Fine-needle aspiration (FNA) cytology of axillary lymph nodes is a simple, minimally invasive technique that can be used to improve preoperative determination of the status of the axillary lymph nodes in patients with breast cancer, thereby serving as a tool with which to triage patients for sentinel versus full lymph node dissection procedures. The aim of the current study was to determine the sensitivity and specificity of FNA cytology to detect metastatic breast carcinoma in axillary lymph nodes.
A total of 115 FNAs of axillary lymph nodes of breast cancer patients with histologic follow-up (subsequent sentinel or full lymph node dissection) were included in the current study. The specificity and sensitivity, as well as the positive and negative predictive values, were calculated.
The positive and negative predictive values of FNA cytology of axillary lymph nodes for metastatic breast carcinoma were 1.00 and 0.60, respectively. The overall sensitivity of axillary lymph node FNA in all the cases studied was 65% and the specificity was 100%. The sensitivity of FNA was lower in the sentinel lymph node group than in the full lymph node dissection group (16% vs 88%, respectively), which was believed to be attributable to the small size of the metastatic foci in the sentinel lymph node group (median, 0.25 cm). All false-negative FNAs, with the exception of 1 case, were believed to be the result of sampling error. There was no ‘true’ false-positive FNA case in the current study.
The detection of metastases in axillary lymph nodes is a very important prognostic factor in breast cancer because of its impact on subsequent management and overall survival. Different methods are used to detect metastases in axillary lymph nodes preoperatively.1–19 Although some studies have used ultrasonographic imaging techniques alone,2–6, 12, 17, 18 others have used fine-needle aspiration (FNA) cytology with or without ultrasound guidance to improve the detection of metastases.1, 7–13, 15, 19
FNA cytology of axillary lymph nodes is a simple, minimally invasive technique that can be utilized as a tool with which to select patients for sentinel versus full lymph node dissection procedures in cases of breast carcinoma. Indeed, by objectively documenting the presence of metastatic carcinoma by preoperative FNA cytology, many patients can be directed toward full lymph node dissection, thereby preventing unnecessary sentinel lymph node procedures and reducing the involved costs and time.
The objective of the current study was to determine the accuracy of FNA in detecting metastatic breast carcinoma in axillary lymph nodes.
MATERIALS AND METHODS
From the database of the Department of Pathology of the McGill University Health Center (MUHC), we retrieved all FNAs of axillary lymph nodes (N = 314) performed with or without ultrasound guidance and received at the MUHC Cytopathology Laboratory over a 3-year period between January 1, 2004 and December 31, 2006. From those, we identified 153 cases (48.7%) as coming from breast cancer patients by the presence of recent, concurrent, or subsequent breast histologic material in the same database. Of the latter 153 axillary lymph node FNAs, 25 (16.3%) were unsatisfactory and 13 (8.4%) had no subsequent relevant histologic follow-up. Therefore, 115 FNAs (75.1%) remained with histologic follow-up, either sentinel or full axillary lymph node dissection, and these were included in the current study. Of the total 115 axillary FNAs included in this study, 49 were performed under ultrasound guidance, whereas 66 were performed with palpation only.
For each of the 115 cases, the following data were gathered: the original FNA cytologic diagnosis for the axillary lymph node, the histologic diagnosis of the primary malignancy (site of origin, size of the tumor, and type and grade of breast carcinoma), the type of diagnostic procedure for the axillary lymph node (full lymph node dissection vs sentinel lymph node dissection) with the number of positive lymph nodes of the total number of resected lymph nodes, and the maximum size of the largest metastasis.
All the slides from the cases for which there was a discrepancy between the cytology and the histology (false-positive and false-negative cases) were reviewed. The specificity, sensitivity, positive predictive values (PPVs), and negative predictive values (NPVs) were calculated.
An axillary sentinel lymph node dissection was performed in 53 cases (46%), whereas a full axillary lymph node dissection was performed in 62 cases (53.9%).
The histologic characteristics of the primary breast carcinomas and of the axillary lymph nodes related to the 115 cases included in this study were as follows. The primary breast tumors included 90 cases of ductal carcinoma (78.2%), 13 cases of lobular carcinoma (11.3%), and 12 cases with mixed ductal and lobular features (10.4%). According to the Bloom-Scarff-Richardson grading for breast carcinoma, there were 50 cases of grade 3 of 3 (43.1%), 51 cases of grade 2 of 3 (44.8%), and 14 cases of grade 1 of 3 (12.0%). The median size of the primary tumor was 1.3 cm. The median size of the metastatic foci in the axillary lymph nodes (with full and sentinel lymph node dissection cases grouped together) was 1.0 cm (range, 0.09–3.0 cm); however, when broken down by the type of axillary lymph node dissection performed, the median size of the metastatic foci in the sentinel lymph node group was 0.25 cm and was 1.5 cm in the full lymph node dissection group.
The breakdown of the FNA diagnoses in the 115 axillary lymph node FNAs was as follows: 49 (42.6%) demonstrated metastatic breast carcinoma, and all but 1 were confirmed histologically. The 1 FNA that was not confirmed histologically (ie, all 28 axillary lymph nodes resected were negative) was originally believed to be a false-positive result; however, it became clear that it actually represented a ‘false’ false-positive result. Indeed, review of the FNA smears and the FNA cell block demonstrated unequivocally malignant cells (Fig. 1, top and middle). The discrepancy in that case was explained by the finding that, after the performance of the axillary lymph node FNA, the patient was treated with neoadjuvant chemotherapy before the full lymph node dissection was performed. Indeed, in our institution neoadjuvant chemotherapy is typically administered to patients with stage III breast cancer (determined according to the Tumor, Node, Metastasis staging system), which was the case in this particular patient. Histologic section of 1 of the lymph nodes demonstrated extensive scarring, presumably corresponding to a ‘burnt-out’ focus of metastatic carcinoma (Fig. 1, bottom).
Of the 66 FNAs diagnosed as benign lymph nodes, 40 (60.6%) demonstrated no evidence of metastasis in the resected lymph nodes (ie, true-negative), whereas 26 FNAs (39.4%) demonstrated metastases in the resected lymph nodes (ie, false-negative cases). Of the 26 false-negative FNAs, 20 (76.9%) were in the sentinel lymph node group compared with only 6 cases (23%) in the full lymph node dissection group. All false-negative FNAs, with the exception of 1 case, were the result of sampling error because no malignant cells were identified, even on careful retrospective review of the FNA material. Indeed, only 1 FNA was determined to be a false-negative result from a diagnostic standpoint (ie, diagnostic miss), because occasional malignant cells, originally missed, were found on review of the FNA slides.
The overall sensitivity and specificity of the axillary lymph node FNAs in all the cases (including both sentinel and full lymph node dissection cases) were 65% and 100%, respectively. When considering only the 49 ultrasound-guided FNAs, the sensitivity was 62.5% and the specificity was 100%, whereas the sensitivity and specificity were 67.4% and 100%, respectively, in the 66 FNAs performed by palpation only.
It is interesting to note that the sensitivity of FNA was lower in the sentinel lymph node group compared with the full lymph node dissection group (16% vs 88%, respectively), a finding that was attributable to the small size of the metastatic foci in the sentinel lymph node group (median of 0.25 cm vs a median of 1.5 cm in the full lymph node dissection group). The majority of the false-negative cases (81.5%) was associated with lymph node metastasis measuring <1 cm (median, 0.25 cm); in fact, the metastatic foci were so small in 8 cases of sentinel lymph node dissection that they were detected only after examination of deeper levels and/or immunohistochemical staining of the histologic slides.
The overall PPV of the axillary lymph node FNAs in all the cases (including both sentinel and full lymph node dissection cases) was 1.00 and the NPV was 0.60 at the 95% confidence interval. When considering only the 49 ultrasound-guided FNAs, the PPV was 100% and the NPV was 58.6%, whereas the PPV and the NPV were 100% and 62.2%, respectively, in the 66 FNAs performed by palpation only.
Preoperative knowledge of the axillary lymph node status for metastasis is invaluable in patients with breast carcinoma because it affects the surgeon's selection of the type of axillary lymph node dissection performed, specifically sentinel versus full lymph node dissection. Currently, in a patient known to have at least 1 positive axillary lymph node preoperatively, a full axillary lymph node dissection would be performed directly, without prior sentinel lymph node dissection.7, 11 Avoiding an unnecessary sentinel lymph node procedure is beneficial because it is both time-consuming and cost-consuming.7 Indeed, the sentinel lymph node dissection procedure, which aims at identifying the status of the first (sentinel) lymph node draining the primary tumor within the axilla,7 is complex and lengthy because it requires a multidisciplinary team approach for the preoperative injection and detection of a blue dye and of a radiocolloid agent. Preoperative confirmation of metastatic breast carcinoma in an axillary lymph node (either by FNA cytology or core needle biopsy) constitutes an absolute contraindication to sentinel lymph node dissection.20
Because determination of the axillary lymph node status greatly affects the decision-making for the most appropriate lymph node dissection, it is critical to know the accuracy of the different techniques used in the preoperative assessment of the axillary lymph node status. To our knowledge no known biologic tumor marker to date can predict axillary lymph node metastases preoperatively. Different methods have been used, including palpation, ultrasonography alone, and ultrasonography combined with FNA cytology or core needle biopsy. Clinical examination alone has been shown to be inaccurate, with a sensitivity ranging from 45.4%2 to 68%.5 For clinical examination, de Freitas et al.5 quoted a sensitivity of 68%, a specificity of 68%, a PPV of 82%, and an NPV of 50% for a total accuracy of 68%. Approximately 15% to 60% of patients with clinically nonpalpable lymph nodes have been shown to harbor lymph node metastases on follow-up.19
Similar to clinical examination, the accuracy of evaluation by ultrasound only also has been shown to be variable. Indeed, despite relying on ‘suspicious’ imaging features (such as a lymph node size >10 mm, the absence of a fatty hilum, a hypoechoic internal echo, a circular shape, and cortical thickening1), the reported sensitivity of the evaluation of axillary lymph node status by ultrasound alone has ranged from 35%19 to 82%,10 whereas its specificity has ranged from 73%10 to 97.9%.3 For ultrasound alone, de Freitas et al.5 quoted a PPV of 92%, an NPV of 49%, and a total accuracy of 67%. Therefore, it appears that ultrasonography is not sufficiently accurate to replace histologic examination of the lymph nodes for the reliable determination of metastatic status.9
To improve the accuracy of the preoperative assessment of axillary lymph nodes, others have used FNA and found it to be simple, moderately accurate, and minimally invasive, and a good triage tool for the management of patients.1, 7–13, 15, 19 Indeed, it has been shown that combining axillary ultrasonography with FNA cytology of abnormal lymph nodes can significantly reduce the number of unnecessary sentinel lymph node procedures performed by 8% to 40%, with a concomitant reduction in healthcare costs of up to 20%.7–10, 15, 19
The data from the current study confirm that FNA cytology is a useful procedure with which to detect axillary lymph node metastases from breast carcinoma. Indeed, we found that the overall sensitivity and specificity for axillary lymph node FNAs in all our cases were 65% and 100%, respectively, whereas the PPV was 1.00 and the NPV was 0.60. When considering only the ultrasound-guided FNAs in our series, the sensitivity was 62.5%, the specificity was 100%, the PPV was 100%, and the NPV was 58.6%, whereas when considering the FNAs performed with palpation only, the sensitivity was 67.4%, the specificity was 100%, the PPV was 100%, and the NPV was 62.2%. These results compare favorably with those in the literature, in which the sensitivity of ultrasound-guided FNAs for lymph nodes ranges from 36% to 86.4%, the specificity ranges from 95.7% to 100%, the PPV ranges from 92% to 100%, and the NPV ranges from 67%8 to 70%.7–9, 11, 15
The data from the current study demonstrated that the sensitivity of lymph node FNA was lower in the sentinel lymph node group than in the full lymph node dissection group (ie, 16% vs 88%, respectively). All the false-negative FNAs, with the exception of 1, were because of sampling error. The difference in the sensitivity of the FNA in the sentinel lymph node group compared with the full lymph node dissection group relates to the small size of the metastatic foci in the sentinel lymph node group, decreasing the likelihood of sampling the metastatic foci. Other studies have also shown that the most common cause of false-negative cases is inadequate sampling,11 either because of the small size of the metastases,1, 3, 7 the low number of lymph nodes positive for metastasis,1, 3, 19 or failure to observe the lymph nodes during examination of the axilla by ultrasound.1, 3, 18, 19
Reported causes of false-positive diagnoses in the literature are due mostly to cytologic misinterpretation of cells11 or inadequate sampling of the lymph node dissection.19 In the current study, there were no false-positive FNA cases; indeed, although 1 FNA specimen diagnosed as metastatic breast carcinoma was not confirmed histologically (all of the 28 lymph nodes resected were found to be negative), review of the FNA smears and cell block material demonstrated unequivocally malignant cells. The discrepancy in that case is explained by the fact the patient was treated with neoadjuvant chemotherapy before the full lymph node dissection was performed and scarring was identified on the histologic section of 1 of the lymph nodes, presumably corresponding to a ‘burnt-out’ focus of metastatic carcinoma. Although some studies also report an absence of false-positive cases,13, 14 others have reported false-positive rates of 1.4% to 1.6%.7, 11
FNA is a sensitive and very specific method with which to detect metastases in the axillary lymph nodes of patients with breast cancer. The most common cause of false-negative FNAs is sampling error because of very small metastatic foci, especially in the setting of sentinel lymph node dissection. Because of its excellent positive predictive value, full axillary lymph node dissection can be safely planned after a positive FNA result.
We thank Dr. R. P. Michel for editorial assistance with the article