Abstract
We evaluated the clinical significance of micrometastases in axillary lymph nodes (AxLNs) of breast cancer patients for prediction of prognosis. Archived formalin-fixed paraffin-embedded AxLN specimens from 129 node-negative breast cancer patients diagnosed by routine H&E staining between 1986 and 1990 were subjected to carcinoembryonic antigen-specific reverse transcription-PCR analysis. Micrometastases were detected in 40 of 129 (31.0%) node-negative breast cancer patients. After a median follow-up period of 105.6 months, log-rank test analysis indicated that 10-year disease-free and overall survival rates by Kaplan-Meier methods were significantly better in patients without micrometastases than in patients with micrometastases[ disease-free survival, 87.6% versus 66.1% (P = 0.0008); overall survival, 93.7% versus 67.8% (P = 0.0024)]. The presence of micrometastases in AxLNs was revealed by multivariate analyses to be an independent and significant predictor of clinical outcome. The hazard ratio was 3.992 (95% confidence interval, 1.293–12.323; P = 0.0161) for relapse and 4.293 (95% confidence interval, 1.043–17.675; P = 0.0436) for cancer-related death. The molecular staging of AxLNs using reverse transcription-PCR is useful for prediction of clinical outcome in early-stage breast cancer patients and can provide a powerful and sensitive complement to routine histopathological analysis.
Introduction
AxLN3 status in breast cancer patients is considered the most useful predictor of prognosis because an inverse relationship exists between the number of positive nodes and patient prognosis (1 , 2) . About 80% of node-negative patients diagnosed by standard histopathological examination with H&E staining will have long-term disease-free survival after surgical therapy, whereas 15–20% of node-negative patients will experience recurrence and ultimately die of the disease (1 , 2) . Thus, detecting subgroups of patients at high risk for relapse is of critical importance for providing optimal chemo-endocrine treatment.
Previous studies, such as those reported by the International (Ludwig) Breast Cancer Study Group (3 , 4) , have found that routine H&E staining analysis of AxLNs may miss micrometastases in 9–25% of patients that can be readily detected by serial sectioning (3 , 5 , 6) and immunohistochemical staining (4, 5, 6, 7) . Moreover, prospective studies suggest that the presence of micrometastases in AxLNs, detected by any of the above-mentioned methods, is associated with significantly poor disease-free and overall survival rates (3, 4, 5, 6, 7) . However, these methods are too cumbersome, time-consuming, and costly to be used in routine examinations (8) .
Recently, RT-PCR has been used to detect micrometastases in regional lymph nodes (9, 10, 11, 12, 13, 14, 15) , bone marrow (16 , 17) , and peripheral blood (14 , 16 , 18 , 19) by amplification of mRNA markers specific for epithelial cells, such as CK19 (9 , 10 , 13 , 16 , 17) , CK20 (20) , MUC-1 (13) , CEA (11 , 12 , 14 , 15) , and so on. Moreover, screening for micrometastatic disease in AxLNs by RT-PCR amplification has been shown to be more sensitive and cost effective than serial sectioning and immunohistochemical staining (8) . However, to our knowledge, the RT-PCR detection method has not been proven clinically applicable due to a paucity of prospective data on the long-term prognosis of patients with micrometastases.
We recently succeeded in efficiently extracting RNA from archived formalin-fixed surgical specimens (21) . Adding a simple heating process to eliminate the majority of chemical modifications of mRNA by formaldehyde before reverse transcription enabled us to use archived AxLN specimens for RT-PCR analysis of patients with proven prognoses (21) .
The purpose of this study was to evaluate the effectiveness of RT-PCR-based detection of micrometastases in AxLNs in predicting clinical outcome of breast cancer patients.
Materials and Methods
Patients.
We considered 215 consecutive patients diagnosed with primary unilateral breast cancer who had undergone radical mastectomy or quadrantectomy with AxLN dissection at Osaka University Hospital between 1986 and 1990. Ten patients were excluded from this study because distant metastatic lesions (stage IV disease) were found at operation. There were 203 women and 2 men. The mean age of the patients was 50.7 years at operation, with an age range of 24–87 years. Histopathological examination by H&E staining revealed metastases in AxLNs of 74 patients and no metastases in the remaining 131 patients. Histological tumor grade and the presence or absence of peritumoral vascular or lymphatic invasion were reexamined and recorded in a blind fashion by a single pathologist. Tumor types were classified histopathologically according to the Japanese general rules for clinical and pathological recording of breast cancer (22) .
RNA Extraction from AxLN Specimens.
Paraffinembedded blocks of AxLN tissue from 131 node-negative breast cancer patients were subjected to RNA extraction. The median number and mean number of AxLNs examined per patient were 17.0 and 18.6, respectively (the overall range was 5–50 AxLNs examined per patient). For any given patient, the number of AxLN specimens ranged between one and six paraffin blocks. Extraction of RNA and subsequent analysis were performed on pooled specimens from each patient, not on each individual lymph node.
Tissue blocks from 131 patients were trimmed of excess paraffin, finely minced into 1-mm3 cubes, weighed, and transferred to sterile 15-ml tubes. Paraffin was extracted twice by 20-min incubations in xylene at room temperature. Total RNA was extracted according to our original method, as described previously (21) . Briefly, each sample was homogenized in 1.0 ml of digestion buffer with 500 μg/ml proteinase K (Sigma Chemical Co., St. Louis, MO) per 50-mg tissue sample and incubated at 45°C for 1 h. The tissue lysate was extracted with 1.0 ml of liquified phenol-chloroform-isoamyl alcohol (25:24:1) followed by 1.0 ml of chloroform. After ethanol precipitation, the extracted RNA was purified by using the RNeasy mini kit (Qiagen GmbH, Hilden, Germany). To ensure elimination of genomic DNA, each microgram of total RNA was incubated with 1 unit of DNase I (Life Technologies, Inc., Gaithersburg, MD) at 37°C for 1 h and then inactivated at 75°C for 15 min. To check RNA for degradation and to adjust the quantity of RNA, a portion of each sample was electrophoresed on an agarose gel and stained with SYBR Green II (FMC Bioproducts, Rockland, ME); RNA yields were also measured spectrophotometrically.
RT-PCR Amplification of CEA and PBGD.
cDNA was synthesized from 1.0 μg of total RNA and 0.5 μg of oligo(dT) primers (Life Technologies, Inc.) in a 20-μl reaction mixture as described previously (21) . The preheating demodification process in which RNAs were incubated in TE buffer [10 mm Tris-HCl (pH 7.0), 1 mm EDTA] at 70°C for 30 min was done before annealing with oligo(dT) primers (21) . Reverse transcription was carried out by incubation with 200 units of SuperScript II (Life Technologies, Inc.) at 42°C for 60 min.
Primers for PCR amplification of CEA cDNA, which serves as a breast cancer cell marker, and PBGD, which serves as a positive control for RT-PCR, were designed with the aid of Oligo software version 4.0 (National Biosciences Inc., Plymouth, MN) and the results of earlier studies (11 , 12 , 14 , 15 , 23) . Primer sequences were as follows: (a) CEA forward primer, 5′-GAGCGAACCTCAACCTCTCCTGCCACT-3′; (b) CEA reverse primer, 5′-TGTAGCTGTTGCAAATGCTTTAAGGAAGAAGC-3′; (c) PBGD forward primer, 5′-GACTGGAGGAGTCTGGAGTC-3′; and (d) PBGD reverse primer, 5′-AATCCCTGGAAGGCTTGAAC-3′. These primer sets for each gene were designed to be intron-spanning. The primer for CEA forward and CEA reverse is located in exon 8 and exon 10 of CEA gene, respectively. The PBGD forward and PBGD reverse primers bond with regions of exon 14 and exon 15, respectively. RT-PCR for CEA yields a 366-bp fragment, and RT-PCR for PBGD yields a 186-bp fragment. Because the PBGD reverse primer is located at a distance of 281 bp upstream from the polyadenylation site, the detection of the 186-bp fragment from PBGD indicates that cDNA synthesis has been successfully performed with a length of at least 467 bp.
PCR amplification was performed in a 20 μl volume containing 1× PCR buffer (Perkin-Elmer, Foster City, CA); 0.4 μm of each specific primer; 144 μm each of dATP, dCTP, dGTP, and dTTP; 1 unit of AmpliTaq DNA polymerase (Perkin-Elmer); and 2 μl of cDNA from the reverse transcription reaction. After an initial incubation at 94°C for 5 min, the reaction mixtures were subjected to 35 cycles of amplification in the following sequence: 94°C for 30 s; 69°C for 30 s; and 72°C for 60 s. This was followed by a final extension step at 72°C for 10 min. PCR amplification was performed with a Genamp PCR System 9700 (Perkin-Elmer). For amplification of PBGD cDNA, the annealing temperature was set at 60°C instead of 69°C. Five μl of the reaction mixture were run on a 1.5% agarose gel and visualized with ethidium bromide. Samples were considered positive for micrometastases when cDNA of the predicted length for CEA was detected.
We performed all analyses in duplicate after the extraction of total RNA. If the results differed, we performed the experiment again. In experiments performed three times, results obtained more than twice were considered to be the final evaluation.
Sensitivity and Specificity of RT-PCR.
The sensitivity of this assay was initially established by performing RT-PCR analysis on RNA extracted from serial dilutions of MCF-7 cells with lymphocytes obtained from normal control lymph nodes. To imitate the assay’s detection of micrometastatic cancer cells in formalin-fixed paraffin-embedded lymph nodes, these cell pellets were treated with 10%-buffered-formalin for 16 h at 4°C and then washed with PBS. After RNA extraction and RT-PCR according to the above-mentioned method, the sensitivity of the assay in detecting CEA mRNA was determined.
To check the sensitivity and specificity in the case of clinical samples, we also examined 27 control lymph nodes obtained from noncancer patients who had undergone splenectomy, cholecystectomy, or vascular surgery. We then tested 40 tumor-involved lymph nodes from node-positive breast cancer patients diagnosed by routine H&E staining between 1986 and 1990. Paraffin-embedded blocks of these AxLN specimens were subjected to RNA extraction by the proteinase K-based method and the above-mentioned RT-PCR assay.
Statistical Analysis.
The χ2 test (or Student’s t test for continuous variables) and logistic regression model were used to assess the correlation of clinicopathological risk factors and micrometastases in AxLNs. The primary end point of this study was survival, which was measured from the date of surgery to the date of last follow-up or death. The disease-free survival period was defined as the period of time before the first relapse. The overall survival period was defined as the period of time before breast cancer-related death.
Univariate survival curves were calculated by the Kaplan-Meier method and compared by using the log-rank test. We used the Cox proportional hazards regression model in multivariate analysis to estimate the contribution of positive micrometastases in AxLNs to disease-free and overall survival after accounting for the potential effects of other explanatory clinicopathological risk factors. All tests were two-sided, and Ps < 0.05 were considered significant. Computations were carried out using StatView software version 5.0 (SAS Institute, Inc., Cary, NC).
Results
The sensitivity of this assay was established by performing RT-PCR analysis on RNA extracted from serial dilutions of MCF-7 cells with lymphocytes obtained from normal control lymph nodes and was determined to be 1 cancer cell/105 lymphocytes (Fig. 1⇓ A). PCR amplifications in various experiments in this study were carried out with the dilution of 1 MCF-7 cell/105 lymphocytes as the positive control sample that is the detection limit under our conditions. In the case of clinical samples, as shown in the Fig. 1⇓ B, CEA mRNA was not detected in any of the 27 normal lymph nodes under our experimental conditions. All 40 (100%) cases showed positive results for CEA-specific amplification (Fig. 1⇓ C).
Sensitivity and specificity of RT-PCR detection of micrometastatic cancer cells in AxLNs. The sensitivity of the assay was established by performing RT-PCR analysis on RNA extracted from serial dilutions of MCF-7 cells with lymphocytes obtained from normal control lymph nodes and was determined to be 1 cancer cell/105 lymphocytes (A). The sensitivity and specificity were calculated to be 100% by evaluating 27 control lymph nodes obtained from the noncancer patients (B) and 40 tumor-involved lymph nodes from node-positive breast cancer patients (C). Lane M, 1-kb DNA ladder (Life Technologies, Inc.). Lane N, cDNA synthesized from 1 of the 27 normal lymph nodes was used as a negative control sample. Lane P, cDNA synthesized from the diluted solution of 1 MCF-7 cancer cell/105 lymphocytes was used as a positive control sample.
Of 131 node-negative patients, 2 were excluded from this study because we could not obtain sufficient amounts of intact mRNA to carry out the RT-PCR assay. The PCR band of PBGD mRNA was detected in samples from the remaining 129 patients.
A 366-bp PCR product, which corresponded to CEA mRNA, was amplified in 40 of 129 (31.0%) node-negative patients. Examples of RT-PCR results for the 21 patients who experienced recurrence of the disease are shown in Fig. 2⇓ . Micrometastasis in AxLNs was found in 13 of the 21 (61.9%) patients.
Examples of RT-PCR results of 21 patients who experienced recurrence of disease during follow-up. Lane M, 1-kb DNA ladder. Lane N, negative control. Lane P, positive control.
Table 1⇓ shows the clinical and pathological details of the 74 node-positive breast cancer patients and the 129 node-negative breast cancer patients. The node-positive patients were more likely to have large tumors and were treated with adjuvant therapies more frequently. Adjuvant therapies were used in 168 patients according to various treatment protocols: (a) 137 patients received endocrine treatment, such as oral administration of tamoxifen and/or oophorectomy; and (b) 120 patients received chemotherapy (combination chemotherapy, such as treatment with cyclophosphamide, doxorubicin, and 5FU, for the node-positive group and oral administration of anticancer drugs, such as 5FU, for the node-negative group). The remaining 35 patients did not undergo adjuvant therapy.
%Characteristics of patients and tumors according to H&E staining
The median follow-up period was 105.6 months (range, 3.4–161.8 months). Of the 74 node-positive patients, 32 (43.2%) experienced relapse, and 26 (35.1%) died from progressive disease. Of the 129 node-negative patients, 21 (16.3%) relapsed, and 15 (11.6%) died.
Table 2⇓ shows the clinical characteristics of the 129 node-negative patients according to the presence or absence of micrometastases. No significant difference in age distribution, proportion of menopausal patients, or tumor characteristics was detected between micrometastasis-positive and -negative patients. Relapse and cancer-related death occurred in 13 (32.5%) and 10 (25.0%) patients, respectively, with evidence of micrometastases and in only 8 (9.0%) and 5 (5.6%) patients without evidence of micrometastases (χ2 test: P = 0.0008 and P = 0.0024, respectively).
%Characteristics of 129 node-negative patients and tumors according to the results of RT-PCR
Correlations between patient characteristics and the results of univariate analysis of clinical outcome estimated by the Kaplan-Meier method are shown in Table 3⇓ . The 10-year disease-free survival rate was 87.6% in patients without micrometastases and 66.1% in patients with micrometastases (P = 0.0008). Ten-year overall survival was significantly worse (P = 0.0024) in patients with micrometastases (67.8%) than in patients without micrometastases (93.7%). According to univariate analysis of the effect of micrometastases on disease-free and overall survival, the hazard ratio was 4.007 (95% CI, 1.660–9.675; P = 0.0020) and 4.553 (95% CI, 1.556–13.327; P = 0.0057), respectively. Candidate negative prognostic factors for disease-free survival in addition to micrometastases included high histological grade and lymphatic invasion. For overall survival, scirrhous type, negative estrogen receptor status, high histological grade, and vascular invasion were significantly associated with the worst outcomes.
%Univariate analysis of the prognostic value of micrometastases in AxLNs and other predictive factors for disease-free and overall survival in 129 node-negative breast cancer patients
Multivariate Cox regression analysis was performed with parameters measured in 106 patients with invasive ductal carcinoma. In addition to micrometastases in AxLNs, parameters included pathologically confirmed tumor type, estrogen receptor status, histological grade, lymphatic invasion, and vascular invasion, which were strong prognostic factors as determined by univariate analysis. As shown in Table 4⇓ , patient prognosis was complementary and attributable to the presence of micrometastases in AxLNs, with a hazard ratio of 3.992 (95% CI, 1.293–12.323; P = 0.0161) for relapse and a hazard ratio of 4.293 (95% CI, 1.043–17.675; P = 0.0436) for cancer-related death. No other prognostic factors were independent.
%Multivariable Cox regression analysis of disease-free and overall survival for 106 node-negative invasive ductal carcinoma patients
The survival curves according to nodal status are shown in Figs. 3⇓ and 4⇓ . The clinical outcome for patients with micrometastases was significantly worse than that for patients without micrometastases. The survival curve for patients with micrometastases fell between the micrometastasis-negative and node-positive patient populations.
Probability of disease-free survival for 203 breast cancer patients according to AxLN status.
Probability of overall survival for 203 breast cancer patients according to AxLN status.
Discussion
Although node-negative breast cancer patients have a favorable prognosis, recurrent disease occurs in approximately 20% of patients within 10 years after surgery (1 , 2) . Therefore, considerable efforts have been made to find markers conclusively associated with risk of relapse. In addition to menopausal status, tumor size, estrogen and progesterone receptor status, and histological grade, new prognostic candidates for markers in breast cancer tumors have been recommended to help make accurate estimates of relapse risk. These include c-erbB2 (24) , p53 (24 , 25) , cathepsin D (26) , plasminogen activators (27) , vascular endothelial growth factor (28) , and so forth. However, whether factors associated with tumor characteristics reflect prognosis independently remains controversial.
Another approach to identifying high-risk node-negative patients is to detect AxLN micrometastases, which cannot be detected by routine H&E analysis. Several studies have shown that H&E staining of serially sectioned specimens and immunohistochemical analysis can be used to detect tumor cells and increase the accuracy of metastasis detection (3 , 7) . These methods have been proven to be important for determining patient prognosis by large-scale prospective studies (3 , 4) . Recently reported RT-PCR methods show promise as diagnostic tools because they are highly sensitive, require relatively less time to perform, and are cost effective (8) . To our knowledge, however, little is known about the clinical outcome of patients with AxLN micrometastases and the clinical significance of AxLN micrometastases, possibly because follow-up periods are currently too short to consider use in a prospective study and because the prognosis of node-negative patients is relatively good.
To circumvent this problem, we used archived formalin-fixed samples from patients whose clinical outcomes were known. Formalin-fixed samples are generally considered to be poor material for molecular biological applications. However, in a previous study (21) , we successfully established a new method for extracting RNA from formalin-fixed samples fairly efficiently for use in RT-PCR analysis. In addition, we devised a method to eliminate the majority of chemical modifications by formaldehyde and demonstrated that the resulting mRNA was suitable for reverse transcription. As a result, formalin-fixed specimens could be used to detect micrometastatic cancer cells in AxLNs by RT-PCR amplification. We chose CEA mRNA as a marker to detect breast cancer cells. CEA has been the preferred molecular marker for detection of micrometastases in lymph nodes in almost all carcinomas (11 , 12 , 14 , 15) . Some evidence suggests that false positives may arise due to several factors, including illegitimate CEA mRNA expression in lymphocytes, the presence of noncancerous epithelial cells in lymph nodes, and the use of too many PCR amplification cycles (20) . In fact, it was difficult to find the proper experimental conditions even in our laboratory. When primers A, B, and C of CEA, which is used commonly in other study groups (11 , 12 , 14 , 15) , were applied to this micrometastasis detection assay, we could not find the proper experimental conditions that resulted in positive results for cancer cells and negative results for normal lymph nodes. We coped with this problem by designing our original CEA fwd primer that provided a 366-bp amplimer in the pair with primer B. In our hands, CEA was not detected in any of the 27 control lymph nodes (Fig. 1⇓ B), whereas CEA was able to be detected in all (100%) of the 40 tumor-involved lymph nodes (Fig. 1⇓ C).
Our study demonstrated that the detection of micrometastases in AxLNs by RT-PCR amplification of CEA mRNA might have clinical value for node-negative patients. We found that micrometastases were associated with a significant reduction in the 5-year and 10-year disease-free survival rates from 93.0% and 87.6%, respectively, in patients without micrometastases to 72.1% and 66.1%, respectively, in patients with micrometastatic disease in AxLNs. A similar result was observed for overall survival. Multivariate analysis revealed that RT-PCR detection of micrometastases represented an independent prognostic factor for both disease-free survival and overall survival. Moreover, the existence of micrometastases was recognized as the most important predictor of outcome in node-negative breast cancer patients (the hazard ratio was 3.992 for relapse and 4.293 for cancer-related death).
In the overall survival analysis, there was no significant difference by 5 years between micrometastasis-negative and micrometastasis-positive populations (96.5% versus 89.6%; Fig. 4⇓ ), whereas the disease-free survival curve of the micrometastasis-positive patients was located midway between the curves for the micrometastasis-negative patients and node-positive patients (Fig. 3)⇓ . Interestingly, the overall survival curve of the micrometastasis-positive patients descends toward that of the node-positive patients during the follow-up period after 5 years. A statistically significant difference in overall survival between micrometastasis-negative and micrometastasis-positive patients was detected at 10 years (93.7% versus 67.8%). Thus, these findings suggest that most of the effect of micrometastases on relapse occurred within the first 5 years after surgery, whereas the effect of micrometastases on cancer-related death occurred 5 years after surgery. In short, tumor cell dissemination occurred comparatively early in the patients with AxLN micrometastases. However, the prognosis of micrometastasis-positive patients after relapse seemed to be better than that of patients with macrometastases detected by routine H&E staining. This might be a result of real differences in biological characteristics, such as tumor growth rate and responsiveness to chemotherapy.
The routine histopathological assessment in our hospital was performed on only one section of the middle of each lymph node. Therefore, this pathological examination of H&E staining might miss macrometastases when the nests of cancer cells were located in the periphery rather than the middle of lymph nodes. In case 57 in Fig. 2⇓ , for example, the amplimer of CEA was very strong, possibly due to a sampling error missing macrometastases or a putative high level of CEA expression. Unfortunately, however, we do not have the means of distinguishing which of the 40 patients with positive RT-PCR results for CEA have positive results due to a sampling error and which have micrometastases in a true sense.
In our study, although there was no statistical significance, the clinical outcome of patients with micrometastases tended to be better than that of node-positive patients (i.e., patients with macrometastases). As shown in Figs. 3⇓ and 4⇓ , the survival curve for patients with micrometastases fell midway between the curve for patients without micrometastases and the curve for patients with histologically positive lymph nodes. From these points of view, almost all of the 40 cases might be considered to have AxLNs with true micrometastases but not macrometastases.
Cote et al. (4) reported a large-scale study on the prognostic impact of immunohistochemically detectable tumor cells in the lymph nodes of breast cancer patients. They compared H&E staining of serial multisections with immunohistochemistry. They concluded that the immunohistochemical examination of AxLNs is a reliable, prognostically valuable, and simple method to detect occult micrometastases, especially in postmenopausal patients. In our study, we experimented mainly with the aim of comparing routine H&E staining with the RT-PCR method. Because we extracted RNA from the whole archived lymph node that had been cut in half and paraffin embedded, it was not possible to compare the RT-PCR method with the immunohistochemical method from the viewpoint of sensitivity and specificity. This comparison is regarded as important and a necessity in clinical application. Although we did not compare the results of immunohistochemistry with those of RT-PCR, we also consider immunohistochemical examination to be a valuable method for better selection of patients and risk assessment in clinical practice, and we agree with the interpretations of Cote et al. (4) . However, as reported by Lockett et al. (8) , RT-PCR methods show promise as diagnostic tools because they are highly sensitive, require relatively less time to perform, and are cost effective in comparison with immunohistochemistry.
We emphasize that it is a simple, useful, and worthwhile method to examine the middle section of lymph node by H&E staining and analyze the remainder by the molecular-based technique that is RT-PCR in the clinical field.
With respect to adjuvant chemo-endocrine therapy, node-negative breast cancer patients had been given mainly oral administration of 5FU or similar anticancer drugs instead of combination chemotherapy such as CMF (cyclophosphamide, methotrexate, and 5FU) and CAF (cyclophosphamide, doxorubicin, and 5FU). The prognosis of patients with micrometastases was worse than that of patients without micrometastases, and, although there was no statistical significance, the clinical outcome of the patients with micrometastases tended to be better than that of node-positive patients. If more effective and adequate adjuvant therapies had been selected for the patients with micrometastases, their clinical prognosis may have improved.
In summary, the RT-PCR amplification method may be most useful for planning combination therapy and for identifying patients that are likely not in need of further treatment. Moreover, to determine risk assessment and treatment of breast cancer patients with precision, we advise the use of this method for analyzing AxLN status in combination with other factors, such as ploidy pattern (29) , c-erbB2 amplification (24) , protease expression (26 , 27) , and detection of disseminated cancer cells in peripheral blood and bone marrow (14 , 16, 17, 18, 19 , 30) .
In conclusion, micrometastasis detection in AxLNs examined by CEA-specific RT-PCR is useful for predicting high risk for relapse in early-stage breast cancer patients and is a powerful complement to routine histopathological analysis in the clinical field. It will be necessary to ascertain whether high-risk patients identified by this method can benefit from adjuvant therapy. Prospective, randomized, and multicenter trials with large numbers of patients are recommended.
Acknowledgments
We are grateful to Dr. Makoto Miyamoto and Dr. Takashi Matsumoto of Osaka University for their valuable advice.
Footnotes
-
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.
-
↵1 Supported in part by Grant 96L00104 from Research for the Future of Japan Society for the Promotion of Science and Grant 08283105 for Scientific Research on Priority Areas from the Ministry of Education, Science, Sports and Culture of Japan.
-
↵2 To whom requests for reprints should be addressed, at Department of Surgery and Clinical Oncology, Graduate School of Medicine, Osaka University, 2-2-E2, Yamada-oka, Suita, Osaka 565-0871, Japan. Phone: 81-6-6879-3251; Fax: 81-6-6879-3259; E-mail: nmasuda{at}surg2.med.osaka-u.ac.jp
-
↵3 The abbreviations used are: AxLN, axillary lymph node; RT-PCR, reverse transcription-PCR; CEA, carcinoembryonic antigen; PBGD, porphobilinogen deaminase; CI, confidence interval; 5FU, 5-fluorouracil.
- Received May 31, 2000.
- Revision received August 28, 2000.
- Accepted August 29, 2000.
References
Volume 6, Issue 11
