Prognostic value of serum CA 125 bi-exponential decrease during first line paclitaxel/platinum chemotherapy: a French multicentric study.

BACKGROUND: CA 125 assays enable treatment-response monitoring in ovarian cancer. METHODS: A multicentric study of CA 125 kinetics under paclitaxel/platinum-based chemotherapy was performed in 130 stage IIc-IV patients. CA 125 half-life and nadir concentration were compared to patient outcome. Some patients (n=38, 29.2%) presented a CA 125 bi-exponential decrease and its clinical implication was studied. Survival analyses for disease-free survival (DFS) and overall survival (OS) used univariate (Kaplan-Meier) and multivariate (Cox model). RESULTS: During a median follow-up time of 29 months (range 5-106 months), 111 patients (85%) relapsed and 94 (72%) died from ovarian cancer. Patients were split into 4 groups according to their pattern of CA 125 decrease: non-assessable half-life because of a low pre-chemotherapy CA 125 level (n=38), half-life < or = 14 days and mono-exponential CA 125 decay (n=18), half-life < or = 14 days and bi-exponential CA 125 decay (n=21), and half-life > 14 days (n=53). In Cox models, nadir concentration, residual tumour volume and number of chemotherapy courses were found to be independent prognostic factors for DFS and OS. The group classification was found to be an independent prognostic factor only for DFS. However, when nadir was not introduced in the models, the CA 125 kinetics groups were the most important prognostic factor for OS. CONCLUSION: Characteristics of CA 125 kinetics during first line paclitaxel/platinum chemotherapy have a strong and independent prognostic value. A CA 125 bi-exponential decrease is an indicator of bad prognosis.

CA 125 half-life and CA 125 nadir during induction chemotherapy are independent predictors of epithelial ovarian cancer outcome: results of a French multicentric study.

BACKGROUND: CA 125 assays enable treatment-response monitoring in ovarian cancer. PATIENTS AND METHODS: A multicentric study of CA 125 kinetics under induction chemotherapy was performed in 631 patients. CA 125 half-life was calculated by mono-compartmental logarithmic regression. Nadir CA 125 concentration and time to nadir were also studied. Survival analyses for disease-free survival (DFS) and overall survival (OS) used univariate (Kaplan-Meier) and multivariate (Cox) models. RESULTS: For 553 stage IIC-IV patients, 459 (83.0%) relapsed and 444 (80.3%) died from cancer. Median (range) follow up time was 32 months (2-214 months). Median (range) for CA 125 kinetics were: 263 kU/l (5-52000 kU/l) before 1st course, 15.8 days (4.5-417.9 days) for CA 125 half-life, 16 kU/l (3-2610 kU/l) for nadir and 85 days (0-361 days) for time to nadir. Pre-chemotherapy CA 125, its half-life, nadir concentration and time to nadir all had a univariate prognostic value for DFS and OS (P<0.0001). In Cox models, CA 125 half-life, residual tumour (P<0.0001 for both), nadir concentration (P=0.0002) and stage (P=0.0118) were the most powerful prognostic factors for DFS. For OS, the significant variables were similar, with age ranking last (P=0.0319). CONCLUSION: Among well-established prognostic factors in ovarian cancers, CA 125 half-life and nadir concentration bear a strong and independent prognostic value.

Drug-induced cardiotoxicity studied by longitudinal B-type natriuretic peptide assays and radionuclide ventriculography.

BACKGROUND: To study the longitudinal variations of plasma B-type natriuretic peptide (BNP) with reference to left ventricular ejection fraction (LVEF) during and after chemotherapy with cardiotoxic drugs. PATIENTS AND METHODS: We prospectively measured plasma BNP using an immunoradiometric assay in 12 anthracycline-treated breast cancer patients monitored for a mean time of 880+/-293 days (pilot group). Prior to each cycle and throughout the following year, LVEF and cardiac output were measured by radionuclide ventriculography. Anthracycline pharmacokinetics was studied during the first cycle. Relationships between serial observations were analysed with the general linear mixed effects model. Identical methods were subsequently applied to a test group of 67 anthracycline or trastuzumab-treated patients. RESULTS: Five out of 70 (6.33%) patients developed anthracycline-induced heart failure. BNP concentrations were found to be positively correlated to anthracycline cumulative dose and negatively to LVEF values. Variables entering the mixed models were cumulative anthracycline dose, time and cardiac output. CONCLUSION: An infra-clinical cardiotoxicity of anthracyclines as defined by BNP elevation is frequent but reversible. Patients who developed heart failure showed a continuous BNP increase and concentrations over 100 ng/ml.

Tissue expression and serum levels of the oncoprotein HER-2/neu in 157 primary breast tumours.

BACKGROUND: We studied HER-2 expression in paired serum and tissue samples, in 157 selected cases from 701 consecutive primary breast cancer patients with pre-treatment HER-2 extracellular domain (ECD) > or = 10 ng/ml, or < 10 ng/ml but showing a HER-2 ECD lead time before first metastasis. PATIENTS AND METHODS: HER-2 ECD was measured by the Immuno 1 automated ELISA (Bayer). Tumour tissue was analysed by immunohistochemistry (IHC) with Dako A 0485 and CB 11 antibodies and scored with the Dako scoring system. RESULTS: Mean HER-2 ECD was 12.48+/-7.08 ng/ml and 21/157 (13.4%) sera were > or = 15 ng/ml (cut-off). Forty tumours (25.48%) showed both invasive and intraductal components, 3 (1.91%) were pure in situ carcinomas and 114 (72.61%) were pure invasive tumours. Elevated HER-2 ECD concentration was related only to pT (p=0.0008), histological grade (p=0.0465), presence of comedonecrosis (p=0.0123) or comedo-type carcinoma (p=0.041) and was unrelated to the presence of an intraductal component. HER-2 ECD was > or = 15 ng/ml in 48% of Dako 3+ and 60% of CB 11 2+ and 3+ tumours. By logistic regression analysis, the significant parameters associated with HER-2 ECD concentration were pT (p=0.0038) and Dako 3+ scores (p=0.0005). In Dako 3+ or CB 11 2+3+ tumours, elevated mean HER-2 ECD concentrations were observed only when pT exceeded 28-30 mm (p=0.0062 and p=0.0036, respectively). CONCLUSION: In breast tumours, a threshold in size and HER-2 overexpression is necessary to observe elevated concentrations of HER-2 ECD at diagnosis. This information may be useful when the primary tumour is not available for IHC.

Management and clinical outcomes of pregnant patients with invasive cervical cancer.

BACKGROUND: The objective of this study was to evaluate the clinical outcomes and to discuss the management of women presenting with an invasive cervical cancer during pregnancy. PATIENTS AND METHODS: We retrospectively reviewed patients treated for an invasive cervical cancer diagnosed during pregnancy between 1985 and 2000 in our institution. RESULTS: Twenty-one pregnant patients among a total of 487 women were treated. Thirteen, five, two and one, respectively, were diagnosed during the first, second and third pregnancy trimester and post-partum. The FIGO stage was IB in 15 cases, IIB in five cases and IVA in one case. Mean follow-up was 64 months (range 2-165). Overall and disease-free survival at 5 years were 82% and 79%, respectively. All five patients diagnosed in the second trimester were alive. Two of the 13 patients and one of the two patients diagnosed during the first trimester and the third trimester, respectively, died of their disease. No difference was observed between the nine patients whose treatment was delayed or not. CONCLUSIONS: Invasive cervical cancer during pregnancy is rare but is a dilemma for women and their physicians. The present study and review of the literature suggest that pregnancy does not seem to influence the prognosis of cervical cancer. Delayed treatment could be proposed to selected patients diagnosed at the end of the second trimester or at the beginning of the third trimester, with a small tumor (<2 cm) and negative nodes, after a multidisciplinary approach.

Serum HER-2 extracellular domain: relationship with clinicobiological presentation and prognostic value before and after primary treatment in 701 breast cancer patients.

PURPOSE: To determine the clinical correlations and prognostic value of serum HER-2 (sHER-2) before and after primary breast cancer treatment. METHODS: sHER-2 from 701 consecutive patients with stage I-III tumors (median follow-up 7.7 years) was assayed by an enzyme-linked immunosorbent assay (Immuno 1, Bayer Diagnostics). RESULTS: The median pretreatment sHER-2 concentration was 8.30 ng/mL (range 3.15-82.00 ng/mL). Forty-seven patients (6.7%) had sHER-2 concentrations >12 ng/mL (cutoff level). Pretreatment sHER-2 correlated positively with CA 15.3 (p=0.0169), pathological tumor size (p=0.0082), number of invaded lymph nodes (pN, p=0.0160) and histological grading (p=0.0086). Kaplan-Meier analyses indicated that pretreatment sHER-2 was of prognostic value for contralateral breast cancer (p=0.0018), metastasis-free survival (MFS) (p=0.0008) - particularly lung (p=0.0082) and liver metastases (p=0.0035) - and overall disease-specific survival (DSS) (p=0.0020). According to pN status, pretreatment sHER-2 was of prognostic value only for pN-positive patients (p=0.0017). When combined with estradiol or progesterone receptor status, patients with elevated sHER-2 and receptor-negative tumors had a significantly shorter DSS (p<0.0001 for both receptors). Post-treatment sHER-2 also had individual prognostic value for MFS (p=0.0144) and DSS (p=0.0212). In multivariate analysis, only sHER-2 after primary treatment was an independent prognostic variable for MFS and DSS (p=0.0078 and p=0.0058, respectively). CONCLUSION: sHER-2 elevation in early breast cancer correlates with the principal criteria of tumor aggressiveness, thus permitting selection of patients with a high risk of visceral metastases and contralateral breast tumors. Post-treatment sHER-2 is an independent prognostic factor enabling to identify patients likely to benefit from aggressive adjuvant treatments.

Limited prognostic value of c-erbB-2 compared to uPA and PAI-1 in primary breast carcinoma.

In a retrospective study of 488 women with primary breast cancer, after a median follow-up of 10 years, we sought interactions between disease-free survival (DFS) and overall survival (OS) and tumor antigen levels of two components of the plasminogen system, urokinase-type plasminogen activator (uPA) and its inhibitor PAI-1, and the transmembrane growth factor receptor c-erbB-2. We used ELISAs (American Diagnostica, Greenwich, CT, USA) to quantify uPA and PAI-1 antigen levels in cytosols, and a double monoclonal antibody-based assay (EIA) (Ciba Corning Diagnostics, Alameda, CA, USA) to quantify c-erbB-2 in membrane extracts of the same tissues. Weak positive correlations were found between uPA and c-erbB-2 (r(s) = 0.146; p = 0.001) and between PAI-1 and c-erbB-2 (r(s) = 0.154; p < 0.001). In the overall population, using univariate analyses, c-erbB-2 overexpression and high uPA and PAI-1 antigen levels (> 300 IU/mg, > 1.40 ng/mg and > 5.53 ng/mg, respectively) were significantly associated with shorter DFS (p = 0.003, p < 0.001 and p < 0.001, respectively) and OS (p < 0.001 in all cases). Using multivariate analyses, PAI-1, node status and tumor size were independent predictors of DFS and c-erbB-2 was retained in the model only for OS. In the node-negative subgroup, PAI-1 was the strongest significant survival predictor both for OS (p = 0.003; HR 2.52) and DFS (p < 0.001; HR 2.39). This study shows that in primary breast cancer c-erbB-2 offers no additional prognostic information when uPA and/or PAI-1 are candidates in the multivariate analyses.

Prognostic value of CA 15.3 kinetics for metastatic breast cancer.

Up to 80% of breast cancer patients developing metastases have high levels of CA 15.3. We studied the prognostic implications of CA 15.3 kinetics in 119 patients before and at first metastasis by univariate and multivariate statistics. At first metastasis, CA 15.3 was elevated in 82.4% of patients, with a lead time (median 162 days) in 42.0% of them. Kaplan-Meier analysis showed overall survival (median 1477 days) to be significantly related to estrogen receptor (ER) and progesterone receptor (PgR) status (p=0.0001) and tumor size (p=0.025). The interval between diagnosis and first abnormal CA 15.3 (p=0.0001), the CA 15.3 concentration (p=0.013), and the presence or absence of a lead time (p=0.001) also had prognostic value. ER and PgR status (p=0.0005 and p=0.0103, respectively), metastasis-free interval (p=0.0003), existence of a CA 15.3 lead time (p=0.0028), and days from diagnosis to first abnormal CA 15.3 (p=0.0055) entered in the Cox model. After first metastasis (median survival 573 days), ER and PgR status (p=0.0001 and p=0.0004, respectively), existence of a lead time for CA 15.3 (p=0.0138), and the concentration of first abnormal CA 15.3 (p=0.0145) had individual prognostic value. In the Cox model ER status (p=0.0001), nodal status (p=0.0191), existence of a lead time for CA 15.3 (p=0.0033), days from diagnosis to first abnormal CA 15.3 (p=0.0132), and concentration of first abnormal CA 15.3 (p=0.0320) were found to be independent prognostic variables. Compared to a matched historical control group that was not monitored by CA 15.3 assaying (n=140), the study group had a significantly longer survival after the first metastasis (p=0.0005). In conclusion, the kinetics of CA 15.3 before the first metastasis is of prognostic value. When associated with 18-fluorodeoxyglucose imaging, serial CA 15.3 assays may help to implement early treatment of metastases.

ERBB2 status and benefit from adjuvant tamoxifen in ERalpha-positive postmenopausal breast carcinoma.

We examined the relation between ERBB2 gene expression (as determined by a real-time quantitative RT-PCR assay) and the response to adjuvant tamoxifen therapy in a well-defined cohort of 125 ERalpha-positive postmenopausal patients with breast cancer. Although ERBB2 overexpression was associated with shorter relapse-free survival in univariate analysis (P=0.00029), ERBB2 did not persist as an independent prognostic factor in multivariate analysis. Nevertheless, when we analyzed the ERBB2 mRNA level as a continuous variable, the higher the ERBB2RNA level, the poorer the outcome (P=0.00036). The results point to the need for a quantitative ERBB2 expression assay for use in future studies of ERBB2-based clinical management of breast cancer.

Comparison of enzyme immunoassay and immunohistochemical measurements of estrogen and progesterone receptors in breast cancer patients.

Before replacing enzyme immunoassay of estrogen and progesterone receptors by immunohistochemistry, results of both methods were compared on 437 samples obtained from breast cancer patients (342 primary breast carcinomas, 16 local recurrences, 49 biopsies, and 30 tumor specimens obtained after neoadjuvant treatment). Immunohistochemistry (IHC) results were first assessed semiquantitatively on the basis of the estimated proportion of positive tumor cells, and then quantitatively using the "quick score." Semiquantitative IHC hormone receptors results (positive > or = 10%) correlated well with enzyme immunoassay status (positive >15 fmol/mg protein) in 358 surgical samples (342 primary tumors and 16 recurrences), with overall concordance rates of 89.9% and 82.1%, respectively. Among the 100 discordant cases, a large intraductal carcinoma component was observed in 7 of 36 cases for estrogen receptor (ER) and 15 of 64 for progesterone receptor (PR). Thirty-five discordant cases also were observed near the cut-off values. Hormone receptor levels by enzyme immunoassay correlated strongly with the quantitative IHC "quick score." Whatever the method, hormone receptor status was associated with histologic grade (SBR) and tumor size, whereas age correlated strongly with ER positivity. Similar results were obtained for biopsy specimens and posttreatment samples. This comparison improved the reliability of the IHC technique, which is currently routinely used for ER and PR determination in the authors' institution.

Prognostic factors in localized primary synovial sarcoma: a multicenter study of 128 adult patients.

PURPOSE: To identify most significant and therapeutically relevant prognostic factors in adults with localized primary synovial sarcomas (SS) and to confirm the usefulness of the French Federation of Cancer Centers (FNCLCC) grading system, the prognostic impact of which has been already proven in soft tissue sarcomas. PATIENTS AND METHODS: Data on 128 patients with nonmetastatic SS collected from a cooperative database by the FNCLCC Sarcoma Group between 1980 and 1994 were studied retrospectively. Immunohistochemistry was performed at diagnosis in 77 cases (61%). The tumors were classified as biphasic (n = 45), monophasic fibrous (n = 72), and poorly differentiated (n = 10) subtypes. Histologic grade was determined according to the FNCLCC method, and vascular invasion was assessed in every case. RESULTS: The 5-year disease-specific survival (DSS) rate for this series of patients with localized SS was 62.9% (+/- 9.6% [SD]) with a median follow-up time of 37 months (range, 8 to 141 months). In multivariate analysis, the adverse risk factors associated with decreased DSS were International Union Against Cancer/American Joint Committee on Cancer stage III/IVA disease, male sex, and truncal tumor locations. For metastasis-free survival (MFS), disease stage III/IVA, tumor necrosis, and monophasic subtypes were the major factors associated with a less favorable prognosis. Separately, when not using disease stage, tumor necrosis, and mitotic activity, histologic grade became the most significant prognostic factor for both DSS and MFS. In addition, larger tumors and older patients become associated with a significantly worse prognosis. Independent adverse risk factors for local recurrence-free survival included histologic grade 3 and truncal tumor location. CONCLUSION: These data confirm that not all SS present the same severe outcome. High-risk patients identified on the basis of these parameters may qualify for an aggressive treatment approach.

Comparison of Cyfra 21-1 and SCC assays in head and neck tumours.

Patients with head and neck tumours (HNT) have a high risk of early locoregional relapse that is difficult to diagnose. This study evaluated the usefulness of the serum Cyfra 21-1 assay compared to squamous cell carcinoma antigen (SCC) assay for monitoring such patients. Three hundred and twelve HNT patients, including 204 newly diagnosed patients, were followed up for a median of 446 days with serial serum assays for SCC and Cyfra 21-1. Untreated patients showed SCC and Cyfra 21-1 serum levels correlated with each other: concentration was correlated to clinical stage, tumour size (as T1 + T2 vs. T3 + T4) and nodal status. Cyfra 21-1, but not SCC, was related to the presence of metastases and the primary tumour site, with a univariate prognostic value for disease-free survival (p = 0.015). Cox's regression analysis showed that only Cyfra 21-1 was associated with a risk of relapse (p = 0.027). The random coefficient growth curve model applied to serial SCC and Cyfra 21-1 measurements of 111 patients showed that only Cyfra 21-1 exhibited a significant difference between patients with and without relapses. We found Cyfra 21-1 to be more closely related to initial clinical data and disease evolution than SCC, and therefore propose the use of Cyfra 21-1 for monitoring head and neck cancers.

[Sem: a suitable statistical software adaptated for research in oncology]. / Sem: un outil de gestion informatique et statistique adapté à la recherche en cancérologie.

Many softwares have been adapted for medical use; they rarely enable conveniently both data management and statistics. A recent cooperative work ended up in a new software, Sem (Statistics Epidemiology Medicine), which allows data management of trials and, as well, statistical treatments on them. Very convenient, it can be used by non professional in statistics (biologists, doctors, researchers, data managers), since usually (excepted with multivariate models), the software performs by itself the most adequate test, after what complementary tests can be requested if needed. Sem data base manager (DBM) is not compatible with usual DBM: this constitutes a first protection against loss of privacy. Other shields (passwords, cryptage...) strengthen data security, all the more necessary today since Sem can be run on computers nets. Data organization enables multiplicity: forms can be duplicated by patient. Dates are treated in a special but transparent manner (sorting, date and delay calculations...). Sem communicates with common desktop softwares, often with a simple copy/paste. So, statistics can be easily performed on data stored in external calculation sheets, and slides by pasting graphs with a single mouse click (survival curves...). Already used over fifty places in different hospitals for daily work, this product, combining data management and statistics, appears to be a convenient and innovative solution.

Immunohistochemistry on cell blocks from fine-needle cytopunctures of primary breast carcinomas and lymph node metastases.

We assessed the reliability of prognostic biologic markers by means of immunohistochemistry on cell blocks obtained from diagnostic fine-needle cytopunctures of breast carcinomas and their lymph node metastases. Immunohistochemical studies of MIB-1 (Ki-67), estrogen receptors (ER), progesterone receptors (PR), p53, and c-erb-B-2 were performed in 55 cases of primary breast carcinoma on cell blocks (cytoblock technique) and on their corresponding tissue samples (46 mastectomy specimens and 9 Trucut biopsies) and in 38 cases on cell blocks from fine-needle cytopunctures of both the primary breast tumors and their concurrent lymph node metastases. Interobserver reproducibility ranged from 87 to 100%, depending on the marker. A good correlation was observed between immunostaining assessment on cell blocks and on the corresponding tumor tissues as follows: Ki-67 (85%), ER (96%), PR (82%), p53 (76%), and c-erb-B-2 (84%). An excellent correlation was observed between cell-block results for primary tumors and node metastases; however, a far higher percentage of Ki-67-positive nuclei was observed in the nodes than in the corresponding tumors in seven cases. All nodes corresponding to ER- or PR-negative tumors were also negative, whereas the nodes corresponding to two ER-positive and one PR-positive tumor were negative. Marked discrepancies were also noted with p53 in two cases and with c-erb-B-2 in two cases. Most discrepancies occurred with Trucut biopsies and with breast tumors that contained a large intraductal component. We conclude that cell blocks prepared from fine-needle cytopuncture specimens of breast carcinomas and their node metastases are useful when planning neoadjuvant treatment.

Is adjuvant tamoxifen recommended in post-menopausal node-negative breast cancer patients with high estrogen receptor values?

The impact of ER levels on the response to tamoxifen was evaluated in 1,623 postmenopausal primary breast cancer patients treated at our center (median follow-up 8.2 years). In patients receiving adjuvant tamoxifen a significantly longer disease-free survival (DFS) was observed when ER levels were elevated (p<0.00001). Very high ER (>424 fmol/mg protein) appeared to be detrimental in node-negative patients not treated with tamoxifen.

Relationship between c-erbB-2 and other tumor characteristics in breast cancer prognosis.

The aim of this study was to evaluate c-erbB-2 overexpression by means of a quantitative biochemical technique in 488 primary breast cancer patients with long-term follow-up (median, 10 years) and its relation to other biochemical prognostic factors (uPA, p53, and epidermal growth factor receptor) and adjuvant therapy. High levels of c-erbB-2 (>500 IU/mg protein) were associated with estrogen receptor (ER) and progesterone receptor negativity, high histoprognostic SBR grade and high levels of uPA and p53. Univariate analyses showed shorter metastasis-free survival (MFS) and overall survival (OS) in patients whose tumors overexpressed c-erbB-2 in the overall population, in subgroups defined by ER and uPA status, and in patients with positive pathological nodal status, SBR grade II, progesterone receptor, and p53-negative tumors. Patients with ER-positive, c-erbB-2-positive tumors had a shorter MFS and OS than those patients with c-erbB-2-negative tumors. No difference was observed between adjuvant-treated and untreated patients (chemotherapy and/or hormone therapy) in the c-erbB-2-negative subgroup. There was a trend toward a longer short-term MFS in c-erbB-2-positive patients treated with chemotherapy, whereas an opposite effect was observed with hormone therapy. Cox multivariate analyses showed that high levels of c-erbB-2 negatively influenced MFS in the overall population as well as in node-positive patients and in tamoxifen-treated patients, along with pN and uPA. Results for OS were comparable with those obtained for MFS. These results suggest that c-erbB-2 overexpression in breast cancer may be a better predictor of the response to tamoxifen than is ER status alone.

[Breast cancer: prognostic value of a dissemination index based on 4 components of the urokinase-type plasminogen activator system]. / Cancer du sein: valeur pronostique d'un index de dissémination basé sur quatre composants du système activateur du plasminogène type urokinase.

Among the proteases involved in the tumor invasion process, components of the plasminogen activator system (plasminogen activator type-urokinase uPA, its membrane receptor uPAR and its two inhibitors PAI-1 and PAI-2) appear to define high risk patients in primary breast cancer. As individual analysis of each component of the plasminogen activator system does not reflect the complex interactions between the different components, we studied the prognostic impact of a dissemination risk index combining the four variables. We found that this index was the most powerful prognostic factor, particularly in node-negative patients.

Dissemination risk index based on plasminogen activator system components in primary breast cancer.

PURPOSE: To study interactions between disease-free survival (DFS) and four components of the plasminogen activator system: urokinase-type plasminogen activator (uPA), its two inhibitors (PAI-1 and PAI-2), and its membrane receptor uPAR. PATIENTS AND METHODS: We conducted a retrospective study of 499 primary breast cancer patients (median follow-up, 6 years). uPA, PAI-1, and PAI-2 were determined on cytosols and uPAR on solubilized pellets, using enzyme-linked immunoadsorbent assay kits (American Diagnostica, Greenwich, CT). Classical univariate and multivariate statistical methods were used together with multiple correspondence analysis to graphically examine interactions between the variables and outcome. RESULTS: By univariate analysis, higher uPA and PAI-1 values were significantly related to shorter DFS (P =.002; P <.00002). PAI-2 was not significantly related to DFS, although patients with high and very low PAI-2 values had a longer DFS. Multiple correspondence analysis showed the parallel impact of uPA and PAI-1 on outcome, and the clearly different behavior of PAI-2 compared with PAI-1. The prognostic contribution of uPAR seemed weak by both methods. A dissemination risk index [uPA x PAI-1/(PAI-2 + 1)], taking into account the modulation of uPA proteolytic activity by the ratio of its two inhibitors, was then tested. Dissemination risk index was selected as an independent variable in the Cox model in the overall population (P <.000001) and in node-positive patients (P <.00001). It was the only variable selected in node-negative patients (P =. 003). CONCLUSION: A dissemination risk index determined on primary tumor and taking into account the different effects of PAI-1 and PAI-2 on uPA can be of major help in clinical management of breast cancer, particularly in node-negative patients.