- Email: [email protected]
Role of axillary ultrasound examination in the selection of breast cancer patients for sentinel node biopsy
The American Journal of Surgery 193 (2007) 16 –20
Clinical surgery–International
Role of axillary ultrasound examination in the selection of breast cancer patients for sentinel node biopsy Jacopo Nori, M.D.a, Ermanno Vanzi, M.D.a, Massimo Bazzocchi, M.D.b, Filippo Nori Bufalini, M.D.a, Vito Distante, M.D.c, Francesco Branconi, M.D.d, Tommaso Susini, M.D., Ph.D.d,* a
Diagnostic Senology Unit, Azienda Ospedaliera-Universitaria Careggi, Florence, Italy b Department of Radiology, University of Udine, Udine, Italy c Department of General Surgery, University of Florence, Florence, Italy d Department of Gynecology, Perinatology, and Human Reproductive Medicine, University of Florence, Viale Morgagni, 85 - 50134, Firenze, Italy Manuscript received December 7, 2005; revised manuscript February 20, 2006
Abstract Background: Sentinel node biopsy (SNB) is a time-consuming procedure that can be avoided in presence of axillary metastases. The aim of this study was to assess the accuracy of ultrasound scan (US) in the prediction of axillary nodes status in patients scheduled for SNB. Methods: Axillary US was performed and when feasible, a core biopsy of suspicious nodes was taken. The nodal status as assessed by US and/or core biopsy was compared with final histology. Results: Of the 132 patients enrolled, 31 (23.5%) had suspicious axillary nodes according to US; 19 (61.3%) were true positive, whereas 12 cases (38.7%) were not. In 14 of 31 suspicious cases an US-guided core-biopsy was taken, which in 11 of 14 cases (78.5%) confirmed the neoplastic involvement. Overall, core biopsy of the nodes correctly predicted the final histology in 13 of 14 cases (92.8%). Conclusions: The US of axillary nodes, possibly associated with core biopsy, improved the preoperative evaluation of breast cancer patients scheduled for SNB. © 2007 Excerpta Medica Inc. All rights reserved. Keywords: Breast cancer; Axillary lymph nodes; Ultrasound; Core biopsy; Sentinel node biopsy
The presence of lymph node metastasis, in the absence of distant recurrences, is the single most important prognostic factor in breast cancer [1]. Currently, as many as 70% of diagnosed breast cancers do not exhibit lymph node metastases, making total axillary dissection often unjustified. For this reason, in recent years sentinel node biopsy (SNB) has met with general acceptance, as an attempt to minimize overtreatment of healthy axillae [2–5]. On the other hand, the widespread use of SNB has increased the cost of the overall procedure because of the need of preoperative lymphoscintigraphy, intraoperative ␥–probe use, and prolonged operative time, while waiting for frozen section results. In addition, an expensive and time-consuming intensive pathological study of the sentinel node is required, with up to 20% false negatives on frozen sections that reveal micrometastasis at final histopatologic evaluation, thus requiring a second surgical procedure for axillary node dissection. * Corresponding author. Tel.: ⫹39.055.411879; fax: ⫹39.055.434330. E-mail address: [email protected]
Therefore, accurate selection of patients with high probability of having uninvolved axillary nodes would be crucial in the optimization of the SNB procedure. The clinical examination of the axilla is unfortunately unreliable for correct preoperative staging of the lymph nodes [6]. The ultrasound scan (US) study of the axilla may provide a significant improvement in the preoperative assessment of lymph node status. In expert hands, axillary US almost always allows identification of the lymph nodes, assessment of even the smallest nodes, and in-depth evaluation of node morphology [7–11]. If US criteria to recognize metastatic nodes can be consistently identified, axillary US, possibly in association with US-guided core biopsy, may significantly impact on the choice of the surgical procedure [12]. Obviously, patients with core biopsy– documented positive nodes or US highly suspicious nodes might undergo straight to total axillary node dissection, saving financial resources and time. The aim of our study was to evaluate the preoperative diagnostic accuracy of axillary US, associated when necessary
0002-9610/07/$ – see front matter © 2007 Excerpta Medica Inc. All rights reserved. doi:10.1016/j.amjsurg.2006.02.021
J. Nori et al. / The American Journal of Surgery 193 (2007) 16 –20
17
with US-guided core biopsy, in breast cancer patients scheduled for quadrantectomy combined with SNB. Methods Between July 2001 and December 2003 we performed preoperative axillary US with detailed study of the lymph nodes in 147 women scheduled for conservative breast surgery for whom SNB had been planned. Fifteen patients were excluded from the study: 6 because they were found to have benign breast disease and 9 because they failed to meet the sonographic requirement of at least 4 visible axillary nodes. The US examinations were performed by a single radiologist (J.N.), using a Technos MP unit (Esaote, Genova, Italy), with a high-frequency probe (10 to 13 MHz) and axial and longitudinal scans. To better characterize lymph node status in the patients with doubtful or suspicious US findings, US-guided core biopsy with a semiautomated biopsy gun and Tru-cut needles (Precisa, Hospital Service, Rome, Italy) was performed under local anaesthesia (5 mL subcutaneous Carbocaine, AstraZeneca, Milan, Italy). We used 14- or 16-G needles, with a 23-mm throw. Three samples per node were taken through a single cutaneous approach, from the cortical region, site of possible neoplastic infiltration, avoiding the highly vascular hilum region. In some cases, color-Doppler US was used during the biopsy procedure to avoid dissecting any large vessels. The biopsy samples were fixed in formalin and sent to the pathology laboratory. At the end of the procedure the biopsy site was compressed for 5 minutes before applying sterile tape and dry ice. The biopsy was not performed if the lesion was in close proximity to the large vessels owing to the risk of bleeding. In the first years of the study, also some patients in whom US-guided core biopsy was technically feasible actually were not submitted to the procedure, because our policy towards percutaneous biopsy of the nodes was less “aggressive.” With increasing experience, we are currently able to perform US-guided core biopsy of suspicious axillary nodes in approximately 80% of the cases. The patients’ ages ranged between 28 and 88 years (mean 56.4 years). They all had histologically or cytologically proven breast carcinoma with diameters ranging from 4 to 28 mm (mean diameter 12 mm) and clinically nonpalpable axillary lymph nodes. Histologic type was as follows: ductal carcinoma-in-situ, 3 cases; infiltrating ductal carcinoma, 83 cases; infiltrating lobular carcinoma, 30 cases; and infiltrating ductal and lobular carcinoma, 16 cases. Ultrasound criteria for axillary node evaluation The study consisted in evaluating the presence of at least 4 lymph nodes, which were assessed for size, shape, visibility and morphology of the hilum, vascularity, and focal cortical thickening. In particular, concerning shape, 2 parameters were considered: the ratio between longitudinal diameter and transverse diameter (L/T ratio), which is ⱖ2 in normal lymph nodes, and the ratio between hilum diameter (H) and longitudinal diameter (L) of the lymph node (H/L ratio), normal lymph nodes being considered those with a hilum diameter ⱖ50% of the overall longitudinal diameter (H/L ratio ⱖ50%). Metastatic nodes tend to have globular
Fig. 1. Schematic representation of lymph node morphology. (A) The ratio between the total longitudinal axis (L) and the longitudinal diameter at the hilum (H) is ⱖ50% (normal). (B) The H/L ratio is ⬍50%, a condition more often associated with malignant infiltration of the node.
morphology (L/T ratio ⫽ 1) and compression or disappearance of the hilum (H/L ratio ⬍50%) (Fig. 1). The sonographic criteria considered suspicious for malignancy included: (1) globular morphology and/or H/L ratio less than 50%; (2) disappearance of hilum fat hyperechogenicity; and (3) eccentric focal thickening and denting of the cortex. Patients with a core biopsy– documented lymph node metastasis proceeded directly to total axillary dissection and breast-conserving surgery. All other patients underwent primary surgical treatment including SNB, and when a positive sentinel node was found, total axillary dissection was performed. The final histologic results concerning lymph nodal involvement were compared with the preoperative axillary node status as assessed by US or by US-guided core biopsy. Statistical analysis The results were analyzed using descriptive statistical methods. Sensitivity, specificity, overall accuracy, and negative predictive values were calculated by comparing the results of US and core biopsy with the final histologic nodal findings. Results US findings versus final histology Axillary US was performed in 132 breast cancer patients in whom at least 4 lymph nodes were visualized. Metastases in 1 or more nodes were documented in 42 patients (31.8%) at final histology. Of the 132 patients, 101 (76.5%) had sonographically normal or nonsuspicious lymph nodes according to the previously mentioned US criteria. The definitive histologic examination of these 101 cases confirmed the negative results in 78 true-negative cases (77.2%), whereas in the remaining 23 false-negative cases (22.8%) the histologic examination demonstrated: in 20 cases (86.9%) the presence of micrometastasis in the sentinel lymph node only; in 1 case (4.3%) the presence of metastasis in two lymph nodes; in 1 case (4.3%) the presence of metastasis in four lymph nodes; and in 1 case (4.3%) the presence of metastasis in eight lymph nodes. These 23 patients underwent surgery for primary breast lesions, which, at the histologic examination of the surgical specimen, presented diameters between 7.0 and 18 mm (pT1) (mean diameter 13.5 mm). The remaining 31 patients had sonographically suspicious lymph nodes (Tables 1 and 2), in particular, 26 cases (83.8%) showed hilum alterations: partially visible hilum in 21 cases (80.7%); nonvisible hilum in 5 cases (19.3%); 6
18
J. Nori et al. / The American Journal of Surgery 193 (2007) 16 –20
Table 1 Distribution of true positives and false positives according to hilar morphology and cortical morphology among 31 cases with suspicious axillary ultrasound 31 cases sonographically doubtful or suspicious
Hilar morphology
Cortical morphology
Normal
Partially visible
Not visible
Normal
Focal thickening
Diffuse thickening
19 true positives 12 false positives
0 5
14 7
5 0
10 10
5 1
4 1
cases (19.3%) showed focal cortical thickening and 5 cases (16.1%) diffuse cortical thickening; 19 cases (51.6%) showed H/L ratio less than 50% (Fig. 1). There was agreement between the sonographic findings and the definitive histology in 19 of these 31 patients (61.3% true positive), but no agreement in the remaining 12 cases (38.7% false positive) (Tables 1 and 2). The definitive histologic examination of the 19 true-positive cases demonstrated: in 10 cases (52.6%) positivity of the sentinel node only; in 3 cases (15.8%) positivity of 2 lymph nodes; in 2 cases (10.5%) positivity of 4 lymph nodes; and in 4 cases (21.0%) positivity of 8, 9, 15, and 24 nodes, respectively. In the 12 false-positive cases (38.7%), in which there was disagreement between suspicious US findings and negative histology, the sonographic findings that had raised suspicion were as follows: hilar morphology (partially visible in 7 cases); cortical morphology (focal cortical thickening in 1 case, diffuse cortical thickening in 1 case); and H/L ratio (⬍50% in 3 cases and equal to 50% in 9 cases, but associated with the above-mentioned hilar and cortical morphology alteration). According to these numbers, sensitivity and specificity of axillary US alone in the detection of metastatic nodes were 45.2% and 86.6% respectively, and positive predictive value (PPV) and negative predictive value (NPV) were 61.3% and 77.2%, respectively. Overall accuracy was 73.5%. Core biopsy of suspicious nodes Of the 31 patients with sonographically suspicious lymph nodes, 14 were submitted to US-guided histologic core biopsy with 2 to 3 samplings per node so as to gain further information on lymph node status. In 7 cases, at the beginning of our experience, we did not propose the procedure. The remaining 10 patients did not undergo preoperative percutaneous biopsy because of unfavorable node location (too deep or too close to the large axillary vessels, 6 cases), or patient’s refusal of the procedure (4 cases). By taking these precautions we had no cases of post-biopsy complications. Table 2 Distribution of true positives and false positives according to the ratio between hilar diameter and overall longitudinal diameter of the node (H/L ratio) among 31 cases with suspicious axillary ultrasound 31 cases sonographically doubtful or suspicious
H/L ratio Less than 50%
Equal to 50%
Greater than 50%
19 true positives 12 false positives
16 3
3 9
0 0
The percutaneous node biopsies showed the following results: 11 lymph nodes with neoplastic infiltration; and 3 lymph nodes free of neoplastic lesions. Definitive histology on the surgical specimen confirmed the biopsy findings in all of the 11 positive cases (100% true positives); among the 3 negative cases at the preoperative examination, there was agreement in 2 cases (66.6% true negatives) and disagreement in 1 case (33.3% false negatives) that final histology showed to be involved. Sensitivity, specificity, PPV, and NPV were 91.6%, 100%, 100%, and 66.6%, respectively. Overall accuracy was 92.8%. Comments The sonographic study of the axilla may be used as a non-invasive and fast method for assessing lymph node status in the preoperative staging of breast cancer patients and in the subsequent planning of local and systemic treatment. In metastatic lymph nodes, the process mainly involves the entire subcortical or cortical region, which, being invaded by the neoplastic tissue, becomes dishomogeneously hypertrophic deforming or even cancels the hyperechogenicity of the hilum or lymphatic sinusoids [6,7,9]. The sonographic parameters to be assessed in the study of the lymph nodes are: dimensions; lymph node shape; morphology of the hilum and cortex; color-Doppler patterns. The longitudinal diameter of normal lymph nodes varies widely, which means it cannot constitute the only criterion to guide the level of US suspicion. Hence, reactive lymph nodes may have diameters of up to 2.5 cm, while lymph nodes with diameters as small as 5 mm may exhibit micrometastasis. Node morphology is an important criterion: reactive or hyperplastic nodes have a reniform, ovoid shape, with longer longitudinal diameter (L) than transverse diameter (T), whereas neoplastic nodes tend to have a globular appearance, with a L/T ratio of approximately 1. The ratio between these 2 diameters has been defined as the “roundness index” by Sakai et al, who demonstrated the L/T ratio to be greater than 2 in approximately 80% of histologically benign nodes, but less than 2 in malignant nodes [13]. This index also has moderate but not absolute specificity. As regards morphology of the hilum region, it should be recalled that the hilum is normally located centrally with respect to the node and that its level of conspicuity may be suggestive of an alteration that requires further investigation. Even if compressed and smaller than usual, the hilum can be documented in about 40% of neoplastic nodes. The cortical region appears as a peripheral hypoechoic rim of regular homogeneous thickness; thickening, irregularity, or interruption of the cortex is a sign of extracapsular spread and involvement of the perinodal fat, which
J. Nori et al. / The American Journal of Surgery 193 (2007) 16 –20
may be related to metastatic cells penetrating the node via the afferent lymphatic vessels [11,14]. The careful study of the hilum and cortical regions may give rise to a sonographic pattern of doubtful or suspicious nodal status. In the current study, among false-positive cases (12 women) the most misleading parameter was increased cortical thickness and consequently a ratio between overall node size (total longitudinal diameter) and size of the hyperechoic hilar portion (H/L ratio) equal to 50%. In contrast, no patient with lymph node metastases at definitive histology had a normal hilar morphology by US. This confirms that alterations of the hilar region are the most important parameters in establishing a sonographic diagnosis of suspicion or doubt, as already reported by others [9,15,16]. In our study, we considered only morphologic parameters, and in particular nodal dimensions (longitudinal and transverse diameters), hilar and cortical morphology, and H/L ratio (Fig. 1). The information obtained from color-Doppler examination was available in only few cases, but may represent a further improvement of US examination of axillary nodes [10,17]. The H/L ratio was measured in all cases, and this measurement, combined with the morphologic parameters, may, in our opinion, improve specificity. We considered all lymph nodes with H/L ratio less than 50% to be doubtful or suspicious. On the basis of the echo structural morphologic features (morphology of hilum and cortex; H/L ratio) the nodes were classified as benign or suspicious; in addition, sonography also visualized nonpalpable nodes that could not be clinically assessed, thereby increasing diagnostic sensitivity in preoperative staging. Of the parameters described, the H/L ratio showed moderate specificity, especially if combined with the morphologic assessment of the hilar and cortical regions. Of 31 patients with positive sonography in our series, in only 61.3% (19/31 cases) did the sonographic suspicion correspond to a histopathologic alteration in 1 or more lymph nodes (true positives), whereas in 38.7% of cases (12/31 cases) there was no concordance between sonography and histopathology (false positives). The H/L ratio was less than 50%, and therefore a sign of doubtful or suspicious node status, in 84.2% (16/19 cases) of the true-positive cases confirmed by histology, in comparison with 25% (3/12 cases) among the false positives, with doubtful sonographic findings, but negative histology for tumor infiltration. Our findings in this initial case series appear to indicate that this index (H/L ratio ⬍50%) can be added to the sonographic parameters orienting towards possible metastatic involvement of lymph nodes in patients with breast cancer. In 101 women in whom sonography detected normal or at any rate nonsuspicious lymph nodes, sonography was unable to identify the alterations demonstrated by the pathologist in 22.8% of cases (false negatives), whereas there was agreement between sonography and histology in 77.2% of cases (true negatives). Therefore, while axillary US is fairly effective for identifying the true-negative cases, it also introduces a certain number of false negatives. This may be related either to the inability to recognize all nodes at sonography or to the presence of micrometastasis that are too small to produce sonographically detectable echo structural alterations. Indeed, in the current series, most of the
19
sonographic false negatives were related to micrometastasis, ranging from .1 to .5 mm in size. Sonography can also be used to guide the performance of fine-needle aspiration cytology (FNAC) or, in this series, histologic core biopsy on suspicious nodes, further increasing specificity. Several reports have been published on the use of US-guided FNAC on axillary nodes in breast cancer patients, reporting sensitivity values ranging from 86.4% to 94.7%, specificity values from 97.1% to 100%, and diagnostic accuracy values from 85.7% to 100% [14,18 –21]. The false-negative results were ascribed to difficulties in interpreting the slides, difficulties in sampling small-size lymph nodes, or sampling errors. When technically feasible, the patients in our series underwent percutaneous histologic core biopsy of the suspicious axillary node to enable a preoperative histologic diagnosis, which allowed the surgeon to plan directly axillary dissection in positive cases, without having to perform SNB. In 100% of the cases (11/11) with a positive core biopsy, the diagnosis was confirmed by the final histopathologic examination. On the other hand, in 66.6% (2/3) of cases both the core biopsy and the final histopathologic examination excluded metastatic involvement of sonographically suspicious nodes. Core biopsy yielded only 1 false-negative result, in a case recognized as a metastatic node at definitive histology. Hence, the concordance between core biopsy on sonographically suspicious nodes and final histology was very high (92.8%; 13/14 cases). The only previous report of US-guided core biopsy in the preoperative diagnosis of axillary node metastasis is, to our knowledge, that of Damera et al [12]. In their study, out of 48 nodal core biopsies performed, 26 (54%) were positive for metastatic invasion. Although in the current series the number of core biopsy procedures was lower, we found a higher concordance between US and histology, with 78.5% (11/14) of the biopsies confirming the US suspicion of metastatic invasion. Despite the experience with core biopsy of axillary nodes is still limited, we believe that this approach will receive increasing attention in the near future, for preoperative diagnosis of both primary breast cancer and axillary node metastases. Indeed, the core biopsy technique, in comparison with FNAC, allows a substantial reduction of the inadequate (C1–B1), doubtful (C3–B3), and suspicious (C4 –B4) categories of the pathologic reports [22,23]. The availability of a true histologic diagnosis seems us preferable when planning whether or not to perform total axillary node dissection. Overall, our study confirms that a negative axillary sonography does not allow one to hypothesize about the absence of nodal involvement. This limitation is due to the possible presence of micrometastasis in lymph nodes with an apparently normal echo structure. On the other hand, a positive or suspicious axillary sonography was confirmed at final histologic examination in only 61.3% of the cases, with as many as 38.7% of cases being false positive. The most important sonographic alteration was the disappearance of hilar echogenicity, observed in 100% of true metastatic nodes (Table 1), in agreement with the findings of Sato et al [16]. Conversely, hilar denting or irregularities, as well as dimensional criteria, proved to be poorly specific. In conclusion, the sonographic assessment of axillary nodes, possibly in combination with core biopsy, may improve
20
J. Nori et al. / The American Journal of Surgery 193 (2007) 16 –20
the preoperative evaluation of breast cancer patients scheduled for conservative surgery of the breast and SNB. US findings strongly suspicious for metastatic involvement of axillary lymph nodes, namely, extreme reduction or disappearance of hilar echogenicity, should be considered an indication for preoperative percutaneous biopsy. A more widespread use of US-guided axillary node core biopsy could increase the preoperative identification of axillary node metastasis, allowing a significant reduction in the number of SNB, with considerable savings of time and financial resources. References [1] Fisher B, Bauer M, Wickerham DL, et al. Relation of number of positive axillary nodes to the prognosis of patients with primary breast cancer. An NSABP update. Cancer 1983;52:1551–7. [2] Giuliano AE, Kirgan DM, Guenther JM, et al. Lymphatic mapping and sentinel lymphadenectomy for breast cancer. Ann Surg 1994;220: 391– 8. [3] Veronesi U, Paganelli G, Galimberti V, et al. Sentinel-node biopsy to avoid axillary dissection in breast cancer with clinically negative lymph-nodes. Lancet 1997;349:1864 –7. [4] Krag D, Weaver D, Ashikaga T, et al. The sentinel node in breast cancer—a multicenter validation study. N Engl J Med 1998;339: 941– 6. [5] Veronesi U, Galiberti V, Zurrida S, et al. Sentinel lymph node biopsy as an indicator for axillary dissection in early breast cancer. Eur J Cancer 2001;37:454 – 8. [6] de Freitas R, Costa MV, Schneider SV, et al. Accuracy of ultrasound and clinical examination in the diagnosis of axillary lymph node metastases in breast cancer. Eur J Surg Oncol 1991;17:240 – 4. [7] Vassallo P, Wernecke K, Roos N, et al. Differentiation of benign from malignant superficial lymphadenopaty: the role of high-resolution US. Radiology 1992;183:215–20. [8] Yang WT, Ahuja A, Tang A, et al. Ultrasonographic demonstration of normal axillary lymph nodes: a learning curve. J Ultrasound Med 1995;14:823–7. [9] Feu J, Tresserra F, Fabregas R, et al. Metastatic breast carcinoma in axillary lymph nodes: in vitro US detection. Radiology 1997;205: 831–5.
[10] Yang WT, Chang J, Metreweli C. Patients with breast cancer: differences in color doppler flow and gray-scale us features of benign and malignant axillary nodes. Radiology 2000;215:568 –73. [11] Rajesh YS, Ellenbogen S, Banerjee B. Preoperative axillary ultrasound scan: its accuracy in assessing the axillary nodal status in carcinoma of the breast. Breast 2002;11:49 –52. [12] Damera A, Evans AJ, Cornford EJ, et al. Diagnosis of axillary nodal metastases by ultrasound-guided core biopsy in primary operable breast cancer. Br J Cancer 2003;89:1310 –3. [13] Sakai F, Kiyono K, Sone S, et al. Ultrasonic evaluation of cervical metastatic lymphadenopathy. J Ultrasound Med 1998;7:305–10. [14] Krishnamurthy S, Sneige N, Bedi DG, et al. Role of ultrasoundguided fine needle aspiration of indeterminate and suspicious axillary lymph nodes in the initial staging of breast carcinoma. Cancer 2002; 95:982– 8. [15] Vassallo P, Edel G, Roos N, et al. In-vitro high-resolution ultrasonography of benign and malignant lymph nodes. A sonographic-pathologic correlation. Invest Radiol 1993;28:698 –705. [16] Sato K, Tamaki K, Tsuda H, et al. Utility of axillary ultrasound examination to select breast cancer patients suited for optimal sentinel node biopsy. Am J Surg 2004;187:679 – 83. [17] Tschammler A, Ott G, Schang T, et al. Lymphadenopathy: differentiation of bening from malignant disease— colour doppler US assessment of intranodal angioarchitecture. Radiology 1998;208:117–23. [18] Bonnema J, vanGeel A, van Ooijen B, et al. Ultrasound-guided aspiration biopsy for detection of non palpable axillary node metastases in breast cancer patients: new diagnostic method. Worl J Surg 1997;21:270 – 4. [19] Oruwari JU, Chung MA, Koelliker S, et al. Axillary staging using ultrasound-guided fine needle aspiration biopsy in locally advanced breast cancer. Am J Surg 2002;184:307–9. [20] Deurloo EE, Tanis PJ, Gilhuijs KGA, et al. Reduction in the number of sentinel lymph node procedures by preoperative ultrasonography of the axilla in breast cancer. Eur J Cancer 2003;39:1068 –73. [21] Kuenen-Boumeester V, Menke-Pluymers M, de Kanter AY, et al. Ultrasound-guided fine needle aspiration citology of axillary lymph nodes in breast cancer patients. A preoperative staging procedure. Eur J Cancer 2003;39:170 – 4. [22] Clarke D, Sudhakaran N, Gateley CA. Replace fine needle aspiration citology with automated core biopsy in the triple assessment of breast cancer. Ann R Coll Surg Engl 2001;83:110 –2. [23] Parikh J, Tickman R. Image-guided tissue sampling: where the radiology meets pathology. Breast J 2005;11:403–9.
Clinical surgery–International
Role of axillary ultrasound examination in the selection of breast cancer patients for sentinel node biopsy Jacopo Nori, M.D.a, Ermanno Vanzi, M.D.a, Massimo Bazzocchi, M.D.b, Filippo Nori Bufalini, M.D.a, Vito Distante, M.D.c, Francesco Branconi, M.D.d, Tommaso Susini, M.D., Ph.D.d,* a
Diagnostic Senology Unit, Azienda Ospedaliera-Universitaria Careggi, Florence, Italy b Department of Radiology, University of Udine, Udine, Italy c Department of General Surgery, University of Florence, Florence, Italy d Department of Gynecology, Perinatology, and Human Reproductive Medicine, University of Florence, Viale Morgagni, 85 - 50134, Firenze, Italy Manuscript received December 7, 2005; revised manuscript February 20, 2006
Abstract Background: Sentinel node biopsy (SNB) is a time-consuming procedure that can be avoided in presence of axillary metastases. The aim of this study was to assess the accuracy of ultrasound scan (US) in the prediction of axillary nodes status in patients scheduled for SNB. Methods: Axillary US was performed and when feasible, a core biopsy of suspicious nodes was taken. The nodal status as assessed by US and/or core biopsy was compared with final histology. Results: Of the 132 patients enrolled, 31 (23.5%) had suspicious axillary nodes according to US; 19 (61.3%) were true positive, whereas 12 cases (38.7%) were not. In 14 of 31 suspicious cases an US-guided core-biopsy was taken, which in 11 of 14 cases (78.5%) confirmed the neoplastic involvement. Overall, core biopsy of the nodes correctly predicted the final histology in 13 of 14 cases (92.8%). Conclusions: The US of axillary nodes, possibly associated with core biopsy, improved the preoperative evaluation of breast cancer patients scheduled for SNB. © 2007 Excerpta Medica Inc. All rights reserved. Keywords: Breast cancer; Axillary lymph nodes; Ultrasound; Core biopsy; Sentinel node biopsy
The presence of lymph node metastasis, in the absence of distant recurrences, is the single most important prognostic factor in breast cancer [1]. Currently, as many as 70% of diagnosed breast cancers do not exhibit lymph node metastases, making total axillary dissection often unjustified. For this reason, in recent years sentinel node biopsy (SNB) has met with general acceptance, as an attempt to minimize overtreatment of healthy axillae [2–5]. On the other hand, the widespread use of SNB has increased the cost of the overall procedure because of the need of preoperative lymphoscintigraphy, intraoperative ␥–probe use, and prolonged operative time, while waiting for frozen section results. In addition, an expensive and time-consuming intensive pathological study of the sentinel node is required, with up to 20% false negatives on frozen sections that reveal micrometastasis at final histopatologic evaluation, thus requiring a second surgical procedure for axillary node dissection. * Corresponding author. Tel.: ⫹39.055.411879; fax: ⫹39.055.434330. E-mail address: [email protected]
Therefore, accurate selection of patients with high probability of having uninvolved axillary nodes would be crucial in the optimization of the SNB procedure. The clinical examination of the axilla is unfortunately unreliable for correct preoperative staging of the lymph nodes [6]. The ultrasound scan (US) study of the axilla may provide a significant improvement in the preoperative assessment of lymph node status. In expert hands, axillary US almost always allows identification of the lymph nodes, assessment of even the smallest nodes, and in-depth evaluation of node morphology [7–11]. If US criteria to recognize metastatic nodes can be consistently identified, axillary US, possibly in association with US-guided core biopsy, may significantly impact on the choice of the surgical procedure [12]. Obviously, patients with core biopsy– documented positive nodes or US highly suspicious nodes might undergo straight to total axillary node dissection, saving financial resources and time. The aim of our study was to evaluate the preoperative diagnostic accuracy of axillary US, associated when necessary
0002-9610/07/$ – see front matter © 2007 Excerpta Medica Inc. All rights reserved. doi:10.1016/j.amjsurg.2006.02.021
J. Nori et al. / The American Journal of Surgery 193 (2007) 16 –20
17
with US-guided core biopsy, in breast cancer patients scheduled for quadrantectomy combined with SNB. Methods Between July 2001 and December 2003 we performed preoperative axillary US with detailed study of the lymph nodes in 147 women scheduled for conservative breast surgery for whom SNB had been planned. Fifteen patients were excluded from the study: 6 because they were found to have benign breast disease and 9 because they failed to meet the sonographic requirement of at least 4 visible axillary nodes. The US examinations were performed by a single radiologist (J.N.), using a Technos MP unit (Esaote, Genova, Italy), with a high-frequency probe (10 to 13 MHz) and axial and longitudinal scans. To better characterize lymph node status in the patients with doubtful or suspicious US findings, US-guided core biopsy with a semiautomated biopsy gun and Tru-cut needles (Precisa, Hospital Service, Rome, Italy) was performed under local anaesthesia (5 mL subcutaneous Carbocaine, AstraZeneca, Milan, Italy). We used 14- or 16-G needles, with a 23-mm throw. Three samples per node were taken through a single cutaneous approach, from the cortical region, site of possible neoplastic infiltration, avoiding the highly vascular hilum region. In some cases, color-Doppler US was used during the biopsy procedure to avoid dissecting any large vessels. The biopsy samples were fixed in formalin and sent to the pathology laboratory. At the end of the procedure the biopsy site was compressed for 5 minutes before applying sterile tape and dry ice. The biopsy was not performed if the lesion was in close proximity to the large vessels owing to the risk of bleeding. In the first years of the study, also some patients in whom US-guided core biopsy was technically feasible actually were not submitted to the procedure, because our policy towards percutaneous biopsy of the nodes was less “aggressive.” With increasing experience, we are currently able to perform US-guided core biopsy of suspicious axillary nodes in approximately 80% of the cases. The patients’ ages ranged between 28 and 88 years (mean 56.4 years). They all had histologically or cytologically proven breast carcinoma with diameters ranging from 4 to 28 mm (mean diameter 12 mm) and clinically nonpalpable axillary lymph nodes. Histologic type was as follows: ductal carcinoma-in-situ, 3 cases; infiltrating ductal carcinoma, 83 cases; infiltrating lobular carcinoma, 30 cases; and infiltrating ductal and lobular carcinoma, 16 cases. Ultrasound criteria for axillary node evaluation The study consisted in evaluating the presence of at least 4 lymph nodes, which were assessed for size, shape, visibility and morphology of the hilum, vascularity, and focal cortical thickening. In particular, concerning shape, 2 parameters were considered: the ratio between longitudinal diameter and transverse diameter (L/T ratio), which is ⱖ2 in normal lymph nodes, and the ratio between hilum diameter (H) and longitudinal diameter (L) of the lymph node (H/L ratio), normal lymph nodes being considered those with a hilum diameter ⱖ50% of the overall longitudinal diameter (H/L ratio ⱖ50%). Metastatic nodes tend to have globular
Fig. 1. Schematic representation of lymph node morphology. (A) The ratio between the total longitudinal axis (L) and the longitudinal diameter at the hilum (H) is ⱖ50% (normal). (B) The H/L ratio is ⬍50%, a condition more often associated with malignant infiltration of the node.
morphology (L/T ratio ⫽ 1) and compression or disappearance of the hilum (H/L ratio ⬍50%) (Fig. 1). The sonographic criteria considered suspicious for malignancy included: (1) globular morphology and/or H/L ratio less than 50%; (2) disappearance of hilum fat hyperechogenicity; and (3) eccentric focal thickening and denting of the cortex. Patients with a core biopsy– documented lymph node metastasis proceeded directly to total axillary dissection and breast-conserving surgery. All other patients underwent primary surgical treatment including SNB, and when a positive sentinel node was found, total axillary dissection was performed. The final histologic results concerning lymph nodal involvement were compared with the preoperative axillary node status as assessed by US or by US-guided core biopsy. Statistical analysis The results were analyzed using descriptive statistical methods. Sensitivity, specificity, overall accuracy, and negative predictive values were calculated by comparing the results of US and core biopsy with the final histologic nodal findings. Results US findings versus final histology Axillary US was performed in 132 breast cancer patients in whom at least 4 lymph nodes were visualized. Metastases in 1 or more nodes were documented in 42 patients (31.8%) at final histology. Of the 132 patients, 101 (76.5%) had sonographically normal or nonsuspicious lymph nodes according to the previously mentioned US criteria. The definitive histologic examination of these 101 cases confirmed the negative results in 78 true-negative cases (77.2%), whereas in the remaining 23 false-negative cases (22.8%) the histologic examination demonstrated: in 20 cases (86.9%) the presence of micrometastasis in the sentinel lymph node only; in 1 case (4.3%) the presence of metastasis in two lymph nodes; in 1 case (4.3%) the presence of metastasis in four lymph nodes; and in 1 case (4.3%) the presence of metastasis in eight lymph nodes. These 23 patients underwent surgery for primary breast lesions, which, at the histologic examination of the surgical specimen, presented diameters between 7.0 and 18 mm (pT1) (mean diameter 13.5 mm). The remaining 31 patients had sonographically suspicious lymph nodes (Tables 1 and 2), in particular, 26 cases (83.8%) showed hilum alterations: partially visible hilum in 21 cases (80.7%); nonvisible hilum in 5 cases (19.3%); 6
18
J. Nori et al. / The American Journal of Surgery 193 (2007) 16 –20
Table 1 Distribution of true positives and false positives according to hilar morphology and cortical morphology among 31 cases with suspicious axillary ultrasound 31 cases sonographically doubtful or suspicious
Hilar morphology
Cortical morphology
Normal
Partially visible
Not visible
Normal
Focal thickening
Diffuse thickening
19 true positives 12 false positives
0 5
14 7
5 0
10 10
5 1
4 1
cases (19.3%) showed focal cortical thickening and 5 cases (16.1%) diffuse cortical thickening; 19 cases (51.6%) showed H/L ratio less than 50% (Fig. 1). There was agreement between the sonographic findings and the definitive histology in 19 of these 31 patients (61.3% true positive), but no agreement in the remaining 12 cases (38.7% false positive) (Tables 1 and 2). The definitive histologic examination of the 19 true-positive cases demonstrated: in 10 cases (52.6%) positivity of the sentinel node only; in 3 cases (15.8%) positivity of 2 lymph nodes; in 2 cases (10.5%) positivity of 4 lymph nodes; and in 4 cases (21.0%) positivity of 8, 9, 15, and 24 nodes, respectively. In the 12 false-positive cases (38.7%), in which there was disagreement between suspicious US findings and negative histology, the sonographic findings that had raised suspicion were as follows: hilar morphology (partially visible in 7 cases); cortical morphology (focal cortical thickening in 1 case, diffuse cortical thickening in 1 case); and H/L ratio (⬍50% in 3 cases and equal to 50% in 9 cases, but associated with the above-mentioned hilar and cortical morphology alteration). According to these numbers, sensitivity and specificity of axillary US alone in the detection of metastatic nodes were 45.2% and 86.6% respectively, and positive predictive value (PPV) and negative predictive value (NPV) were 61.3% and 77.2%, respectively. Overall accuracy was 73.5%. Core biopsy of suspicious nodes Of the 31 patients with sonographically suspicious lymph nodes, 14 were submitted to US-guided histologic core biopsy with 2 to 3 samplings per node so as to gain further information on lymph node status. In 7 cases, at the beginning of our experience, we did not propose the procedure. The remaining 10 patients did not undergo preoperative percutaneous biopsy because of unfavorable node location (too deep or too close to the large axillary vessels, 6 cases), or patient’s refusal of the procedure (4 cases). By taking these precautions we had no cases of post-biopsy complications. Table 2 Distribution of true positives and false positives according to the ratio between hilar diameter and overall longitudinal diameter of the node (H/L ratio) among 31 cases with suspicious axillary ultrasound 31 cases sonographically doubtful or suspicious
H/L ratio Less than 50%
Equal to 50%
Greater than 50%
19 true positives 12 false positives
16 3
3 9
0 0
The percutaneous node biopsies showed the following results: 11 lymph nodes with neoplastic infiltration; and 3 lymph nodes free of neoplastic lesions. Definitive histology on the surgical specimen confirmed the biopsy findings in all of the 11 positive cases (100% true positives); among the 3 negative cases at the preoperative examination, there was agreement in 2 cases (66.6% true negatives) and disagreement in 1 case (33.3% false negatives) that final histology showed to be involved. Sensitivity, specificity, PPV, and NPV were 91.6%, 100%, 100%, and 66.6%, respectively. Overall accuracy was 92.8%. Comments The sonographic study of the axilla may be used as a non-invasive and fast method for assessing lymph node status in the preoperative staging of breast cancer patients and in the subsequent planning of local and systemic treatment. In metastatic lymph nodes, the process mainly involves the entire subcortical or cortical region, which, being invaded by the neoplastic tissue, becomes dishomogeneously hypertrophic deforming or even cancels the hyperechogenicity of the hilum or lymphatic sinusoids [6,7,9]. The sonographic parameters to be assessed in the study of the lymph nodes are: dimensions; lymph node shape; morphology of the hilum and cortex; color-Doppler patterns. The longitudinal diameter of normal lymph nodes varies widely, which means it cannot constitute the only criterion to guide the level of US suspicion. Hence, reactive lymph nodes may have diameters of up to 2.5 cm, while lymph nodes with diameters as small as 5 mm may exhibit micrometastasis. Node morphology is an important criterion: reactive or hyperplastic nodes have a reniform, ovoid shape, with longer longitudinal diameter (L) than transverse diameter (T), whereas neoplastic nodes tend to have a globular appearance, with a L/T ratio of approximately 1. The ratio between these 2 diameters has been defined as the “roundness index” by Sakai et al, who demonstrated the L/T ratio to be greater than 2 in approximately 80% of histologically benign nodes, but less than 2 in malignant nodes [13]. This index also has moderate but not absolute specificity. As regards morphology of the hilum region, it should be recalled that the hilum is normally located centrally with respect to the node and that its level of conspicuity may be suggestive of an alteration that requires further investigation. Even if compressed and smaller than usual, the hilum can be documented in about 40% of neoplastic nodes. The cortical region appears as a peripheral hypoechoic rim of regular homogeneous thickness; thickening, irregularity, or interruption of the cortex is a sign of extracapsular spread and involvement of the perinodal fat, which
J. Nori et al. / The American Journal of Surgery 193 (2007) 16 –20
may be related to metastatic cells penetrating the node via the afferent lymphatic vessels [11,14]. The careful study of the hilum and cortical regions may give rise to a sonographic pattern of doubtful or suspicious nodal status. In the current study, among false-positive cases (12 women) the most misleading parameter was increased cortical thickness and consequently a ratio between overall node size (total longitudinal diameter) and size of the hyperechoic hilar portion (H/L ratio) equal to 50%. In contrast, no patient with lymph node metastases at definitive histology had a normal hilar morphology by US. This confirms that alterations of the hilar region are the most important parameters in establishing a sonographic diagnosis of suspicion or doubt, as already reported by others [9,15,16]. In our study, we considered only morphologic parameters, and in particular nodal dimensions (longitudinal and transverse diameters), hilar and cortical morphology, and H/L ratio (Fig. 1). The information obtained from color-Doppler examination was available in only few cases, but may represent a further improvement of US examination of axillary nodes [10,17]. The H/L ratio was measured in all cases, and this measurement, combined with the morphologic parameters, may, in our opinion, improve specificity. We considered all lymph nodes with H/L ratio less than 50% to be doubtful or suspicious. On the basis of the echo structural morphologic features (morphology of hilum and cortex; H/L ratio) the nodes were classified as benign or suspicious; in addition, sonography also visualized nonpalpable nodes that could not be clinically assessed, thereby increasing diagnostic sensitivity in preoperative staging. Of the parameters described, the H/L ratio showed moderate specificity, especially if combined with the morphologic assessment of the hilar and cortical regions. Of 31 patients with positive sonography in our series, in only 61.3% (19/31 cases) did the sonographic suspicion correspond to a histopathologic alteration in 1 or more lymph nodes (true positives), whereas in 38.7% of cases (12/31 cases) there was no concordance between sonography and histopathology (false positives). The H/L ratio was less than 50%, and therefore a sign of doubtful or suspicious node status, in 84.2% (16/19 cases) of the true-positive cases confirmed by histology, in comparison with 25% (3/12 cases) among the false positives, with doubtful sonographic findings, but negative histology for tumor infiltration. Our findings in this initial case series appear to indicate that this index (H/L ratio ⬍50%) can be added to the sonographic parameters orienting towards possible metastatic involvement of lymph nodes in patients with breast cancer. In 101 women in whom sonography detected normal or at any rate nonsuspicious lymph nodes, sonography was unable to identify the alterations demonstrated by the pathologist in 22.8% of cases (false negatives), whereas there was agreement between sonography and histology in 77.2% of cases (true negatives). Therefore, while axillary US is fairly effective for identifying the true-negative cases, it also introduces a certain number of false negatives. This may be related either to the inability to recognize all nodes at sonography or to the presence of micrometastasis that are too small to produce sonographically detectable echo structural alterations. Indeed, in the current series, most of the
19
sonographic false negatives were related to micrometastasis, ranging from .1 to .5 mm in size. Sonography can also be used to guide the performance of fine-needle aspiration cytology (FNAC) or, in this series, histologic core biopsy on suspicious nodes, further increasing specificity. Several reports have been published on the use of US-guided FNAC on axillary nodes in breast cancer patients, reporting sensitivity values ranging from 86.4% to 94.7%, specificity values from 97.1% to 100%, and diagnostic accuracy values from 85.7% to 100% [14,18 –21]. The false-negative results were ascribed to difficulties in interpreting the slides, difficulties in sampling small-size lymph nodes, or sampling errors. When technically feasible, the patients in our series underwent percutaneous histologic core biopsy of the suspicious axillary node to enable a preoperative histologic diagnosis, which allowed the surgeon to plan directly axillary dissection in positive cases, without having to perform SNB. In 100% of the cases (11/11) with a positive core biopsy, the diagnosis was confirmed by the final histopathologic examination. On the other hand, in 66.6% (2/3) of cases both the core biopsy and the final histopathologic examination excluded metastatic involvement of sonographically suspicious nodes. Core biopsy yielded only 1 false-negative result, in a case recognized as a metastatic node at definitive histology. Hence, the concordance between core biopsy on sonographically suspicious nodes and final histology was very high (92.8%; 13/14 cases). The only previous report of US-guided core biopsy in the preoperative diagnosis of axillary node metastasis is, to our knowledge, that of Damera et al [12]. In their study, out of 48 nodal core biopsies performed, 26 (54%) were positive for metastatic invasion. Although in the current series the number of core biopsy procedures was lower, we found a higher concordance between US and histology, with 78.5% (11/14) of the biopsies confirming the US suspicion of metastatic invasion. Despite the experience with core biopsy of axillary nodes is still limited, we believe that this approach will receive increasing attention in the near future, for preoperative diagnosis of both primary breast cancer and axillary node metastases. Indeed, the core biopsy technique, in comparison with FNAC, allows a substantial reduction of the inadequate (C1–B1), doubtful (C3–B3), and suspicious (C4 –B4) categories of the pathologic reports [22,23]. The availability of a true histologic diagnosis seems us preferable when planning whether or not to perform total axillary node dissection. Overall, our study confirms that a negative axillary sonography does not allow one to hypothesize about the absence of nodal involvement. This limitation is due to the possible presence of micrometastasis in lymph nodes with an apparently normal echo structure. On the other hand, a positive or suspicious axillary sonography was confirmed at final histologic examination in only 61.3% of the cases, with as many as 38.7% of cases being false positive. The most important sonographic alteration was the disappearance of hilar echogenicity, observed in 100% of true metastatic nodes (Table 1), in agreement with the findings of Sato et al [16]. Conversely, hilar denting or irregularities, as well as dimensional criteria, proved to be poorly specific. In conclusion, the sonographic assessment of axillary nodes, possibly in combination with core biopsy, may improve
20
J. Nori et al. / The American Journal of Surgery 193 (2007) 16 –20
the preoperative evaluation of breast cancer patients scheduled for conservative surgery of the breast and SNB. US findings strongly suspicious for metastatic involvement of axillary lymph nodes, namely, extreme reduction or disappearance of hilar echogenicity, should be considered an indication for preoperative percutaneous biopsy. A more widespread use of US-guided axillary node core biopsy could increase the preoperative identification of axillary node metastasis, allowing a significant reduction in the number of SNB, with considerable savings of time and financial resources. References [1] Fisher B, Bauer M, Wickerham DL, et al. Relation of number of positive axillary nodes to the prognosis of patients with primary breast cancer. An NSABP update. Cancer 1983;52:1551–7. [2] Giuliano AE, Kirgan DM, Guenther JM, et al. Lymphatic mapping and sentinel lymphadenectomy for breast cancer. Ann Surg 1994;220: 391– 8. [3] Veronesi U, Paganelli G, Galimberti V, et al. Sentinel-node biopsy to avoid axillary dissection in breast cancer with clinically negative lymph-nodes. Lancet 1997;349:1864 –7. [4] Krag D, Weaver D, Ashikaga T, et al. The sentinel node in breast cancer—a multicenter validation study. N Engl J Med 1998;339: 941– 6. [5] Veronesi U, Galiberti V, Zurrida S, et al. Sentinel lymph node biopsy as an indicator for axillary dissection in early breast cancer. Eur J Cancer 2001;37:454 – 8. [6] de Freitas R, Costa MV, Schneider SV, et al. Accuracy of ultrasound and clinical examination in the diagnosis of axillary lymph node metastases in breast cancer. Eur J Surg Oncol 1991;17:240 – 4. [7] Vassallo P, Wernecke K, Roos N, et al. Differentiation of benign from malignant superficial lymphadenopaty: the role of high-resolution US. Radiology 1992;183:215–20. [8] Yang WT, Ahuja A, Tang A, et al. Ultrasonographic demonstration of normal axillary lymph nodes: a learning curve. J Ultrasound Med 1995;14:823–7. [9] Feu J, Tresserra F, Fabregas R, et al. Metastatic breast carcinoma in axillary lymph nodes: in vitro US detection. Radiology 1997;205: 831–5.
[10] Yang WT, Chang J, Metreweli C. Patients with breast cancer: differences in color doppler flow and gray-scale us features of benign and malignant axillary nodes. Radiology 2000;215:568 –73. [11] Rajesh YS, Ellenbogen S, Banerjee B. Preoperative axillary ultrasound scan: its accuracy in assessing the axillary nodal status in carcinoma of the breast. Breast 2002;11:49 –52. [12] Damera A, Evans AJ, Cornford EJ, et al. Diagnosis of axillary nodal metastases by ultrasound-guided core biopsy in primary operable breast cancer. Br J Cancer 2003;89:1310 –3. [13] Sakai F, Kiyono K, Sone S, et al. Ultrasonic evaluation of cervical metastatic lymphadenopathy. J Ultrasound Med 1998;7:305–10. [14] Krishnamurthy S, Sneige N, Bedi DG, et al. Role of ultrasoundguided fine needle aspiration of indeterminate and suspicious axillary lymph nodes in the initial staging of breast carcinoma. Cancer 2002; 95:982– 8. [15] Vassallo P, Edel G, Roos N, et al. In-vitro high-resolution ultrasonography of benign and malignant lymph nodes. A sonographic-pathologic correlation. Invest Radiol 1993;28:698 –705. [16] Sato K, Tamaki K, Tsuda H, et al. Utility of axillary ultrasound examination to select breast cancer patients suited for optimal sentinel node biopsy. Am J Surg 2004;187:679 – 83. [17] Tschammler A, Ott G, Schang T, et al. Lymphadenopathy: differentiation of bening from malignant disease— colour doppler US assessment of intranodal angioarchitecture. Radiology 1998;208:117–23. [18] Bonnema J, vanGeel A, van Ooijen B, et al. Ultrasound-guided aspiration biopsy for detection of non palpable axillary node metastases in breast cancer patients: new diagnostic method. Worl J Surg 1997;21:270 – 4. [19] Oruwari JU, Chung MA, Koelliker S, et al. Axillary staging using ultrasound-guided fine needle aspiration biopsy in locally advanced breast cancer. Am J Surg 2002;184:307–9. [20] Deurloo EE, Tanis PJ, Gilhuijs KGA, et al. Reduction in the number of sentinel lymph node procedures by preoperative ultrasonography of the axilla in breast cancer. Eur J Cancer 2003;39:1068 –73. [21] Kuenen-Boumeester V, Menke-Pluymers M, de Kanter AY, et al. Ultrasound-guided fine needle aspiration citology of axillary lymph nodes in breast cancer patients. A preoperative staging procedure. Eur J Cancer 2003;39:170 – 4. [22] Clarke D, Sudhakaran N, Gateley CA. Replace fine needle aspiration citology with automated core biopsy in the triple assessment of breast cancer. Ann R Coll Surg Engl 2001;83:110 –2. [23] Parikh J, Tickman R. Image-guided tissue sampling: where the radiology meets pathology. Breast J 2005;11:403–9.