Combination of Grading and New Biological Factors (S-Phase Fraction and Epidermal Growth Factor Receptor) Can Predict Relapse and Survival in Patients with Node-Negative Primary Breast Cancer

Abstract

Background: In most cases of node-negative breast cancer, clinical outcome is relatively good after surgical treatment. It is necessary, however, to determine prognostic factors for the identification of high- and low-risk subgroups concerning recurrence and death, because only patients with excellent prognosis require no further treatment. The objective of the study was to select a few potential factors for a prognostic index that can be used in clinical practice. Patients and Methods: In 108 patients with primary node-negative breast cancer we examined simultaneously the impact of new prognosticators on disease-free survival (DFS) and overall survival (OAS): DNA ploidy, S-phase fraction (SPF), cycling index encoded by Ki 67, estrogen and progesterone receptor status, epidermal growth factor receptor (EGF-R), Her 2b/neu oncoprotein, and GP170 glycoprotein. Only fresh-frozen tissue was used. Median time of follow-up was 42 months. Results: By means of the log-rank analysis we found that patients with a SPF > 5% of the tumor had a significantly shorter DFS (p = 0.01) and OAS (p = 0.01) than patients with a SPF < 5%. Similar results for OAS were obtained for tumors with EGF-R levels > 15 fmol/mg versus < 15 fmol/mg (p = 0.05) and tumors of grade III versus grade I/II (p = 0.03). A low progesterone receptor level ( < 20 fmol/mg) indicated a short DFS (p = 0.06) but was of no prognostic value for OAS. Using a combination of SPF and EGF-R, a group of patients with extremely good prognosis (no patients have died) could be identified, if at least one factor was favorably expressed. If both factors were elevated clinical outcome was poor (DFS: p = 0.003; OAS: p = 0.002). By including histological grading in the analysis, a prognostic index could be described when discriminated patients with good, medium, and high risk for relapse and survival (DFS: p = 0.009; OAS: p = 0.001). Conclusions: Out of 11 parameters, grading, SPF, and EGF-R have been selected for a prognostic index. This prognostic index can be used to classify patients with node-negative primary tumors in different risk categories, so that an individual and risk-adapted adjuvant systemic therapy becomes possible, according to the treatment recommendations outside clinical trials.

Cost-effectiveness of using prognostic information to select women with breast cancer for adjuvant systemic therapy

“…As shown in Figure 3, many studies had fewer than 10 events per candidate variable. An extreme case is the study of Kaufmann and colleagues, 186 who investigated 11 variables in a data set of 57 patients with only four events (EPV <0.5).…”

“…That such omissions could be of major significance is illustrated by the study of Kaufmann and colleagues. 186 They reported the relevant numbers so we know that they began with 108 patients of whom 11 had died, but their prognostic model (derived from 11 variables) was based on only 57 patients and just four deaths. A few studies excluded patients lost to follow-up early in order to have a fixed-length follow-up (e.g.…”

“…Many studies had fewer than 10 events per candidate variable. An extreme case is the study of Kaufmann and colleagues, 186 who investigated 11 variables in a data set of 57 patients with only four events. Table 10 presents the characteristics of 54 prognostic models developed for breast cancer.…”

Sonographic blood flow measurements in malignant breast tumors

“…The correlation of a low blood flow with the ploidy, S-phase fraction, and receptor status-and thus a favorable prognosis-is also evident: a low S-phase fraction was mostly evident in the blood flow class I, whereas tumors with an aneuploid genome mostly had a lot of color pixels or areas. The detection of estrogen and progesterone receptors also correlated with less blood flow within the tumor, a lack of receptor correlating with a good blood supply [7,15,16]. These results are not so clear as shown in our first studies on this field, but show a correlation between color pixels and prognostic factors in malignant breast tumors [15,16].…”