| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Clinical Investigations |
Department of Radiation and Cellular Oncology, The Pritzker School of Medicine, The University of Chicago, Chicago, Illinois 60637
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
20–30% probability of developing metastatic disease
(2
, 3) . The presently available markers of prognosis do
not allow us to distinguish at diagnosis who the minority of the
node-negative patients are who will actually develop distant
metastases. For metastases to occur, several progressive changes in the phenotype are needed. These include neovascularization, decreased adherence of the tumor cells to each other, increased motility, adhesion to the extracellular matrix, and degradation of the extracellular matrix (4, 5, 6) . We have developed a strategy to identify prognostic markers based on biomarkers for these metastatic phenotypes.
Folkman and colleagues (7, 8, 9) have shown that tumor growth, progression, and metastasis require angiogenesis. Growth of a tumor beyond 1–2 mm3 is dependent on angiogenesis (9) . Tumor spread to distant sites is also dependent on access to the vasculature. The higher the count of microvessels and the larger the surface area of these vessels, the higher the probability that tumor cells will enter the circulation (10 , 11) . We, as well as others, have shown that when angiogenesis is measured by counting immunohistochemically stained microvessels, there is an excellent correlation between the MVC2 , the metastatic propensity, and the long-term DFS (12, 13, 14, 15) .
Nm23 is a tumor suppressor family of genes that inhibits metastases. Nm23-H1 has been shown to suppress metastatic potential in human carcinoma cell lines (16 , 17) . It has also been implicated in regulating basement protein deposition and restoring the normal phenotype to metastatic breast cells in culture. Transfection of the nm23-H1 gene suppresses the cytokine-stimulated motility of human breast carcinoma cells (18) . A correlation between decreased levels of nm23-H1 expression and lymph node metastases (19, 20, 21, 22, 23) , tumor grade (21 , 24) , or outcome (21 , 25 , 26) have been demonstrated. We have shown that, particularly in tumors that have high angiogenesis, if the expression of nm23-H1 is intact, the outcome is still excellent (27) . None of these markers allowed us to distinguish a very poor prognosis group because even in our previously shown "bad" group, i.e., with high angiogenesis and low nm23-H1, two-thirds of the patients survived long-term. This indicated to us the need for additional markers.
E-cadherin is a calcium-regulated homophilic adhesion molecule. The gene has been cloned and is located on chromosome 16q22.1 (28) . The extracellular domain is involved in cell-cell adhesion, whereas the intracellular domain connects to the actin cytoskeleton via catenins. It has a significant function in the epithelial intercellular junction complex, the establishment of epithelial polarization, glandular differentiation, and stratification (29) . It is a component of the adherent junctions, and it concentrates the urokinase plasminogen activator and the epidermal growth factor receptor to cell contact sites (30 , 31) . In development, it is an important regulator of morphogenesis (32 , 33) . E-cadherin knockouts have been shown not to be viable and demonstrate abnormal epithelial morphogenesis (34) .
Decreased E-cadherin-mediated adhesion is also one of the changes characterizing the invasive phenotype (29) . In cell lines, an inverse relationship between levels of E-cadherin expression and invasion has been shown (35) . In vitro data provide evidence that E-cadherin acts as an invasion suppressor molecule and that the level of expression of E-cadherin is related to invasive characteristics. Transfection of cDNA encoding E-cadherin into highly invasive mouse mammary tumor cell lines (36) resulted in decreased invasiveness in vitro and decreased metastasis in vivo (37) .
Down-regulation of E-cadherin expression has been observed in many human carcinomas (29 , 38) . In general, tumors with lower expression of E-cadherin are more infiltrating, lower grade, and are more likely to have spread to lymph nodes. In breast cancer, a correlation between decreased levels of E-cadherin expression and lymph node metastases (39, 40, 41) , tumor grade (40 , 42 , 43) , or outcome (40 , 42, 43, 44) have been demonstrated. The majority of these studies have either a short follow-up or include heterogeneous patient populations treated with varied adjuvant therapies that confound meaningful prognostic conclusions.
We hypothesized that E-cadherin, an epithelial-epithelial adhesion molecule, will add prognostic information to that of nm23-H1 and angiogenesis because it characterizes a different step in the malignant progression. We used archival material from our previously described database of patients treated with mastectomy between 1927 and 1987 (3 , 45, 46, 47, 48, 49, 50) . None of the patients received adjuvant hormone therapy or chemotherapy, which may confound the natural history, and there is sufficient follow-up (median 14 years) so that the entire natural history of the disease has been expressed. This is particularly important in breast cancer because of its long natural history (48 , 50) .
|
PATIENTS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. The median age of the patients is 57 years (range, 29–82 years). The
treatment consisted of surgery only: radical or extended radical
mastectomy in 73%, modified radical mastectomy in 23%, and other type
of surgery in 4% of the patients. The median number of lymph nodes
removed is 23 (range, 1–68). Tumor size was determined by gross
measurements of the excised lesion or by the largest tumor diameter as
measured on a histological section. The tumor size distribution is
shown in Table 1
. The majority of the tumors (88%) are 
3 cm. Tumor
nuclear grading was done using the pathological criteria described by
Fisher et al. (51)
. In six patients, the tissue
preservation was not satisfactory for nuclear grading.
|
E-cadherin.
Standard immunohistochemical detection with minor modification was
performed on sections from archival paraffin-embedded tissue to detect
the expression of E-cadherin and nm23-H1 and to assess angiogenesis
(anti-CD34 monoclonal antibody; Refs. 14
, 27
, 52 ). To
detect E-cadherin, five-micrometer sections mounted onto pretreated
slides were deparaffinized and rehydrated in graded alcohols and
distilled water. The slides were rinsed in PBS and microwaved at high
power for two cycles of 5 min each with a 10-min break between cycles
(41)
in citrate buffer (pH 6.0). The samples were allowed
to cool to room temperature, rinsed with PBS, and incubated in 10%
normal horse serum in PBS containing 1% BSA. After the PBS rinse, the
tumor sections were incubated at 4°C overnight with primary antibody
(mouse monoclonal antihuman E-cadherin antibody, clone HECD-1, Zymed
Laboratories, San Francisco, CA) diluted 1:100 in PBS (40
, 53)
. After rinsing in PBS, the slides were incubated with
biotinylated antimouse immunoglobulin G (VectorBA-2000; Vector
Laboratories, Burlingame, CA) for 1 h at room temperature. After
rinsing with PBS, endogenous peroxidase activity was quenched with
0.3% H2O2 in PBS for 10
min at room temperature. The slides were again rinsed with PBS and than
incubated for 1 h at room temperature with the
avidin-biotin-peroxidase complex (Vectastain Elite ABC, Vector
Laboratories). The slides were again rinsed with PBS, developed with
3,3-diaminobenzidine (Sigma Chemical, St. Louis, MO) chromogen solution
(0.05% 3,3'-diaminobenzidine in PBS and 0.006%
H2O2) dipped in 0.125%
osmium tetroxide (Sigma Chemical) to enhance positivity, counterstained
with 1% methyl green (Trevigen, Gaithersburg, MD), dehydrated in
graded alcohol, air dried, and mounted using Pro-Tex (American
Scientific Products, McGraw Park, IL) mounting medium under coverslips.
Some variability in staining intensity was present. Previously
described semiquantitative criteria for scoring, with minor
modifications, was used (39
, 40 , 54)
. The tissue was
scanned for areas of well-preserved tumor. Necrotic areas and areas
where the tissue had deteriorated morphology were excluded. In areas of
well-preserved tissue, the fraction of the positive staining cell was
scored. The staining of normal duct epithelium was used as the internal
control for each section. Normal skin was also included as the control.
Staining of >75% of the cells and comparable to normal glands was
scored as high = 3, clearly recognizable but weaker than
normal, or heterogeneous in >25%, but <75% of the cells was scored
as intermediate = 2, just identifiable staining
in < 25%, or none was scored as = 1.
Five consecutive high power fields images were captured and evaluated
using Image Pro (Media Cybernetics, Silver Spring, MD) image
analysis software.
Nm23-H1.
To detect nm23-H1 protein, 5-µm sections mounted onto pretreated
slides were deparaffinized and rehydrated in graded alcohols and
distilled water. Tumor sections were incubated at 4°C overnight with
primary antibody (mouse monoclonal anti-nm23-H1, Novocastra
Laboratories, Newcastle, United Kingdom) diluted 1:50. The
immunoperoxidase detection system (Vectastain Elite ABC) was used with
3,3-diaminobenzidine as chromogen as described for E-cadherin
detection. Some variability in staining intensity was present. The
previously described scoring (24
, 25 , 27)
based on the
proportion of stained cells was used. High nm23 was defined as all
cells expressing nm23-H1. If there were any tumor cells not staining,
i.e., "cold spots," the tumor was scored as low nm23.
Sufficient archival material was available in 147 patients.
Angiogenesis (MVC).
To assess angiogenesis, we used our previously described
immunohistochemical method (33)
. In short, from the
archival tissue block in which tumor was confirmed and graded, 5-µm
sections mounted onto pretreated slides were deparaffinized and
incubated at 4°C overnight with the primary antibody (monoclonal
anti-CD34/QB-END, Novocastra Laboratories), diluted 1:25 in PBS. The
immunoperoxidase detection system used for E-cadherin and nm23-H1 was
also used to detect anti-CD34 binding (Vectastain Elite ABC;
3,3-diaminobenzidine as chromogen). Vascularity was determined by the
number of vessels per field counted in the area of highest vascular
density ("hot spots") at 400x magnification (0.1452 mm; ref.
2
). Previously recommended guidelines were followed
(15
, 55, 56, 57)
. Single endothelial cells, endothelial cell
clusters, and microvessels in the tumor clearly separated from adjacent
microvessels were counted. Peritumoral vascularity and vascularity in
areas of necrosis were not scored. Branching structures were counted as
a single vessel. The presence of lumen or erythrocytes in the lumen was
not required to classify a structure as a vessel. If the vascularity
was uniform, microvessels in three fields were counted and averaged. If
the vascularity in different fields was not uniform, up to 10 fields
were counted, and the three highest counts were averaged. Low
angiogenesis as previously detailed was defined as <15
microvessels/endothelial cells at 400x magnification (14
, 15)
. E-cadherin, nm23-H1, nuclear grading, and the MVC were
performed without the knowledge of the patients’ outcome.
Statistical Analyses.
Actuarial survival curves were calculated according to the Kaplan-Meier
method (58)
, and comparisons were made with the log-rank
test (59)
. Patients were censored at last follow-up if
they were free of disease and were considered dead of disease if they
were dead and were known to have recurrent disease at the last
evaluation. DFS was defined as the elapsed time from mastectomy to
disease recurrence or death. Death of disease or any recurrent disease
local or distant was considered as an event in DFS calculation.
Patients were censored for death due to intercurrent disease.
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. The 14-year DFS of the entire group is 78% (95% CI,
69–84). We choose the 14-year point because the median follow-up is 14
years. The long-term DFS as a function of E-cadherin expression is
shown in Fig. 1B
. There is a significant difference in
outcome as a function of E-cadherin expression (P = 0.005). If E-cadherin is low, the 14-year DFS is 56% (95% CI,
34–71), but if E-cadherin is intermediate or high, the 14-year DFS is
80% (95% CI, 64–89) and 84% (95% CI, 72–92), respectively.
|
. There is a
significant difference in outcome as a function of E-cadherin
expression. The 14-year DFS is 76% (95% CI, 48–91), 61% (95% CI,
34–79), and 44% (95% CI, 22–65), respectively if E-cadherin is
high, intermediate, or low (P = 0.01).
To determine which are the most significant prognostic biomarkers, we
performed multivariate analyses, which are shown in Table 2
. As a first step in a stepwise multivariate analysis, we included
E-cadherin, nm23-H1, MVC, tumor size, tumor grade, age, and ER status.
ER was the least significant; thus, as a first step, we excluded it
from the analysis. Next, we excluded grade because of the high
P (0.6) and age because of the noninformative hazard rate.
This analysis demonstrates that the biomarkers E-cadherin, nm23-H1, and
MVC appear to be more important variables than tumor size, grade, age,
or ER.
|
shows
the long-term DFS in patients with high/intermediate E-cadherin,
comparing patients with either low MVC or high nm23 to those with both
high MVC and low nm23. The 14-year DFS is 92% (95% CI, 82–97)
versus 69% (95% CI, 50–82), respectively
(P = 0.017). Thus, nm23 and MVC add
prognostic information to that obtained from E-cadherin.
To further understand the relative importance of E-cadherin, nm23-H1
and angiogenesis, and breast cancer progression, we analyzed the
outcome as a function of the combinations of the three biomarkers.
Table 3
shows the 14-year DFS and number of patients in the eight possible
combinations of the three biomarkers. This is also further detailed in
Fig. 2
. This figure allows us to identify events that are unlikely to occur
and possibly assign some hierarchy to the events. The highest 14-year
DFS is seen in patients in whom E-cadherin and nm23-H1 are high, and
this is irrespective of the MVC count (93% and 94%). This group
includes a substantial number of patients (54 patients). The group with
slightly lower 14-year DFS is the one with high E-cadherin, low MVC,
and low nm23-H1(88%). There is a more substantial decrease in
long-term DFS if MVC is high, nm23-H1 is low, but E-cadherin is high
(69%). Interestingly, there are very few patients in three groups: low
MVC, high nm23-H1, but low E-cadherin; low MVC, low nm23-H1, and low
E-cadherin; and high MVC, high nm23-H1, and low E-cadherin, indicating
that these combination of events are much less likely to occur. The
worst 14-year DFS is in the group with low E-cadherin, low nm23-H1, and
high MVC (44%).
|
|
. In this table, we combined high and intermediate E-cadherin into one
group and referred to it as high E-cadherin. The expression of
E-cadherin was low in 19% of the tumors tested. There is a trend
(P = 0.07) for decreased expression of
E-cadherin with an increase in tumor size. A larger proportion of
>2-cm tumors (25%) have low E-cadherin compared to 2-cm tumors
(14%). A trend is also noted as a function of age. Older women have a
smaller percent of low E-cadherin tumors. There is a strong correlation
between nuclear grade and E-cadherin. Forty-nine percent of grade 3
tumors compared to 11% of grade 1 tumors have low E-cadherin
expression (P < 0.001). The ER-positive
tumors are more likely to also express E-cadherin
(P = 0.04). Angiogenesis appears to be
inversely correlated with E-cadherin expression (P = 0.05), whereas a trend for a direct correlation with nm23-H1 is
noted.
|
|
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Several studies have demonstrated a correlation between E-cadherin expression, tumor pathological features, and outcome in breast cancer. Lipponen et al. (44) show a trend toward better relapse-free survival in node-negative patients who have high E-cadherin expression. The follow-up was >10 years, and some patients received systemic adjuvant therapy. Their entire patient population includes 208 patients (some have node-positive disease or metastatic disease), and in 13% of the cases, the pathological lymph node status was unknown. Thus, it is difficult to determine the number of node-negative patients. Guriec et al. (42) observed an association between mRNA expression and survival in 42 node-negative patients followed for a median of 90 months. Siitonen et al. (40) also show a significantly higher DFS with invasive breast carcinoma expressing high E-cadherin. The prognostic value of E-cadherin was maintained in multivariate analysis. But the patient population in whom clinical follow-up was available is quite heterogeneous and includes 109 node-positive and node-negative patients, some of whom received various adjuvant therapies. Charpin et al. (43) limit their analysis to patients who did not receive adjuvant systemic therapy. At 10 years in the 82 node-negative patients, E-cadherin is a significant prognostic factor for overall survival. These studies indicate that E-cadherin has the potential to be a valuable prognostic marker, but most are heterogeneous, having many confounding variables.
Our patient population consists of node-negative patients who underwent mastectomy and received no other adjuvant treatment; thus, the natural history is not perturbed (48 , 50) . The long-term DFS of the patients in this group is 78%. We find that E-cadherin is the strongest prognostic factor for long-term outcome. Alone, it identifies a group of patients with a long-term DFS of 56%, and in combination with nm23-H1 and MVC, it identifies a group with an even lower long-term DFS of 44%. Hence, loss of E-cadherin expression appears to be a major determinative step in the metastatic progression. The combination of loss of E-cadherin, loss of nm23-H1, and high angiogenesis results in a worse long-term DFS. E-cadherin by itself is less valuable in identifying the good prognosis group. The long-term DFS in the patients with high or intermediate E-cadherin is 82%. In this group, nm23-H1 and angiogenesis identify the subgroups with higher 14-year DFS, 94% if nm23-H1 is high, or 91% if angiogenesis is low.
In Table 3
and Fig. 2
, we attempt to combine the information on the
three biomarkers. Multivariate analysis is excellent in indicating
which is the most powerful prognostic factor, but it is also important
to relate this acquired phenotype to each other. In Fig. 2
, we assume
that the tumors start out as having low angiogenesis and full
expression of nm23-H1 and E-cadherin. The arrows are hypothetical paths
that are proposed based on outcome, number of patients, and correlation
between the markers. A bin with few patients may indicate that path to
be an unlikely sequence of events. For example, there is only one
patient with low angiogenesis and high nm23-H1 who has loss of
E-cadherin; similarly, there are only a few patients (seven) who have
high angiogenesis and high nm23-H1 who have loss of E-cadherin. Thus,
irrespective of the extent of angiogenesis, there are few patients in
whom E-cadherin expression is lost if nm23-H1 is fully expressed. A
possible explanation may be that loss of E-cadherin occurs subsequent
to the loss of nm23-H1.
It appears that among the three biomarkers, E-cadherin, nm23-H1, and MVC, MVC contributes the least prognostic information. If E-cadherin and nm23-H1 are high, even if MVC is high, the long-term survival is excellent. But if nm23-H1 expression is lost, the survival is lower, and if there is also a subsequent loss E-cadherin, there is an even further decrease in the long-term DFS to 44%. Angiogenesis appears to be the least significant of these biomarkers. If both nm23-H1 and E-cadherin expression are normal, an increase in angiogenesis does not result in further metastases. But if nm23-H1 expression is lost, then increased angiogenesis appears to be a facilitator of metastases because the long-term DFS is 88% if MVC is low and only 69% if MVC is high. There are very few patients with low MVC who have loss of E-cadherin, also indicating that loss of E-cadherin is a latter event in the metastatic progression. Both nm23-H1 and MVC are excellent in predicting the good prognosis patients. Although MVC may appear less significant, it still has prognostic value because there is no correlation between MVC and nm23-H1 and thus they identify different patients. However, the worst outcome is when all three biomarkers are "bad." Hence, all three markers contribute prognostic information superior to the traditional prognostic markers, tumor size, and grade.
But even using the combined information from E-cadherin, nm23, and MVC, the worst outcome group still has a 44% long-term survival. Therefore, in this group, 44% of the patients would be receiving systemic therapy unnecessarily. Further biomarkers have to be identified in these patients. The most clinically relevant prognostic marker combination is that which identifies patients with very low or very high long-term DFS because those with very high long-term DFS will not need chemotherapy, whereas those with very low DFS need aggressive systemic treatments.
Promising biomarkers for predicting risk of metastases are likely to come from understanding the metastatic progression in human breast cancer. The markers analyzed in this study represent angiogenesis (MVC), epithelial-epithelial adhesion (E-cadherin), and motility/signal transduction (nm23-H1). Integration of biomarkers representing different steps in the metastatic progression will likely offer the best hope in prognosticating the outcome at diagnosis. Further, useful markers may come from invasion markers, such as proteolytic enzymes, and molecules characterizing the adhesion to the extracellular matrix. As in this study, the most valuable information both for the understanding of metastatic progression and prognosis will be obtained from studies done on patients followed for long periods of time and who received no adjuvant systemic therapy and therefore in whom the natural history is not perturbed.
ACKNOWLEDGMENTS
We thank Janet Riley for secretarial assistance and
Rolando Torres for technical assistance.
|
FOOTNOTES |
|---|
1 To whom requests for reprints should be
addressed, at Department of Radiation and Cellular Oncology, The
University of Chicago, MC 9006, 5758 South Maryland Avenue, Chicago, IL
60637. Phone: (773) 702-5194; Fax: (773) 834-7340; E-mail: heimann{at}rover.uchicago.edu 
2 The abbreviations used are: MVC, microvessel
count; DFS, disease-free survival; CI, confidence interval; ER,
estrogen receptor.
Received 7/28/99. Accepted 11/10/99.
|
REFERENCES |
|---|
|
TopABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  K. L. Andarawewa, A. C. Erickson, W. S. Chou, S. V. Costes, P. Gascard, J. D. Mott, M. J. Bissell, and M. H. Barcellos-Hoff Ionizing Radiation Predisposes Nonmalignant Human Mammary Epithelial Cells to Undergo Transforming Growth Factor {beta} Induced Epithelial to Mesenchymal Transition Cancer Res., September 15, 2007; 67(18): 8662 - 8670. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Brouxhon, S. Kyrkanides, M. K. O'Banion, R. Johnson, D. A. Pearce, G. M. Centola, J.-n. H. Miller, K. H. McGrath, B. Erdle, G. Scott, et al. Sequential Down-regulation of E-Cadherin with Squamous Cell Carcinoma Progression: Loss of E-Cadherin via a Prostaglandin E2-EP2 Dependent Posttranslational Mechanism Cancer Res., August 15, 2007; 67(16): 7654 - 7664. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Harigopal, A. J. Berger, R. L. Camp, D. L. Rimm, and H. M. Kluger Automated Quantitative Analysis of E-Cadherin Expression in Lymph Node Metastases Is Predictive of Survival in Invasive Ductal Breast Cancer Clin. Cancer Res., June 1, 2005; 11(11): 4083 - 4089. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. K. Mell, J. J. Meyer, M. Tretiakova, A. Khramtsov, C. Gong, S. D. Yamada, A. G. Montag, and A. J. Mundt Prognostic Significance of E-Cadherin Protein Expression in Pathological Stage I-III Endometrial Cancer Clin. Cancer Res., August 15, 2004; 10(16): 5546 - 5553. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. C. Park, R. L. Henshall-Powell, A. C. Erickson, R. Talhouk, B. Parvin, M. J. Bissell, and M. H. Barcellos-Hoff Ionizing radiation induces heritable disruption of epithelial cell interactions PNAS, September 16, 2003; 100(19): 10728 - 10733. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Anders, R. Hansen, R.-X. Ding, K. A. Rauen, M. J. Bissell, and W. M. Korn Disruption of 3D tissue integrity facilitates adenovirus infection by deregulating the coxsackievirus and adenovirus receptor PNAS, February 18, 2003; 100(4): 1943 - 1948. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A N J Graham, P Maxwell, K Mulholland, A H Patterson, N Anderson, K G McManus, H Bharucha, and J A McGuigan Increased nm23 immunoreactivity is associated with selective inhibition of systemic tumour cell dissemination J. Clin. Pathol., March 1, 2002; 55(3): 184 - 189. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. G. Corn, E. I. Heath, R. Heitmiller, F. Fogt, A. A. Forastiere, J. G. Herman, and T.-T. Wu Frequent Hypermethylation of the 5' CpG Island of E-Cadherin in Esophageal Adenocarcinoma Clin. Cancer Res., September 1, 2001; 7(9): 2765 - 2769. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Lundin, M. Lundin, K. Holli, V. Kataja, L. Elomaa, L. Pylkkanen, T. Turpeenniemi-Hujanen, and H. Joensuu Omission of Histologic Grading From Clinical Decision Making May Result in Overuse of Adjuvant Therapies in Breast Cancer: Results From a Nationwide Study J. Clin. Oncol., January 1, 2001; 19(1): 28 - 36. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. Bertucci, R. Houlgatte, A. Benziane, S. Granjeaud, J. Adelaide, R. Tagett, B. Loriod, J. Jacquemier, P. Viens, B. Jordan, et al. Gene expression profiling of primary breast carcinomas using arrays of candidate genes Hum. Mol. Genet., December 1, 2000; 9(20): 2981 - 2991. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Cancer Research | Clinical Cancer Research |
| Cancer Epidemiology Biomarkers & Prevention | Molecular Cancer Therapeutics |
| Molecular Cancer Research | Cancer Prevention Research |
| Cancer Prevention Journals Portal | Cancer Reviews Online |
| Annual Meeting Education Book | Meeting Abstracts Online |
