| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Article |
1 General Surgery, Department of Surgery, Tri-Service General Hospital, National Defense Medical Center, No. 325, Sec. 2, Cheng-Kung Road, Nei-Hu 114, Taipei, Taiwan, ROC
2 Department of Radiology, Tri-Service General Hospital, National Defense Medical Center, No. 325, Sec. 2, Cheng-Kung Road, Nei-Hu 114, Taipei, Taiwan, ROC
3 Department of Pathology, Tri-Service General Hospital, National Defense Medical Center, No. 325, Sec. 2, Cheng-Kung Road, Nei-Hu 114, Taipei, Taiwan, ROC
4 Department of Medicine, Tri-Service General Hospital, National Defense Medical Center, No. 325, Sec. 2, Cheng-Kung Road, Nei-Hu 114, Taipei, Taiwan, ROC
Correspondence: Address correspondence and reprint requests to: Jyh-Cherng Yu, MD; E-mail: doc20106{at}ndmctsgh.edu.tw
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Methods: We used a blue dye to identify the SNs in 127 patients with T3 locally advanced breast cancer initially treated with neoadjuvant chemotherapy. After SN biopsy, we used intraoperative ultrasound to explore the non-SN region for additional lymph nodes, followed by at least level II axillary dissection. All the SNs were evaluated histologically and immunohistochemically with anticytokeratin antibodies. All the non-SNs were examined by routine histology.
Results: SNs were identified in 116 (91.3%) of 127 procedures. SN metastases were found in 64 cases (55.2%). Subsequent axillary dissection revealed tumor involvement of non-SNs in 40 (62.5%) of 64 cases. SN biopsy results had a sensitivity of 92.8%, a specificity of 100%, and a false-negative rate of 9.6%. Furthermore, intraoperative ultrasound detected suspicious malignant nodes in the non-SN level in 39 out of 40 cases, and detected 5 cases with non-SN metastases that had false-negative SN mapping. This technique decreased the false-negative rate of SN mapping from 9.6% to only 1.39% for these cases.
Conclusions: SN biopsy results combined with intraoperative ultrasonography can accurately assess the non-SN status and help breast surgeons to decide whether subsequent axillary dissection is warranted after SN biopsy has been performed.
Key Words: Sentinel node • SN biopsy • Non-SN metastasis • Intraoperative ultrasound
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
PATIENTS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The neoadjuvant chemotherapy was a doxorubicin-based regimen of two to four cycles. No preoperative radiotherapy was provided. Patients with an allergy to blue dyes, who had solid organ metastases at initial diagnosis, who initially had palpable axillary lymph nodes, or whose primary tumor did not shrink at least 50% after neoadjuvant chemotherapy were excluded. All patients had tissue diagnosis by needle or incisional biopsy, and staging examinations included tumor markers such as carcinoembryonic antigen, cancer antigen 153, chest films, and abdominal ultrasound. Whole-body bone scan, computed tomographic scan, magnetic resonance imaging, or positron emission tomography were used to exclude the possibility of occult lesions in organs before the neoadjuvant chemotherapy. Breast operations, including mastectomy or breast-preserving surgery, were performed after two to four cycles of induction chemotherapy. Patients’ data, including age, tumor size, operation type, estrogen and progesterone receptor and Erb-2/neu expression, and axillary lymph node status were recorded and analyzed. All patients were enrolled onto this study after their written informed consent had been obtained. The institutional review board of the Tri-Service General Hospital approved this study.
Technique for SN Identification
Under general anesthesia, 5 mL of blue dye was injected into the periresidual breast tumor parenchyma and areolar. The interval between dye injection and axillary incision was approximately 10 minutes. A transverse incision of approximately 2 cm was made just below the hair-bearing region of the axilla. Blunt dissection was performed until a dye-filled lymphatic tract or blue-stained node was identified. The dye-filled tract was traced to the first blue node. We followed the dye-filled lymphatic tract proximally to the tail of the breast to ensure the blue-stained node was the SN. After the SN was identified, it was excised carefully. The SN or SNs were then examined by imprint cytology and frozen section by a cytologist and pathologist. Intraoperative ultrasound was then used to carefully explore the entire axilla before pathological examination of the SNs. We then performed either a modified radical mastectomy or a partial mastectomy and axillary lymph node dissection. Axillary lymph node dissection included nodes at levels I and II, Rotter’s node, and occasionally level III nodes. The pectoralis minor muscle was left intact. If the SN was not identified, we performed standard axillary node dissection. All operations were performed by one surgeon (J.C.Y.).
Touch Imprint Analysis
Before submitting each SN for histopathological examination, we prepared air-dried touch imprints for cytological examination. Depending on size, the node was cut longitudinally in halves or thirds immediately after excision. We gently touched each cut surface onto a microscope slide. Four slides were prepared, each with two impressions, and two were stained with Diff-Quick. The remaining two were stained with immunohistochemical (IHC) stain for cytokeratin (CK; Dako, Denmark) and for epithelial membrane antigen (Dako), with horseradish peroxidase as an indicator. Positive staining for CK or epithelial membrane antigen in morphologically atypical cells in the SN was used to indicate a metastasis. Results of such cytological studies were readily available within 1 hour after tissue procurement. The result was comparable with the subsequent hematoxylin and eosin (HE)-stained, paraffin-embedded sections.
Intraoperative Ultrasonography
Intraoperative sonography was performed by a breast surgeon (J.C.Y.) and double-checked by an experienced breast radiologist (G.C.H.) with 20 years’ experience in breast sonography. All patients in this study had their mammograms and sonograms reviewed before surgery. In the preoperative period, all patients were given a detailed verbal explanation by the surgeons and radiologists so they could clearly understand the surgical and sonographic procedures as well as the reason for its use without prolonging the operation time and its benefits in giving early true-positive diagnosis of metastatic lymph nodes. All patients received axillary sonographic examinations with a high-frequency transducer at a bandwidth of 5 to 12 MHz (Logiq 500; GE Medical System, Bothell, WA). After SN biopsy was performed, the preliminary results of histological examination by pathologists were available. The ipsilateral axilla was examined by overlapping scans in the radial and antiradial planes and of the SN biopsy wound to ensure a complete search of the level I, II, III, and Rotter’s lymph nodes. The procedure time was approximately 5 to 15 minutes (average, 10 minutes).
Sonographic features for suspicious malignant lymph nodes were defined by size, morphology, and internal echogenicity (including the cortex and medulla of the visualized node), as well as previously published sonographic criteria.9–12 Malignant lymphadenopathy was based on the presence of eccentric or uneven cortical thickening (which may be associated with outwardly bulging to perinodal fat or inward indentation to the medulla) and/or disproportionate enlargement of a lymph node that forms an abnormal "rounding" shape.
The sonomorphologic features of the short-axis diameter and the thickness of the peripheral cortex of the lymph node were evaluated and compared with pathological findings. The short-axis diameter of the lymph node was measured and classified into three groups according to different degrees of enlargement: (1) short-axis diameter 
10 mm; (2) short-axis diameter 8 mm but <10 mm; and (3) short-axis diameter <8 mm. If we used the smaller cutoff threshold of 8 mm as the presence of abnormal enlargement of malignant lymph nodes, groups 1 and 2 were considered as abnormal positive findings. Only group 3 had negative sonographic findings. If we used the larger cutoff threshold of 10 mm as indicating the presence of abnormal enlargement of malignant lymph nodes, only group 1 was considered as an abnormal positive finding; groups 2 and 3 were defined as negative findings. Statistical analyses were performed separately according to the different cutoff thresholds for the short-axis diameter (Tables 1
and 2).
|
|
3 mm in width; (2) eccentric cortical thickening of 2 mm but <3 mm in width; and (3) the cortical width of <2 mm. If we used the smaller cortical thickness of 2 mm as the threshold to define a lymph node as a suspicious malignant lymph node, groups 1 and 2 were considered as positive findings, and only group 3 was considered as a negative finding. If we used the thicker cortical thickness of 3 mm as the threshold to define a lymph node as a suspicious malignant lymph node, group 1 was de-fined as a positive finding and groups 2 and 3 were considered as negative findings. Statistical analyses were then performed separately according to the different cutoff thresholds for the cortical thickness (Table 1
and 2).
Pathology
Lymph nodes were marked as SNs and non-SNs. The SNs were cut in halves or thirds longitudinally depending on their size. A touch imprint of each portion of the SN was made immediately. The node was then processed and embedded in paraffin. Sections at a minimum of five to eight additional levels for each SN were examined after staining with HE and for CK. Every non-SN >2 mm in diameter was grossly sectioned, and all nodal tissues were submitted for paraffin embedding and routine histological examination by HE staining. IHC staining for CK was performed as indicated. An experienced histopathologist examined at least two sections of each non-SN. In all cases, the size of the metastasis in the SN was measured with an ocular micrometer. A micrometastasis was defined as a tumor deposit of 
2 mm. Metastases >2 mm in diameter were defined as macrometastases. If multiple tumor deposits were present in the SN, the sum of the tumor deposits was used to classify them collectively as a macro- or micrometastasis. Patients with more than one SN positive for tumor cells were grouped according to the largest metastasis.
Statistical Analysis
All data were reviewed and analyzed by the biostatistical unit at our hospital. We used Student’s t-test for univariate continuous variables and the 
2 test for categorical variables, as well as Fisher’s exact test whenever the 2 expected value of at least one cell was <5 (SAS version 9.13; SAS Institute, Cary, NC, USA).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Among the 116 study patients, there were 60 patients with nodal size of 8 mm in the non-SN level nodes, whereas the others had nodal size of <8 mm. Fifty-one of these 60 patients had a nodal size of 10 mm. Sixty-four of the 116 patients had a cortical thickness of 2 mm in the non-SN nodes, and 43 of these 64 patients had a cortical thickness of 3 mm. In 44 patients, the nodal size was 8 mm and the cortical thickness 2 mm in the non-SN level nodes. They all had metastatic disease in the non-SN nodes according to pathologic examination; however, in 5 of these 44 cases, the SNs did not have a tumor. There was one case of SN metastasis, but this was only a micrometastatic lesion in the non-SN; however, this node could not be detected by intraoperative ultrasound.
The validity of the use of either nodal size (8 mm vs. <8 mm, or 10 mm vs. <10 mm) or cortical thickness in the non-SN level nodes (2 mm vs. <2 mm, or 3 mm vs. <3 mm) to predict the presence of non-SN metastases was highly statistically significant (P < .0001, respectively; Table 1). However, use of the smaller cutoff thresholds—nodal size 8 mm and cortical thickness 2 mm—to predict the non-SN status had a higher sensitivity, negative predictive rate, and false-positive rate but a lower specificity, positive predictive rate, and false-negative rate than did the larger cutoff threshold. Therefore, use of the larger cutoff thresholds to predict non-SN metastasis may increase the specificity and positive predictive rate, and decrease the sensitivity and negative predictive rate, but give a higher false-negative rate.
In addition, nodal size 8 mm vs. <8 mm predicted non-SN metastasis with a higher specificity and positive predictive rate but a lower false-positive rate than for cortical thickness of 2 mm vs. <2 mm (77.46% vs. 71.83%, 73.33% vs. 68.75%, and 26.67% vs. 31.25%, respectively); however, the sensitivity and the negative predictive rates of both parameters were almost the same (97.78% vs. 97.78%, and 98.21% vs. 98.08%, respectively). We thus suggest use of the smaller cutoff thresholds of nodal size 8 mm and/or cortical thickness 2 mm to predict non-SN metastases because of the higher sensitivity, specificity, and positive and negative predictive rates, and the lower false-positive and false-negative rates (Table 3).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
SN biopsy can avoid axillary dissection, especially in small breast cancers.1–4 However, the value of SN biopsy for patients with large breast cancers or with locally advanced breast cancer after neoadjuvant chemotherapy is still controversial because of complicated lymphatic drainage and fibrotic changes in the lymphatic drainage after chemotherapy.7,8 It is clear that some of these patients had no disease in the axilla or downstage axillary nodes after chemotherapy and would not benefit from axillary lymph node dissection. Knowing how to identify patients without the need for sequential axillary dissection among this group of patients is critical for breast surgeons, because of the low detection and high false-negative rates of SN biopsy findings for these cases, especially in those patients with a good response to neoadjuvant chemotherapy.7,8 SN biopsy is still a reasonable procedure for them. Some of the inconsistent results found in previous studies are due to the small numbers of patients evaluated, so that even one failure greatly contributes to a high false-negative rate.7,8,17,18 Many of these studies were reported during early experiences with SN biopsy when the technique was not standardized, and identification and false-negative rates have improved with experience. With appropriate experience and institutional validation, SN biopsy after neoadjuvant chemotherapy is a reasonable approach. With extensive experience with SN biopsy, our data show we still had a relatively low but acceptable SN identification rate (91.3%), but a higher false-negative rate (9.6%) compared with our cases with small breast cancer.13 We used SN biopsy findings to select 71 (61.2%) of 116 cases that had no disease in the axilla (47 cases) or only localized disease in the SN (24 cases). SN biopsy without sequential axillary dissection is valuable and adequate for these patients.
SN biopsy findings have a higher false-negative rate for patients receiving neoadjuvant chemotherapy.7,8,17,18 Theoretically, preoperative chemotherapy can result in fibrosis of the lymphatic channels, breast parenchyma, and fat tissue, leading to difficulty in identification and a higher false-negative rate. Another plausible problem is that a selective complete response in the SNs but not in the non-SNs could result in higher false-negative rates. SN identification rates in these patients range from 84% to 98%, with an accuracy of 77% to 100% and a false-negative rate of 0% to 33%.7,8,17–19 In this study, we had five patients with false-negative mapping who had meta-static disease at the non-SN level. On retrospective review, we found that four of them only had fibrotic tissue in the SN. This may be due to tumor removal from the SN by the chemotherapy, but not from the non-SN or fibrotic fat tissue, resulting in false-negative SN biopsy findings. However, careful assessment of the remaining axillary region after SN biopsy was necessary and is a reasonable approach in patients who have received neoadjuvant chemotherapy. At this time, no definite method can solve this problem.
Intraoperative ultrasonography is a reasonable alternative technique for evaluating the entire non-SN level while histological examination of an SN by frozen section or imprint cytology with IHC staining is being performed.9–12,20,21 This technique is easily applicable and noninvasive, but depends on experience. We found that nodal size and eccentric cortical thickness predict the possibility of non-SN metastasis when the smaller cutoff thresholds of nodal size 8 mm and/or cortical thickness 2 mm are used as a positive finding in the non-SN level (Table 3). With removal of suspicious nodes in the non-SN level nodes under ultrasound guidance, we could detect 44 (97.78%) of 45 cases exhibiting residual disease in the non-SN level nodes, with only one patient with micrometastatic disease in the non-SN nodes. A lymph node exhibiting only micrometastatic disease may not be detected by ultrasound or even by routine histological examination. As in this study, removal of suspicious nodes 8 mm will result in decreasing the false-negative rate from 9.6% by SN biopsy samples alone to 1.79%. If we removed suspicious nodes of 8 mm and of cortical thickness 2 mm, the false-negative rate would decrease to only 1.39% (Table 3), which is acceptable and similar to SN biopsy for small breast cancers. This technique may increase axillary morbidity slightly but will help breast surgeons to decrease the possibility of false-negative SN biopsy findings. This will help breast surgeons ensure adequate local control in the axillary level.
Furthermore, in SNs without metastasis and where intraoperative ultrasound found no suspicious nodes in the non-SN level, none of our cases had residual disease in the non-SN level; sequential axillary dissection is unnecessary in this instance. Even when the SN biopsy findings showed metastasis but the intra-operative ultrasound could not locate a suspicious node in the non-SN level, only one case had residual micrometastatic foci in a non-SN node. These cases can be treated by postoperative chemotherapy and radiotherapy, which is a reason to omit the necessity of sequential axillary dissection; however, further study and evaluation are necessary.
We conclude that SN biopsy results in findings with a lower identification rate and higher false-negative rate for patients with locally advanced breast cancer who receive neoadjuvant chemotherapy. The SN identification rate can be increased by improving the technique and increasing experience. The identification rate of SN biopsy samples for these patients is 91.3%, which is similar to those series in which SN biopsy is performed before systemic therapy. SN biopsy results can accurately predict residual disease in the axilla in this group, but a 9.6% false-negative rate is higher than for SN biopsy findings for small breast cancer in our hospital.13 However, intraoperative ultrasound can help breast surgeons to evaluate the entire non-SN level during the processing of SN histology, including frozen sections or imprint cytology with IHC staining. This technique can give immediate information about suspicious nodes in the non-SN level, including nodal size or eccentric cortical thickness. Breast surgeons can remove suspicious nodes of 8 mm and/or cortical thickness 2 mm. The false-negative rate of only 1.39% is similar to those series in which the SN biopsy was performed for small breast cancer. Furthermore, in patients with locally advanced breast cancer after neoadjuvant chemotherapy, no SN metastasis, and no suspicious node in the non-SN level by intraoperative ultrasound, sequential axillary dissection is unnecessary.
Received for publication June 9, 2006. Accepted for publication June 14, 2006.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  J.-M. Classe, V. Bordes, L. Campion, H. Mignotte, F. Dravet, J. Leveque, C. Sagan, P. F. Dupre, G. Body, and S. Giard Sentinel Lymph Node Biopsy After Neoadjuvant Chemotherapy for Advanced Breast Cancer: Results of Ganglion Sentinelle et Chimiotherapie Neoadjuvante, a French Prospective Multicentric Study J. Clin. Oncol., February 10, 2009; 27(5): 726 - 732. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
