
Polychemotherapy for early breast cancer: an overview of the randomised trials. Early Breast Cancer Trialists’ Collaborative Group.
Lancet
,
352
:
930
-42,  
1998
.
4
Goldhirsch A, Gelber RD, Price KN, et al Effect of systemic adjuvant treatment on first sites of breast cancer relapse.
Lancet
,
343
:
377
-81,  
1994
.
5
Carmo-Pereira J, Costa FO, Henriques E, et al A comparison of two doses of adriamycin in the primary chemotherapy of disseminated breast carcinoma.
Br J Cancer
,
56
:
471
-3,  
1987
.
6
Tannock IF, Boyd NF, DeBoer G, et al A randomized trial of two dose levels of cyclophosphamide, methotrexate, and fluorouracil chemotherapy for patients with metastatic breast cancer.
J Clin Oncol
,
6
:
1377
-87,  
1988
.
7
Ebbs SR, Saunders JA, Graham H, A’Hern RP, Bates T, Baum M Advanced breast cancer. A randomised trial of epidoxorubicin at two different dosages and two administration systems.
Acta Oncol
,
28
:
887
-92,  
1989
.
8
Focan C, Andrien JM, Closon MT, et al Dose-response relationship of epirubicin-based first-line chemotherapy for advanced breast cancer: a prospective randomized trial.
J Clin Oncol
,
11
:
1253
-63,  
1993
.
9
Rodenhuis S, Bontenbal M, Beex LV, et al High-dose chemotherapy with hematopoietic stem-cell rescue for high-risk breast cancer.
N Engl J Med
,
349
:
7
-16,  
2003
.
10
Goldhirsch A, Gelber RD, Castiglione M Relapse of breast cancer after adjuvant treatment in premenopausal and perimenopausal women: patterns and prognoses.
J Clin Oncol
,
6
:
89
-97,  
1988
.
11
Fredriksson I, Liljegren G, Arnesson L, et al Local recurrence in the breast after conservative surgery-a study of prognosis and prognostic factors in 391 women.
Eur J Cancer
,
38
:
1860
-70,  
2002
.
12
Elston CW, Ellis IO 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
-10,  
1991
.
13
van de Vijver MJ, Mooi WJ, Wisman P, Peterse JL, Nusse R Immunohistochemical detection of the neu protein in tissue sections of human breast tumors with amplified neu DNA.
Oncogene
,
2
:
175
-8,  
1988
.
14
Faneyte IF, Schrama JG, Peterse JL, Remijnse PL, Rodenhuis S, van De Vijver MJ Breast cancer response to neoadjuvant chemotherapy: predictive markers and relation with outcome.
Br J Cancer
,
88
:
406
-12,  
2003
.
15
American Joint Committee on Cancer. .
Cancer staging manual
, 6th Ed. Springer-Verlag New York  
2002
.
16
Chung M, Chang HR, Bland KI, Wanebo HJ Younger women with breast carcinoma have a poorer prognosis than older women.
Cancer (Phila)
,
77
:
97
-103,  
1996
.
17
Kollias J, Elston CW, Ellis IO, Robertson JF, Blamey RW Early-onset breast cancer-histopathological and prognostic considerations.
Br J Cancer
,
75
:
1318
-23,  
1997
.
18
Bertheau P, Steinberg SM, Cowan K, Merino MJ Breast cancer in young women: clinicopathologic correlation.
Semin Diagn Pathol
,
16
:
248
-56,  
1999
.
19
Yildirim E, Dalgic T, Berberoglu U Prognostic significance of young age in breast cancer.
J Surg Oncol
,
74
:
267
-72,  
2000
.
20
Kothari AS, Beechey-Newman N, D’Arrigo C, et al Breast carcinoma in women age 25 years or less.
Cancer (Phila)
,
94
:
606
-14,  
2002
.
21
Slamon DJ, Clark GM, Wong SG, Levin WJ, Ullrich A, McGuire WL Human breast cancer: correlation of relapse and survival with amplification of the HER-2/neu oncogene.
Science (Wash DC)
,
235
:
177
-82,  
1987
.
22
Slamon DJ, Godolphin W, Jones LA, et al Studies of the HER-2/neu proto-oncogene in human breast and ovarian cancer.
Science (Wash DC)
,
244
:
707
-12,  
1989
.
23
Tsuda H, Hirohashi S, Shimosato Y, et al Correlation between histologic grade of malignancy and copy number of c-erbB-2 gene in breast carcinoma. A retrospective analysis of 176 cases.
Cancer (Phila)
,
65
:
1794
-800,  
1990
.
24
Seshadri R, Firgaira FA, Horsfall DJ, McCaul K, Setlur V, Kitchen P Clinical significance of HER-2/neu oncogene amplification in primary breast cancer. The South Australian Breast Cancer Study Group.
J Clin Oncol
,
11
:
1936
-42,  
1993
.
25
Hoff ER, Tubbs RR, Myles JL, Procop GW HER2/neu amplification in breast cancer: stratification by tumor type and grade.
Am J Clin Pathol
,
117
:
916
-21,  
2002
.
26
Allred DC, Clark GM, Elledge R, et al Association of p53 protein expression with tumor cell proliferation rate and clinical outcome in node-negative breast cancer.
J Natl Cancer Inst (Bethesda)
,
85
:
200
-6,  
1993
.
27
Thorlacius S, Thorgilsson B, Bjornsson J, et al TP53 mutations and abnormal p53 protein staining in breast carcinomas related to prognosis.
Eur J Cancer
,
31A
:
1856
-61,  
1995
.
28
Borresen AL, Andersen TI, Eyfjord JE, et al TP53 mutations and breast cancer prognosis: particularly poor survival rates for cases with mutations in the zinc-binding domains.
Genes Chromosomes Cancer
,
14
:
71
-5,  
1995
.
29
Aas T, Borresen AL, Geisler S, et al Specific P53 mutations are associated with de novo resistance to doxorubicin in breast cancer patients.
Nat Med
,
2
:
811
-4,  
1996
.
30
Cariou S, Catzavelos C, Slingerland JM Prognostic implications of expression of the cell cycle inhibitor protein p27Kip1.
Breast Cancer Res Treat
,
52
:
29
-41,  
1998
.
31
Chiarle R, Pagano M, Inghirami G The cyclin dependent kinase inhibitor p27 and its prognostic role in breast cancer.
Breast Cancer Res
,
3
:
91
-4,  
2001
.
32
Nohara T, Ryo T, Iwamoto S, Gon G, Tanigawa N Expression of cell-cycle regulator p27 is correlated to the prognosis and ER expression in breast carcinoma patients.
Oncology
,
60
:
94
-100,  
2001
.
33
Fredersdorf S, Burns J, Milne AM, et al High level expression of p27(kip1) and cyclin D1 in some human breast cancer cells: inverse correlation between the expression of p27(kip1) and degree of malignancy in human breast and colorectal cancers.
Proc Natl Acad Sci USA
,
94
:
6380
-5,  
1997
.
34
Sgambato A, Zhang YJ, Arber N, et al Deregulated expression of p27(Kip1) in human breast cancers.
Clin Cancer Res
,
3
:
1879
-87,  
1997
.