Presented in part at the 2011 Annual Meeting of the American Society for Clinical Oncology; June 3-7, 2011; Chicago, IL.
The objective of this study was to confirm, by means of a multicenter study conducted in Japan, the reliability and usefulness of the one-step nucleic acid amplification (OSNA) assay in routine clinical use for sentinel lymph node biopsy (SLNB) of breast cancer patients.
Patients with Tis-T2N0M0 breast cancer who underwent SLNB before systemic chemotherapy comprised the study cohort. A whole sentinel lymph node (SLN) was examined intraoperatively with the OSNA assay except for a 1-mm-thick, central slice of the lymph node, which underwent pathologic examination after the operation. For patients who underwent axillary dissection, non-SLNs were examined with routine pathologic examination.
In total, 417 SLNBs from 413 patients were analyzed. SLN metastases were detected with greater sensitivity by the OSNA assay than by pathologic examination (22.5% vs 15.8%; P < .001), as expected from the difference in size of the specimens examined. Patients who had SLN metastases assessed with the OSNA assay proved to harbor non-SLN metastases with an overall risk ratio of 33.7%. The risk of non-SLN metastasis was significantly lower for patients who had positive SLNs assessed as OSNA+ than for those who had SLNs assessed as OSNA++ (17.6% vs 44%; P = .012).
Sentinel lymph node biopsy (SLNB) has been a standard procedure for patients with early stage breast cancer.1, 2 However, to date, the method for examining sentinel lymph nodes (SLN) has not been standardized. Hematoxylin and eosin (H&E) staining for multistep sections with or without immunohistochemistry for cytokeratin (CK) generally is recommended,3 although it is not known how many specimens should be examined. To overcome this problem, automated molecular detection systems for lymph node metastases, such as the one-step nucleic acid amplification (OSNA) assay (Sysmex, Kobe, Japan) and the Geneseach breast lymph node (BLN) assay (Veridex, Raritan, NJ) have been developed and are receiving much attention recently.4, 5 Several studies have shown that these new tests can detect lymph node metastases with the same statistically determined accuracy as the conventional pathologic examination,6-15 which indicates that a molecular test may constitute an alternative to pathology. However, in those previous studies, only half the volume of a lymph node was examined with the molecular test, because the remaining half was used for pathologic examination as the standard procedure. Essentially, some results obtained with the 2 methods are discrepant, especially when a lymph node harbors micrometastases. The molecular test originally was supposed to examine a whole lymph node with high sensitivity for detecting cancer deposits and also with much less labor than what is required for a thorough pathologic examination of a great number of sections. Nevertheless, currently, pathology remains the gold standard, and using the molecular tests may generate some anxiety about, for example, technical failure and mechanical trouble. How to use the molecular tests for SLNB in the daily clinical setting is therefore still controversial.
The OSNA assay, a molecular diagnostic system for lymph node metastasis that detects cytokeratin 19 (CK19) messenger RNA (mRNA) of cancer cells, was approved by the Japanese Ministry of Health, Labor and Welfare in June 2008 and has been covered by the Japanese National Health Insurance system since November 2008. In view of these developments, we conducted a multicenter study of the clinical use of the OSNA assay for SLNB, in which most of an SLN was examined with the OSNA assay, and only a central, 1-mm-thick slice of the SLN was preserved as a permanent pathologic section. The reliability and usefulness of the OSNA assay in clinical use and the relation between the OSNA assessment and the risk of non-SLN metastasis are described in this report.
MATERIALS AND METHODS
The objective of this study was to determine the usefulness of the OSNA assay for clinical use in SLNB of breast cancer. The primary endpoint was to examine the superiority of the OSNA assay for detecting metastases in SLN compared with pathologic examination with H&E staining for a single SLN section. The secondary endpoint was to investigate the relation between non-SLN metastasis and the OSNA assessment for CK19 mRNA copy numbers in SLN. SLNs were detected using both radiocolloids and blue dye, radiocolloids only, or blue dye only. Removed SLNs were prepared according to the protocol detailed below and were assessed immediately with the OSNA assay. Patients had axillary lymph node dissection (ALND) recommended when necessary according to the OSNA assessment and/or other clinicopathologic factors. The level of axillary dissection was determined by the surgeon according to the patient's condition and institutional guidelines. Non-SLNs were examined with a routine pathologic examination using H&E staining. Each patient received appropriate postoperative adjuvant therapy and/or radiotherapy based on the clinicopathologic findings and in accordance with guidelines if necessary, and each patient was followed at the treating center.
The study group comprised 11 hospitals, which are the central institutions for breast cancer therapy and research in each area of Japan. The study protocol was approved by the institutional review board of each center.
Patients and Sentinel Lymph Node Biopsy
The enrolment for this study comprised patients with tumor in situ (Tis) through T2, clinically lymph node-negative primary breast cancer who underwent SLNB between August 2009 and December 2010 at 1 of the participating hospitals. Patients who had a preoperative diagnosis of ductal carcinoma in situ (DCIS) were enrolled in the study when a surgeon judged SLNB was needed. Patients who underwent SLNB before receiving preoperative systemic chemotherapy (PSCT) also were eligible for the analysis of sensitivity of the OSNA assay, although those who received chemotherapy or hormone therapy before SLNB were excluded from the study. Men also were excluded. Patients received the necessary information about the study, and only those who gave their consent and underwent SLNB successfully were enrolled.
Preparation of Sentinel Lymph Nodes and the One-Step Nucleic Acid Assay
Preparation of an SLN is shown in Figure 1. Fat tissue surrounding the SLN was trimmed off. A 1-mm-thick slice was then cut out from the longitudinal central part of the SLN, fixed as a permanent section for staining with H&E, and examined postoperatively by a pathologist at one of the hospitals. The remaining part of the lymph node was immediately examined with the OSNA assay by laboratory technicians at the hospitals in the manner described previously.7
An SLN was assessed with the OSNA assay according to the cutoff level of calculated CK19 mRNA copy numbers per microliter determined by Tsujimoto et al, and the results were reported according to the manufacturer's instructions: that is, as negative (<2.5 × 102 copies/μL), + positive (≥2.5 × 102 and <5.0 × 103 copies/μL), ++ positive (≥5.0 × 103 copies/μL), or positive +i (inhibited in the regular sample and ≥2.5 × 102 copies/μL in the diluted sample).4
Sensitivity of the OSNA assay and of pathologic examination for the detection of metastasis was compared and analyzed with the McNemar test. The risk of non-SLN metastases for OSNA-positive patients was calculated with the chi-square test.
In total, 439 patients, including 9 women with bilateral breast cancer, were enrolled in this study. Five of the 9 women with bilateral disease underwent unilateral SLNB, and the remaining 4 women underwent bilateral SLNB, and the biopsy specimens were examined with the OSNA assay. Twenty-one of the originally enrolled patients were excluded from the analysis because of significant violations against the study protocol, including 8 patients who received PSCT before SLNB, 10 patients who were not examined with the OSNA assay, 2 patients whose central sections of the SLN did not undergo pathologic examination as a permanent specimen for H&E staining, and 1 patient who was a man. Two patients who had benign intraductal papilloma confirmed after surgery, 1 who had with a clinical T4 tumor, and 2 who had clinically evident axillary lymph node metastases also were excluded because they did not meet the general criteria for SLNB candidates. Conversely, 2 patients who had T3 tumors that finally were diagnosed as DCIS and T1, invasive cancer were included. The final total enrolment was 413 patients who had 417 SLNBs eligible for analysis.
In total, 775 SLNs were obtained from 417 SLNBs, and the average number of SLNs was 1.86 (1-7 SLNs) per patient. Of those, 762 SLNs (98.3%) were examined successfully with the OSNA assay. In 5 biopsies that had multiple SLNs, >4 excess lymph nodes were assessed by means of pathology (total, 13 SLNs). One hundred and one patients underwent ALND, including 49 patients who underwent level I dissection and 52 patients who underwent level I and II dissections. Of those, 86 patients had positive OSNA assessments, and 15 patients had negative OSNA assessment. Seven OSNA-negative patients underwent delayed ALND based on pathology results after primary surgery. The final axillary status of 11 patients who received PSCT after SLNB was unknown. Patient characteristics are summarized in Table 1.
Of 417 SLNBs, including 11 from patients who received PSCT after SLNB, the OSNA assay identified SLN metastases in 94 biopsies (22.5%), and pathologic examination of a single section identified SLN metastases in 66 biopsies (15.8%) (Table 2). Thus, the OSNA assay detected significantly more metastases than pathologic examination of a single H&E-stained section (P < .001), as expected, because most of each SLN was examined by means of the OSNA.
Table 2. Comparison of the One-Step Nucleic Acid Assay With Pathologya
There were 44 results that were discordant: that is, there were 36 OSNA-positive/pathology-negative (O+/P−) sections and 8 O−/P+ sections (Table 3). In 7 of the O−/P+ patients, only micrometastases were identified in the SLN, and macrometastasis was identified in 1 SLN with a tumor in which further immunohistochemical analysis revealed a low level of CK19 protein expression. Isolated tumor cells were identified in SLNs from 2 of the 36 O+/P− patients, and non-SLN metastases were identified in 7 patients. Therefore, in total, 9 of the O+/P− patients (25%) harbored cancer cells in either SLNs or non-SLNs.
Table 3. Summary of Discordant Cases Between the One-Step Nucleic Acid Assay and Pathology
Of the 86 OSNA-positive biopsies from patients who underwent axillary dissection, 34 were assessed as +, 50 were assessed as ++ and 2 were assessed as +i. In total, 18 of 34 patients with OSNA + results and 22 of 50 patients with OSNA ++ results underwent Level I ALND alone. There was no relation between the level of ALND and OSNA assessment (P = .421). Six patients (17.6%) who had OSNA + results and 22 patients (44%) who had OSNA ++ results had non-SLN metastases (Table 4). The risk of non-SLN metastasis was significantly lower for patients who had positive SLNs assessed as OSNA + versus those who had SLNs assessed as OSNA ++ (P = .012).
Table 4. The Risk of Nonsentinel Lymph Node Metastasis in One-Step Nucleic Acid Assay-Positive Patients Who Undergo Axillary Dissection
Positive OSNA results were scored as + (≥2.5 × 102 copies/μL and <5.0 × 103 copies/μL); ++ (≥5.0 × 103 copies/μL), or i+ (inhibited in the regular sample and ≥2.5 × 102 copies/μL in the diluted sample).
It has been demonstrated that the OSNA assay has the same capability for detecting lymph node metastasis as conventional pathologic examination.6-11 However, only a few studies have presented data regarding clinical use of the assay.10 In our study, most of each SLN was examined intraoperatively by using the OSNA assay, and the decision whether to perform axillary dissection was based in principle on the assay results. Only a single 1-mm-thick, central slice of the lymph node was used for pathologic examination. Therefore, we expected that the OSNA assay would have higher sensitivity for SLN metastasis than pathologic examination, and the results were as expected. There were some discordant cases in our study, which also was expected, because this is inevitable when 2 modalities are used to examine different parts of the lymph nodes. Of the 44 discordant results, 8 were OSNA-negative, in which postoperative pathologic examination identified metastasis. In the 7 patients who had micrometastasis identified, discordance may have occurred because of the uneven allocation of minuscule metastases in an SLN. However, in 1 patient with macrometastasis, low expression of the CK19 protein in the main tumor was confirmed as the result of further immunohistochemical examination performed by a pathologist at the concerned hospital. The incidence of low expression of the CK19 protein in breast cancer was reported previously as 1.6%.16 However, the expression of protein and mRNA can be expected to be different, especially between the main tumor and metastatic sites. In fact, the reported incidence of discordance between OSNA and pathology caused by low expression of CK19 mRNA is very low, from 0.2% to 0.5% of examined lymph nodes in previous studies7, 9 and 0.1% of examined lymph nodes and 0.2% of all patients in our study. Lack of CK19 expression is associated significantly with the triple-negative (estrogen receptor negative, progesterone receptor negative, and human epidermal growth factor receptor 2 [Her2] negative) phenotype.17 Some adjuvant chemotherapy is likely to be used for such patients based on other factors, although SLN is assessed as negative by the OSNA because of low expression of CK19. Therefore, this false-negative aspect may have only a minimal effect on patients' clinical prognosis, because pathologic examination of 1 preserved slice of the lymph node can negate such an effect.
Conversely, there were 36 O+/P− discordant cases, including 2 with isolated tumor cells in the SLNs that were assessed by pathology. Of the 34 patients who underwent axillary dissection, non-SLN metastases were identified in 7 patients. The OSNA assay had made an accurate assessment of these patients. It is interesting to note that there were 6 patients with DCIS among these O+/P− cases. Microinvasion was suspected in a core-needle biopsy specimen from 1 patient. Two patients had wide-spread DCIS that measured >6 cm, and another had multiple lesions. The remaining 2 patients had high-grade DCIS. Ansari et al reported in their review that the estimated incidence of SLN metastases in patients who had a definitive diagnosis of DCIS alone was 3.7%.18 Thus, the OSNA assay can detect metastases with high sensitivity even in tumors diagnosed pathologically as DCIS, and such findings may result in an upgrade of the clinical stage of such tumors.
The clinical significance of micrometastases in SLNs is controversial. de Boer et al reviewed 58 studies concerning this issue and concluded that the presence of metastases measuring ≤2 mm in greatest dimension in axillary lymph nodes detected on single-section examination was associated with poorer disease-free and overall survival.19 Reed et al reported the results from a prospective study indicating a significant association between SLN micrometastasis and distant recurrence.20 Conversely, Hansen et al reported that micrometastatic tumor deposits in SLNs, pN0(i+) or pN1mi, detected by H&E staining or immunohistochemistry do not have clinical significance for disease-free or overall survival.21 In the study, >90% of patients with micrometastases received adjuvant systemic therapy, although only 66% of those without metastases received such therapy. Weaver et al reported that occult metastases were detected by means of further examination using immunohistochemistry in 15.9% of patients with pathologically negative SLNs who were enrolled in The National Surgical Adjuvant Breast and Bowel Project trial B-32.22 That report revealed significant differences in overall survival, disease-free survival, and distant-disease-free survival between patients with and without occult metastases. Nevertheless, the authors concluded that the data did not indicate a clinical benefit of additional evaluation, including immunohistochemical analysis, of initially negative SLNs, because the magnitude of the difference in outcome was so small. However, tumor size, endocrine therapy, and radiation therapy were independent prognostic factors of death or distant disease in the patients studied, which may have reduced the difference in prognostic outcomes. Results from the Micrometastases and Isolated Tumor Cells (MIRROR) study also indicated that both isolated tumor cells and micrometastases in axillary lymph nodes were associated significantly with a worse prognosis for patients who have favorable, early stage breast cancer who did not receive adjuvant systemic therapy.23 That report indicated that adjuvant systemic therapy could improve the 5-year disease-free survival of such patients with micrometastases with a gain in 5-year disease-free survival of nearly 10%. Thus, a precise initial evaluation of SLN metastasis is important for the accurate assessment of clinical stage and the appropriate selection of adjuvant treatment for each patient. The OSNA assay, which can evaluate the volume of metastases in SLNs semiquantitatively, is a useful tool for an accurate assessment of clinical stage of breast cancer patients.
The original objective of SLNB was to avoid axillary dissection and reduce postoperative adverse morbidity for patients without axillary lymph node metastasis. Giuliano et al indicated that axillary dissection may not be needed even for patients with 1 or 2 positive SLNs who have undergone breast-conserving surgery with postoperative whole-breast radiation and systemic adjuvant therapy, as indicated by the results from the American College of Surgeons Oncology Group Z0011 study.24 However, it remains unknown whether axillary dissection also may be omitted for patients who have ≥3 positive SLNs and for those who have positive SLNs and undergo total mastectomy. Therefore, accurate clinical staging and selection of patients who do not need axillary dissection remain the goals of SLNB. Previous reports indicated that approximately 60% of patients with positive SLN did not have any non-SLN metastasis25, 26 and that such patients basically did not need axillary dissection. In our study, 66.3% of patients who had SLN metastases identified by the OSNA assay did not have non-SLN metastases. Conversely, 17.6% of patients with OSNA+ results and 44% of patients with OSNA++ results had non-SLN metastasis, which are ratios similar to those previously reported (range, 13%-22% for patients with SLN micrometastasis; 45%-79% for patients with SLN macrometastasis27), and such patients may have suffered axillary recurrence because they underwent total mastectomy and did not undergo axillary dissection. Thus, how to select patients with a high or low risk of non-SLN metastasis remains an important issue for the use of SLNB. The tumor volume in SLN is considered a significant factor for the prediction of non-SLN metastasis.25, 27, 28 It is easy to assess tumor volume in SLNs semiquantitatively with the OSNA assay, and this ease of operation constitutes a major advantage over conventional pathologic examination. Data from larger numbers of patients are expected to determine the appropriate cutoff level of the OSNA assay for the selection of patients who do not need additional axillary dissection.
In conclusion, the OSNA assay is considered reliable in the clinical setting for the routine intraoperative examination of SLN and is useful because it can be performed easily by a nonpathologist. However, further studies to obtain long-term follow-up data for greater numbers of patients are needed to confirm the clinical significance, especially the prognostic impact, of results of the OSNA assay of SLNB for breast cancer.
We acknowledge the authorship of all members of the Japanese One-Step Nucleic Acid Amplification (OSNA) Study Group for this work, including Niigata Cancer Center Hospital (Nobuaki Sato, MD; Keiichi Homma, MD), National Hospital Organization Shikoku Cancer Center (Daisuke Takabatake, MD; Rieko Nishimura, MD), Osaka University Graduate School of Medicine (Shinzaburo Noguchi, MD; Yasuhiro Tamaki, MD), Osaka Police Hospital (Masahiko Tsujimoto, MD; Katsuhide Yoshidome, MD), National Cancer Center Hospital (Hitoshi Tsuda, MD; Takayuki Kinoshita, MD), National Hospital Organization Kure Medical Center and Chugoku Cancer Center (Hironori Kato, MD; Kiyomi Taniyama, MD), Tokyo Women's Medical University (Takako Kamio, MD), Showa University School of Medicine (Seigo Nakamura, MD), St. Marianna University School of Medicine Hospital (Koichiro Tsugawa, MD), Tokai University School of Medicine (Yutaka Tokuda, MD), Nihon University School of Medicine (Shinobu Masuda, MD), The Cancer Institute of the Japanese Foundation for Cancer Research (Futoshi Akiyama, MD), The Cancer Institute Hospital of the Japanese Foundation for Cancer Research (Takuji Iwase, MD), and Surugadai Nihon University Hospital (Motoo Yamagata, MD).
This study was supported by a grant from the Nakatani Foundation of Electronic Measuring Technology Advancement.
CONFLICT OF INTEREST DISCLOSURES
S. Noguchi has received a research grant from Sysmex Corporation for another study