Multiparametric prognostic evaluation of biological factors in primary breast cancer.

BACKGROUND: An array of biological features related to tumor cell differentiation status, growth rate, and invasive potential have been identified as potential prognostic factors in breast cancer. We were interested in determining their relative importance in predicting patient survival. PURPOSE: We evaluated the relative weight of the following four biological factors in predicting survival of patients with breast cancer: tumor cell DNA content (determined by flow cytometry), tumor cell proliferation rate (determined by thymidine kinase activity), expression levels of cathepsin D and urokinase plasminogen activator, and several "classical" clinical and histological factors. METHODS: Selected from a prospectively updated database, the study population consisted of 319 primary breast cancer patients who received treatment and follow-up care (median, 6 years) in the Centre René Huguenin. To determine the profile of biological factors for each patient, we used frozen tumor specimens and (except for the flow cytometric DNA content assay) commercially available assay kits. We determined by Cox multivariate analysis the relationships of the biological factors to each other, to classical prognostic factors, and to disease-free and metastasis-free survival. RESULTS: In the overall population, disease-free survival was best predicted by node status (P = .004), clinical tumor size (P = .02), and cathepsin D expression (P = .01), whereas metastasis-free survival was best predicted by node status (P = .0004), clinical tumor size (P = .009), and urokinase plasminogen activator expression (P = .04). In node-negative patients, thymidine kinase activity was the only factor selected for disease-free (P = .04) and metastasis-free (P = .05) survival. In node-positive patients, the number of positive axillary lymph nodes was the only factor selected for disease-free (P = .0008) and metastasis-free (P = .00017) survival. CONCLUSIONS: Our retrospective analysis has identified protease expression and tumor cell proliferation rate as important biological prognostic factors in breast cancer. Prospective clinical trials should be undertaken to confirm these results.

Polymerase chain reaction analysis of parathyroid hormone-related protein gene expression in breast cancer patients and occurrence of bone metastases.

Parathyroid hormone-related protein (PTHrP) is associated with the syndrome of humoral hypercalcemia of malignancy. A high incidence of positive staining for PTHrP is observed in breast cancer and positivity is more frequent in patients who develop bone metastases. We assessed the presence of PTHrP mRNA by using the polymerase chain reaction in 38 normocalcemic breast cancer patients with long-term follow-up (minimum, 5 years) selected for the presence or absence of later bone metastasis development. In all the patients except one, the PTHrP gene was expressed in the breast tumor. The level of amplified PTHrP complementary DNA was inversely related to age (P < 0.02) and positively related to the proportion of invaded nodes (P < 0.02) but was not related to the other usual prognostic factors. The level of PTHrP mRNA was not different between the group of patients without recurrence or metastases (n = 11) and the group of patients who later developed metastases in soft tissues (n = 10). By contrast, patients who subsequently developed bone metastases (n = 17) showed higher PTHrP gene expression than patients in the other two groups (P < 0.001). This study suggests that strong PTHrP gene expression in breast tumors is associated with the onset of bone metastases.

Technical evaluation of thymidine kinase assay in cytosols from breast cancers. EORTC Receptor Study Group Report.

Pilot retrospective studies have pointed to the prognostic value of thymidine kinase (TK) in breast cancer. We studied the Prolifigen TK-REA assay (Sangtec Medical, Sweden), usually applied to serum, for TK analysis in breast cancer cytosols. Reproducibility was good, provided that small volume pipetting was performed carefully. The TK assay was not influenced by the short-term storage of cytosols in liquid nitrogen or at -80 degrees C. However, some steps appeared critical for good laboratory practice. The TK level was affected by thawing the cytosols more than twice. Tumour storage in liquid nitrogen should be preferred over storage at -80 degrees C. The components of the homogenisation buffer, especially sodium molybdate and KCl can have a marked influence on results. Finally, linearity problems arose with some cytosols. Thus, although assay of TK in cytosols is apparently simple, care must be taken in practice. The TK-REA kit should be standardised before widespread use in breast cancer.

Influence of the extraction procedure on plasminogen activator inhibitor-2 (PAI-2) and urokinase receptor (uPAR) assays in breast cancer tissues.

The levels of uPA, its inhibitors PAI-1 and PAI-2, and the uPA receptor (uPAR) have prognostic value in breast cancer. However, different extraction methods and assays kits are used in different laboratories and may directly influence the levels observed. To define a buffer suitable for both PAI-2 and uPAR extraction from breast cancer tissue compatible with hormone receptors and other cytosolic prognosticator assays, we compared PAI-2 and uPAR values obtained by immunoenzymatic assays (American Diagnostica, Greenwhich, USA) in several extraction conditions: 1) cytosol obtained with the standard hormone receptor buffer; 2) solubilized pellets obtained by Triton X100 extraction of the pelleted membranes obtained with standard hormone receptor buffer; 3) cytosol obtained by direct extraction in the buffer (containing Triton X100) recommended by the manufacturer, after 2 hours or 12 hours of incubation. Cytosol extracts prepared using the standard procedure recommended for hormone receptors gave the highest PAI-2 values. The highest uPAR values were obtained in the subsequent detergent extraction of the pelleted membranes. PAI-2 levels obtained with the kit manufacturer's method after 12 hours of incubation were lower than those obtained after 2 hours of incubation, whereas uPAR levels were similar. We conclude that the most suitable extraction protocol employs standard hormone receptor extraction buffer to obtain a supernatant cytosol fraction for PAI-2 assay, and subsequent detergent extraction of the pelleted membranes to obtain an extract suitable for uPAR.

Comparison of a new microplate oestrogen receptor (ER) enzyme immunoassay with other ER detection methods.

In a study involving 50 breast cancer tumours, we compared two oestrogen receptor (ER) detection methods developed by us--a microplate immunoenzymometric assay (EIA96) and an immunohistochemistry kit (HistoCIS-ER)--with the radioligand assay (RLA), the Abbott immunoenzymometric assay ER-EIA and the reverse transcriptase polymerase chain reaction technique (RT-PCR). Among the three ER protein cytosolic assays (EIA96, ER-EIA and RLA), the two EIAs showed the best agreement (y = 1.086x - 7.840; r2 = 0.876). At the calculated optimal cut-off values (8 and 14 fmol mg(-1) protein for EIA96 and RLA respectively), EIA96 was more sensitive than RLA (0.94 for EIA96, 0.88 for RLA), but slightly less specific (0.82 for EIA96, 0.94 for RLA). The Cox logistical regression model applied to EIA96, RLA and RT-PCR showed that EIA96 discriminated the best between ER-EIA+ and ER-EIA- samples. The RT-PCR technique and HistoCIS-ER both had a positivity-negativity concordance of 86% with EIA96.