Image-guided fine-needle aspiration (FNA) studies of axillary lymph nodes (LN) to evaluate breast carcinoma have shown high specificity but variable sensitivity. The purposes of this study were to evaluate the performance of axillary LN FNA depending on clinicoradiologic findings and to document how treatment varied according to FNA results.
The study cohort consisted of consecutive axillary LN FNA cases over a 4-year period, in which subsequent treatment was known. Clinicoradiologic assessment was classified as “low suspicion” or “high suspicion” and cytopathologic findings as “positive,” “negative,” or “indeterminate”. The test performance for each, using surgical pathology outcome as the “gold standard,” was calculated. The impact of axillary LN FNA on subsequent management decisions was analyzed.
Of the 163 cases, axillary FNA was positive in 94 of 163 (58%), negative in 55 of 163 (34%), and atypical/nondiagnostic in 14 of 163 (8%). A clinicoradiologic assessment of “high suspicion” had a positive predictive value (PPV) of 88%, whereas a “low suspicion” assessment had a negative predictive value (NPV) of only 68%. In contrast, the PPV and NPV of axillary LN FNA were 98.7% and 81.8%, respectively. Whereas all of the FNA-nonpositive cases were managed surgically, surgery was deferred in 26 of 94 of the FNA-positive cases, including 11 cases of neoadjuvant treatment. Most of the remaining (65 of 68) FNA-positive patients were spared sentinel lymph node biopsy.
A key determinant in the treatment of the breast carcinoma patient is the presence of disease in the locoregional axillary lymph nodes. As clinical palpation for the detection of metastatic disease is suboptimal, with respect to both sensitivity and specificity,1 a variety of imaging modalities have been investigated for detecting lymph node metastases.2, 3 Ultrasound (US) imaging of lymph nodes has become a widely favored imaging technique.3-6 Comparison of the staging of axillary lymph nodes using US alone is difficult because of a lack of standardization; however, studies suggest a wide variation in sensitivity and less than optimal specificity.3-12 This variation is likely attributable to the finding that a number of subjective criteria (size, morphology, color, pattern) are used to assess a lymph node.13
The addition of fine-needle aspiration biopsy (FNA) to the US examination of axillary lymph nodes improves the specificity of the examination and can confirm extramammary metastases.4, 13 Although the specificity of a positive US-guided FNA is high, there remains a wide variation in the reported sensitivity of US-guided FNA of axillary lymph nodes.3, 5, 14, 15 In addition, the role of US-guided FNA in cases of nonpalpable—but sonographically indeterminate/suspicious—axillary nodes remains controversial.16, 17 The role of this procedure in planning surgery and reducing the number of sentinel lymph node biopsy (SLNB) procedures has been well described.3, 5, 14, 15, 17-19 Currently, multiple treatment options are offered to women with breast carcinoma even before they undergo surgery.20 The role of US-guided axillary FNA in this era of multimodality therapy has not been well described.4
This study had 2 objectives. The first purpose of this study was to evaluate and compare the performance of image-guided FNA of axillary lymph nodes (mostly US) in staging breast carcinoma to the performance of combined clinical palpation and radiologic examination. Second, the study sought to identify whether image-guided FNA of axillary lymph nodes is only a determinant of surgical management or whether it also has a role in the overall management of the woman with breast carcinoma.
MATERIALS AND METHODS
Clinical and pathologic data were accessed and recorded in accordance with the study protocol approved by the Mount Sinai Hospital Research Ethics Board. Consecutive cases in which image-guided axillary lymph node FNA had been performed were identified during the period from January 1, 2005 to December 31, 2008. Cases included in the study were those FNAs performed for the presurgical evaluation of a primary breast carcinoma (as determined by subsequent surgical pathology). Excluded were cases without documentation of the subsequent (surgical and/or adjuvant) treatment of the breast carcinoma or cases without availability of the original FNA slides for review.
The medical record was reviewed to determine, for each case, the level of suspicion for axillary lymph node involvement. Whether the lymph node was palpable or nonpalpable and whether the lymph node showed suspicious features on imaging by US (n = 161) or MRI (n = 2) were recorded, as determined from the reported radiologic assessment. Radiologically suspicious features noted in the reports included hypoechoicity, thickened or irregular cortex, and replacement of hilar tissue. The clinicoradiologic finding of either a palpable lymph node or suspicious features on imaging was designated “high suspicion”. The combined clinicoradiologic finding of a nonpalpable lymph node and benign-appearing or indeterminate features reported on imaging was designated “low suspicion”. (Cases in which nodes were neither palpable nor showed any imaging abnormality were typically not sampled by FNA.)
The cytopathologic findings for each case were re-reviewed, in conjunction with the original cytopathology report, to ensure correct classification within 3 cytodiagnostic categories. No discrepancies with the original reports were identified. “Positive cases” were either clearly malignant or highly suspicious for malignancy. Cases in which the specimen was interpretable and showed no evidence of malignancy were labeled as “negative”. “Indeterminate cases” were those cases showing some cellular atypia that was inconclusive for malignancy or cases in which the specimen was paucicellular. Subsequent surgical pathology findings and/or management were documented, with particular notation of lymph node sampling approach(es), any SLNB findings, and any administration of neoadjuvant chemotherapy.
Data were analyzed with 2 separate goals in mind: 1) comparing cytopathologic FNA results with both clinicoradiologic findings and final surgical pathology findings and 2) identifying clinical management approaches after negative, indeterminate, or positive/suspicious FNA results. The performance of both clinicoradiologic and cytopathologic FNA were compared with the final surgical pathologic finding in the axillary nodes (“gold standard”). Only lymph nodes with macrometastases or micrometastases (>0.2 mm) were considered positive in this analysis. Sensitivity, specificity, PPV, and NPV were calculated for clinicoradiologic assessment and for FNA.
Because these parameters were defined only for positive and negative test results, cases in which FNA findings were indeterminate were analyzed separately by comparing them to final surgical pathologic findings. Definitive surgical staging was not performed in all cases; in a subset of cases (n = 15; Table 1), only adjuvant or palliative treatment was undertaken. These cases were excluded from comparisons between testing modalities but not from the overall analysis of clinical follow-up. Finally, cases (n = 2) were identified in which a “discordance” between a positive lymph node FNA and a negative surgical pathologic specimen were attributable to a complete neoadjuvant-treatment response. Although the FNA result presumptively demonstrates pretreatment axillary disease, uncertainty exists as to whether the result was a “true positive” or “false positive”. Therefore, these were omitted from analysis, in accordance with prior approaches to examining isolated postneoadjuvant cases within a case series.16, 21
Table 1. Axillary Lymph Node Fine-Needle Aspiration Biopsy for Primary Breast Carcinoma Assessment: Cytologic, Pathologic, and Clinicoradiologic Characteristics of 163 Cases
Abbreviation: DCIS, ductal carcinoma in situ.
Excludes 15 cases in which no further surgical management was undertaken.
Excludes 2 of 11 post-neoadjuvant cases demonstrating complete pathologic response. (See Methods for “false false-positive FNA results”.)
Result of cytologic evaluation, N=163
Indeterminate, atypical or nondiagnostic
Result of final surgical pathologic assessment of primary breast carcinoma
Grade of undetermined significance (postneoadjuvant)
Result of final surgical pathologic lymph node assessment, n=146b
Degree of clinical and radiologic suspicion, N=163
High, includes 34 cases with palpable lymphadenopathy
Clinical management subsequent to lymph node FNA was identified in all cases, including those undergoing primary neoadjuvant chemotherapy, adjuvant therapy without surgical staging, and palliation only. Statistical differences in distribution between these groups were evaluated by using the chi-square test.
Cytopathologic and Surgical Pathologic Findings of Study
One hundred sixty-three cases met the study criteria of biopsy-proven breast carcinoma with image-guided axillary lymph node FNA sampling and available clinical history. Of the 163 cases, axillary FNA was positive in 94 of 163 (58%), negative in 55 of 163 (34%), and atypical/nondiagnostic in 14 of 163 (8%).
Fifteen of the 163 cases received systemic or palliative treatment alone without surgery. Histopathologic characteristics of the primary breast carcinomas excised are summarized in Table 1. There was no invasive carcinoma (ductal carcinoma in situ only) in 4 cases evaluated, and the remaining 11 cases were postneoadjuvant resections. The latter group was excluded from evaluation of tumor grade. Although postneoadjuvant carcinomas can be graded by using the Nottingham system, it is unclear whether grade is of comparable significance to non-neoadjuvant cases.
Clinical and Diagnostic Imaging Assessment of Lymph Nodes
Before either US or magnetic resonance imaging (MRI) assessment, 34 of 163 (21%) had documented evidence of palpable lymphadenopathy. Subsequent evaluation showed that 33 of these 34 (or 97%) cases with palpable lymph nodes demonstrated suspicious features on imaging as did an additional 67 of the remaining 129 (or 52%) cases with nonpalpable lymph nodes. Among cases with palpable lymphadenopathy in this setting, 30 of 34 (88%) were FNA-positive. In total, 101 of 163 (62%) cases had either palpable lymph nodes or suspicious lymph node(s) on imaging and were, thus, classified as “high clinicoradiologic suspicion”. In 83 of these 101 (82%) cases, the FNA was positive. The remaining 18 cases were either negative or indeterminate. Only 11 of the 62 (18%) cases in which both the clinical and imaging findings were negative, that is “low clinicoradiologic suspicion,” were FNA positive (P < .0001). The remaining 51 cases of the 62 were negative or indeterminate.
The sensitivity and specificity of clinicoradiologic assessment compared with the final surgical pathology of the axillary lymph nodes is shown in Table 2. As described in Methods, not all of the 163 cases were included in this analysis—15 cases remained surgically unstaged and 2 cases were omitted because of uncertainty surrounding complete response to neoadjuvant treatment (n = 146). The finding of either a palpable lymph node and/or suspicious lymph node (“high suspicion”) on imaging had a PPV of 88%, whereas the absence of these clinical and radiologic features (“low suspicion”) had a NPV of only 68%.
Table 2. Clinicoradiologic Versus Surgical Pathologic Assessment of Axillary Lymph Nodes (N=146)
Includes only cases in which fine-needle aspiration (FNA) was performed for lymph node evaluation. Clinical selection criteria are expected to result in variable sensitivities/NPV for axillary FNA. This table does not include 2 cases showing complete pathologic response to neoadjuvant treatment. (See Methods.)
The sensitivity and specificity of axillary lymph node FNA in 132 cases compared with final pathology is shown in Table 3. As described in Methods, 14 cases—in which the FNA result was indeterminate—are analyzed separately below. A positive lymph node FNA had a PPV of 98.7%. In the single “false positive” case, illustrated in Figure 1, the FNA produced a cellular specimen with marked cytologic atypia warranting a malignant classification. Sonographically, this lesion was an 8-mm nodule that could represent either a lymph node or a satellite nodule. The surgical pathologic evaluation showed that the FNA actually sampled an area of periaxillary ductal carcinoma in situ in a node-negative patient. Similarly, a negative lymph node FNA had a NPV of 81.8%. Follow-up of the 14 cases with an indeterminate axillary lymph node FNA revealed that an equal number were lymph node positive (n = 7) as lymph node negative (n = 7). 2
Table 3. FNA Cytopathologic Findings Versus Surgical Pathologic Assessment of Axillary Lymph Nodes (N=132)a
Does not include 14 indeterminate FNA cases or 2 cases showing complete pathologic response to neoadjuvant treatment (see Text).
Axillary FNA positive
Axillary FNA negative
Cytopathologic Impact and Subsequent Management
There were a total of 94 FNA-positive cases, Sixty-eight women subsequently underwent primary surgery, whereas in 26 cases, surgery was deferred for either neoadjuvant treatment or not performed at all in view of metastatic disease or severe comorbidity. The clinical management of a positive FNA differed significantly compared with a negative or indeterminate FNA result. Only patients with positive axillary lymph node FNAs were offered a deferral of surgery for neoadjuvant or palliative treatment.
Sixty-five of the 68 (96%) patients with positive FNAs in the surgical-management group underwent full axillary lymph node dissection without SLNB. In the remaining 3 cases, the SLNB confirmed lymph node positivity, and a subsequent axillary dissection was performed. In contrast, only 15 of 69 (22%, P < .0001) of those with negative/indeterminate FNAs underwent full axillary lymph node dissection. Of the remaining 54 cases, 11 of 54 (20%) had positive sentinel lymph nodes requiring further axillary dissection; 43 of 54 were considered lymph node-negative after SLNB only. Two patients had micrometastases (between 0.2 mm and 2 mm); both had negative FNAs.
Conversely, among the FNA-indeterminate cases, 11 of 14 (79%) proceeded to axillary lymph node dissection. Based on unclear clinical considerations, 4 of 55 (7%) FNA-negative cases were also managed with axillary lymph node dissection; however, none of these demonstrated nodal involvement on pathologic assessment. In this cohort, all patients with negative or indeterminate FNAs were offered definitive surgical management.
The results of this study further support the accuracy of axillary lymph node FNA with respect to final histopathologic lymph node status. As noted by others,14, 21 FNA provided higher sensitivity and specificity compared with clinical or radiologic assessment alone. We note that false positives are very uncommon, and we have further identified a novel cause of a false positive: the presence of high-grade in situ carcinoma within axillary breast tissue. Although this is, in fact, a “true positive” with respect to the presence of disease, for the purposes of lymph node status determination, it did lead to an overestimation of tumor stage. In our series, the proportion of lymph node-positive cases (64%) was significantly higher than the average rate of nodal metastasis (30%).22 This presumably reflects the ability of clinicoradiologic criteria to identify cases with a higher probability of nodal involvement.
In the era of the core-needle biopsy for breast cancer evaluation, axillary lymph node fine-needle aspiration remains an important confirmatory test for locally advanced and/or inoperable carcinoma.23, 24 Previous studies have demonstrated the accuracy of axillary FNA combined with radiologic evaluation to detect lymph node metastases.17, 21 Even compared with core-needle biopsy, axillary lymph node FNA used in this setting has superior sensitivity, lower cost, and equivalent specificity.13 Axillary lymph node FNA has, therefore, been proposed as a routine practice for initial nodal assessment of breast cancer patients, both for initial staging and for triage with respect to SLNB.15, 17
The current era is also one of expanding use of neoadjuvant treatment. The survival benefit of neoadjuvant chemotherapy in inflammatory and locally advanced breast carcinoma has been established since the mid-1980s.25 However, neoadjuvant treatment can also be used in a patient who would require postoperative chemotherapy even in cases of operable disease. The benefits include reducing surgical morbidity and providing oncologists with prognostic and in vivo chemosensitivity data.25 Therefore, preoperative lymph node status has become a valuable parameter for both surgical and medical oncologic triage, regardless of the size and configuration of the primary tumor.
In this study, we have documented the range of treatment modalities undertaken within a full-service breast oncologic center following axillary lymph node FNA during a 4-year period. Our results demonstrate that the findings of an axillary FNA play an important role in triage for both surgical and systemic treatments. Several authors have noted that a positive axillary lymph node FNA should obviate the need for a SLNB.15, 17, 21 Although the vast majority (65 of 68) of patients triaged to surgery through a positive axillary FNA did undergo a full axillary dissection, a decision was made to proceed to SLNB in 3 cases. The reason for SLNB was unclear in these cases. This finding likely reflects the variability of clinical considerations in patient management. Ultimately, it must be recognized that the surgeon is in the best position to decide on the most appropriate procedure.
The high PPV of axillary FNA is well accepted by the radiological, surgical, and medical oncologists at our center. The sustained importance of axillary FNA in the neoadjuvant era is underscored by the significant number of cases in which a patient was triaged to systemic and/or palliative treatment without further (preoperative) lymph node sampling. At our center, 11 of 163 patients presenting for axillary FNA during the study period underwent neoadjuvant therapy, with 2 of 11 demonstrating a complete pathologic response. This overall response rate is similar to those reported within large neoadjuvant trials.26
In summary, axillary lymph node FNA is highly sensitive and specific for lymph node involvement by primary breast carcinoma when compared with clinical and radiologic examination alone. Axillary FNA plays a role in sparing SNLB and in triaging advanced cases for systemic and neoadjuvant therapy. Our results support the routine use of axillary FNA for the breast cancer patient presenting for preoperative evaluation.