Three-dimensional ultrasound-validated large-core needle biopsy: is it a reliable method for the histological assessment of breast lesions?

Transcription

1 Ultrasound Obstet Gynecol 2004; 23: Published online 24 February 2004 in Wiley InterScience ( DOI: /uog.1001 Three-dimensional ultrasound-validated large-core needle biopsy: is it a reliable method for the histological assessment of breast lesions? L. DELLE CHIAIE and R. TERINDE Centre for Ultrasound and Prenatal Diagnosis, Department of Obstetrics and Gynecology, University of Ulm, Ulm, Germany KEYWORDS: 3D ultrasound; 3D ultrasound-validated LCNB; breast biopsy; core needle biopsy ABSTRACT INTRODUCTION Objective The use of three-dimensional (3D) ultrasound may help to determine the exact position of the needle during breast biopsy, thereby reducing the number of core samples that are needed to achieve a reliable histological diagnosis. The aim of this study was to demonstrate the efficacy of 3D ultrasound-validated large-core needle biopsy (LCNB) of the breast. Methods A total of 360 core needle biopsies was obtained from 169 breast lesions in 146 patients. Additional open breast biopsy was performed in 111 women (127/169 breast lesions); the remaining 42 lesions were followed up for at least 24 months. 3D ultrasound visualization of the needle in the postfiring position was used to classify the biopsy as central, marginal or outside the lesion. Based on this classification it was decided whether another sample had to be obtained. Results A median of two core samples per lesion provided for all the lesions a sensitivity for malignancy of 96.9%, specificity of 100%, false-positive rate of 0% and false-negative rate of 3.1%, and for the excised lesions a sensitivity of 96.5%, specificity of 100%, falsepositive rate of 0%, false-negative rate of 3.5% and an underestimation rate of 3.4%. Conclusions 3D ultrasound validation of the postfiring needle position is an efficient adjunct to ultrasound-guided LCNB. The advantages of 3D ultrasound validation are likely to include a reduction in the number of core samples needed to achieve a reliable histological diagnosis (and a possible reduction in the risk of tumor cell displacement), reduced procedure time and lower costs. Copyright 2004 ISUOG. Published by John Wiley & Sons, Ltd. The widespread use of mammographic screening for early diagnosis of breast cancer has increased the need for biopsy taking in order to further evaluate the histology of breast abnormalities. Until recently, fine needle aspiration (FNA) was a widely accepted method of determining the nature of an abnormality observed on sonography or mammography. Apart from the costs and morbidity associated with this method, FNA has the drawback of reducing the reliability of mammographic follow-up due to associated scarring and tissue retractions 1. Open biopsy is a surgical procedure performed solely to establish the diagnosis in cases of malignancy 2 that cannot be 100% sensitive and specific (needle-localized open biopsy: miss rate for all lesions 2.6%; false-negative rate 2%) 3. Percutaneous imaging-guided breast biopsy has been developed as an alternative to surgical biopsy for the histological assessment of breast lesions 4,5 and can be performed under stereotactic mammography and ultrasound guidance. The choice of imaging technique used depends on the visibility of the lesion, but generally ultrasound is preferred if the lesion is visible with both techniques. The main reason for false-negative results in imagingguided biopsy techniques is the sampling error, which can be reduced by using ultrasound. Ultrasound-guided procedures have several advantages over the stereotactic approach, namely they do not involve ionizing radiation, all areas of the breast are accessible, and the needle position can be verified and corrected in real time. A limitation of two-dimensional (2D) ultrasound is the misleading artifactual appearance of correct needle placement when it is positioned at the edge of the lesion. This inaccurate information can be compensated Correspondence to: Dr L. Delle Chiaie, Centre for Ultrasound and Prenatal Diagnosis, Department of Obstetrics and Gynecology, University of Ulm, Prittwitzstrasse 43, D Ulm, Germany ( loredana.delle-chiaie@web.de) Accepted: 14 December 2003 Copyright 2004 ISUOG. Published by John Wiley & Sons, Ltd. ORIGINAL PAPER

2 394 Delle Chiaie and Terinde for with three-dimensional (3D) ultrasound. The aim of the present study was to demonstrate the effectiveness of 3D ultrasound-validated large-core needle biopsy (LCNB) of the breast in obtaining a reliable histological diagnosis despite requiring fewer tissue samples. METHODS From September 2000 to September 2001 we examined 146 patients referred to our clinic because of palpable or non-palpable breast lesions or because of suspicion of breast cancer on mammography. Informed consent was obtained from each patient. We initially performed a 2D ultrasound examination using a 3D ultrasound volume transducer, MHz, 30 volume sector (Voluson 530D, Kretztechnik- Medison, Zipf, Austria) with the patient in a supine position and with elevated arms to identify the suspected lesion and to exclude a second finding in the same or in the contralateral breast. Once located, the lesion was evaluated with 3D ultrasound scanning (3D ultrasound volume dataset: multiplanar scan analysis) to obtain more information about the lesion and its surrounding structures. In 146 patients, 169 solid breast lesions were identified (Table 1). Fourteen women had two lesions and four women had three lesions; overall, 360 LCNBs were performed under ultrasound guidance (Bard Magnum Biopsy Instrument, Bard Magnum, Covington, GA, USA; 14-gauge needle). After injecting local anesthesia and performing a 2-mm skin incision the needle was inserted under 2D ultrasound guidance and was positioned at the margin of the lesion (prefiring position). A 22- or 15-mm core needle throw Table 1 Characteristics of the patients and respective breast lesions that underwent large-core needle biopsy Characteristic Patients (n) Breast lesions (n) Sample Excised lesions Follow-up (24 months) Only chemotherapy 9 11 Insufficient sampling* 2 2 *Patients with insufficient sampling at large-core needle biopsy underwent surgical excision; in the table these cases are included among the excised lesions. (depending on the dimensions and the consistency of the lesion) followed and a 3D ultrasound volume dataset was acquired to check the postfiring position of the needle (3D ultrasound validation) 6. Within approximately 4 s the system acquires the entire 3D volume dataset (10 Mb) and displays the information in a multiplanar imaging mode. The volume with the needle should be rotated in the A- and C-planes to optimize needle visualization. An accurate evaluation of the needle position was possible and the biopsy could be classified as being central (Figure 1), marginal or outside the lesion 7. If the first biopsy was classified as being outside the lesion, additional biopsies were performed until 3D ultrasound of the postfiring needle position confirmed a central or marginal hit. Additional biopsies were obtained if more tissue was necessary for diagnostic purposes. A total of 111 patients (127 breast lesions, 272 LCNBs) underwent surgical excision of their lesion(s) because of a histological diagnosis of malignancy at biopsy (100 cases) or because of discordance between imaging characteristics Figure 1 Three-dimensional multiplanar visualization of the needle in the postfiring position. In this case the needle is classified as central hit.

3 Three-dimensional ultrasound and breast biopsy 395 and/or clinical suspicion and histopathological diagnosis (11 cases). Nine patients (11 breast lesions, 23 LCNBs) underwent solely chemotherapy. Twenty-six patients (31 breast lesions, 65 LCNB) with low or intermediate suspicion and with concordance between the radiographic finding and pathological diagnosis were followed up for at least 24 months (Table 1). The sensitivity, specificity, accuracy, false-positive and false-negative rates, and positive (PPV) and negative predictive values (NPV) of the LCNBs after 3D validation and reduction of core sampling were analyzed for all the lesions. For those patients who underwent chemotherapy only, the diagnostic clinical and imaging criteria of breast cancer were used to calculate the result of the LCNB. In the cases with low or intermediate suspicion that did not undergo surgery, the outcome of the follow-up controls was used to assess the result of the LCNB. In the excised lesions we evaluated the accuracy of the preoperative histological diagnosis with respect to grading (Scarf Bloom Richardson System) and assessment of invasion comparing the results after LCNB and after surgery in 116 lesions diagnosed as malignant. RESULTS In 146 patients, 169 breast lesions were found and 360 LCNBs were performed. The mean diameter of the lesions was 1.5 cm; 29.8% of the lesions had a mean diameter 1 cm, 48.2% had a mean diameter between 1.1 and 2cm,andin22%thelesionwas> 2cm. For each lesion, one to four biopsies were performed: one core was sampled in 22% of the lesions, two cores in 60%, three cores in 15.4% and four cores in 2.9% (median two samples per lesion). In 85 cases an additional biopsy was necessary due to the first hit being outside the lesion. In a further 69 cases biopsy was repeated because more tissue was necessary for determination of prognostic factors. There were two cases of insufficient sampling where it was not possible to establish a diagnosis (1.1% of the lesions). In these cases an open biopsy was performed and in both malignancy was diagnosed (in both cases ductolobular carcinoma). After needle biopsy 124 lesions revealed malignancy and 43 were classified as benign, but in four of these cases classified as benign a carcinoma was detected after surgery (false-negative cases). Overall, after surgery it was possible to detect 130 malignant lesions (including the two cases with non-diagnostic biopsy). The histological findings after surgical excision or after LCNB alone (for the patients who did not undergo surgical excision) are shown in Table 2. After excluding the two cases mentioned above in which a histological result was not obtained we calculated the sensitivity, specificity, accuracy, false-positive and falsenegative rates, and PPVs and NPVs. The sensitivity of the 3D ultrasound-validated LCNB in diagnosing malignancy was 96.9% with 100% specificity, Table 2 Histopathological findings in 169 breast lesions after surgical excision and after solely large-core needle biopsy (35 women, 42 breast lesions) in those cases in which the patient did not undergo surgery Lesion type Malignant lesions 130 Invasive ductal carcinoma (IDC) 83 Ductal carcinoma in situ (DCIS) 4 IDC + DCIS 11 Ductal carcinoma micro-invasive 2 Invasive lobular carcinoma (ILC) 7 ILC + lobular carcinoma in situ 3 Tubular carcinoma 4 Mucinous carcinoma 3 Medullary carcinoma 4 Ductolobular carcinoma 6 Ductotubular carcinoma 1 Invasive cribriform ductal carcinoma 1 Extramedullary plasmocytoma 1 Benign lesions 39 Gynecomastia 1 Non-lactating abscess 2 Fibrocystic disease 4 Fibrosis 4 Granulomatous mastitis 5 Ductal epithelial hyperplasia 3 Mastitis 6 Cholesterol granuloma 2 Normal scar 1 Fatty tissue 3 Fibroadenoma 2 Phyllodes tumor 1 Normal breast parenchyma 3 Fat necrosis 1 Sclerosing adenosis 1 Table 3 Statistics of the three-dimensional ultrasound-validated large-core needle biopsy in diagnosing malignancy* Parameter All lesions (%) n Excised lesions (%) Sensitivity Specificity False-positive rate 0 0 False-negative rate Positive predictive value Negative predictive value Accuracy Insufficient sampling Underestimation at core 3.4 needle biopsy Concordance of grading 67.6 *Data for the complete sample (i.e. 167 breast lesions, of which 125 lesions underwent surgery, 42 cases were evaluated by follow-up and in two cases no histological evaluation was possible after large-core needle biopsy) and exclusively on the excised sample. 100% PPV, 90.9% NPV, 0% false-positive rate and 3.1% false-negative rate. The accuracy of the diagnosis was 97.6%. The equivalent values for the cases that underwent surgery after percutaneous biopsy were 96.5% sensitivity, 100% specificity, 100% PPV, 69.2% NPV,

4 396 Delle Chiaie and Terinde 0% false-positive rate and 3.5% false-negative rate (Table 3). Assessment of the invasion of the basal membrane showed a concordance between the pre- and postoperative diagnoses in 96.6% of the cases and there was underestimation in 3.4% (four cases with diagnosis of ductal carcinoma in situ that revealed an infiltrating ductal carcinoma at surgery) (Table 3). For invasive carcinomas a concordance in the assessment of the grading was found in 67.6% of the cases, an underestimation at biopsy in 27.6% and an overestimation at biopsy in 4.6% of the samples (Table 3). The percentages of the various grades in the percutaneous biopsies were G1 in 5.9%, G2 in 67.6% and G3 in 26.5% of the samples; after surgery 2.9% of the cases were classified as G1, 48.5% as G2 and 48.5% as G3. DISCUSSION For many years, FNA of the breast was considered a reliable sampling method that, in certain instances, could avoid surgical biopsy although many surgeons have been skeptical about basing a definitive decision on the results of FNA. In addition to a significant number of inadequate or non-diagnostic cytological samples (4 36%) and a high false-negative rate (0 32%) there are other drawbacks of FNA. These include the lack of histological diagnosis that is crucial for determining whether a cancer is invasive or not, the impossibility of determining the histological grading, and the need for a highly-trained cytopathologist 8,9. Automated LCNB of the breast could greatly facilitate the management of breast cancer because of its better characterization of benign and malignant lesions and its lower frequency of insufficient samples (< 1 4%). The false-negative rate reported in the literature of 4% 8 is similar to that of surgical biopsy (0 8%) 3. If an LCNB yields a benign diagnosis concordant with the imaging characteristics then surgery can be avoided 4, thereby eliminating one surgical procedure that would have been performed only for diagnostic purposes 10. To ensure accurate tissue sampling it has been suggested that a minimum of five and 10 cores need to be obtained for masses and microcalcifications, respectively 9,11. In order to reduce the drawbacks of the core biopsy devices a new method of tissue acquisition has been developed, namely the vacuum-assisted biopsy instrument (Mammotome, Ethicon Endo-surgery, Inc., Cincinnati, OH, USA). It is inserted once and rotated while in the breast to obtain samples from different areas of the lesion that are double the size of those obtained by conventional biopsy 12. This technique allows the histopathological assessment of microcalcifications and can also be used to remove small benign lesions completely (e.g. fibroadenomas) leading to a reduction in the histological underestimation rate (in lesions containing atypical ductal hyperplasia and ductal carcinoma in situ) which is a problem affecting the LCNB because the samples are not contiguous 13. The use of 3D ultrasound validation to confirm the postfiring position of the needle in relation to the focus could help to achieve a reliable and accurate diagnosis requiring a significantly smaller number of tissue samples. In the present study, if we had obtained five samples from each lesion, we would have acquired a total of 845 cores instead of 360 (the difference being significantly different; t-test, P < ). Thus procedure-related time requirements and costs, as well as levels of discomfort for the patients in our study, were less without reducing the diagnostic accuracy. Insufficient sampling in our series was 1.1% whereas in the literature it ranges from 1% to 4.4% 9,14 ; our results reveal a false-negative rate of 3.1% which is comparable with the results reported in the literature (mean, 4.2%; range, %) 15 ; our sensitivity and specificity (96.9% and 100%, respectively) corresponded to the figures reported elsewhere (85 99% and 100%) 5, We obtained similar results even when analyzing only the cases that underwent surgery after percutaneous biopsy, albeit the NPV was lower (69.2%). The latter was likely to be due to the few true negative cases in this group (patients with a benign lesion for which the preoperative diagnosis was correct). The NPV depends on the rate of healthy cases in the sample. Our hospital is a referral center for breast disease and therefore we deal only with patients at risk: this could be the reason why the NPV cannot evaluate correctly the validity of the biopsy method in our investigation. As outlined by Weismann et al. 6, the additional coronal plane available in 3D ultrasound imaging allows a precise correlation of needle position and breast lesion. This fact will influence the decision whether to obtain another sample or to end the biopsy procedure. In our study population, in four cases we missed the diagnosis and in two cases we obtained inadequate samples despite placement control of the needle in the three planes. We think that besides the necessary learning curve (three of the missed cases were at the beginning of the study), other possible explanations of the problem were also the small dimensions (one case with mean diameter < 1cm) and the blurred edges of the lesions (in two cases). Apart from saving time and costs when fewer specimens are obtained and analyzed, as a new sample requires repositioning of the needle within the breast the risk of bleeding and of tumor cell displacement that may increase the risk of local and skin recurrence will be reduced. Diaz et al. observed tumor cell displacement in 32% of patients who underwent LCNB 20. Seeding of biopsy needle tracks with viable malignant cells was initially thought to promote iatrogenic stromal lymphatic and vascular dissemination of carcinoma, but a significant morbidity deriving from such cell displacement has not been definitively established. The rate of observed tumor displacement can be affected by several factors including type, size and grade of the tumor and type of needle used. Although not specifically investigated, a correlation was suggested with the number of needle passes required by each procedure and the rate of tumor cell displacement. The decreasing incidence of tumor displacement with

5 Three-dimensional ultrasound and breast biopsy 397 increasing interval between biopsy and surgery suggests that this time interval is of importance. In fact most of the displaced cells do not survive displacement, being likely removed by the inflammatory reaction to the biopsy. The clinical influence of the time interval on the prognosis and morbidity is controversial and the excision of the needle tract as well as radiation of the puncture site have been suggested to minimize the risk of tumor recurrence in this area. A reduction in needle-passes (i.e. reducing the number of cores needed to achieve a nevertheless reliable histological diagnosis) could also help to minimize the risk of local recurrence at the site of the puncture REFERENCES 1. Smith W, Surry KJM, Millis GR, Downey DB, Fenster A. Three dimensional ultrasound-guided core needle breast biopsy. Ultrasound Med Biol 2001; 27: Bear H. Image-guided breast biopsy-how, when and by whom? J Surg Oncol 1996; 67: Jackman RJ, Marzoni FA Jr. Needle-localized breast biopsy: why do we fail? Radiology 1997; 204: Libermann L. Percutaneous imaging-guided core breast biopsy. State of the art at the millennium. AJR Am J Roentgenol 2000; 174: Parker SH, Burbanak F, Jackman RJ, Aucreaman CJ, Cardenosa G, Cink TM, Coscia JL Jr, Eklund GW, Evans WP, Garver PR, Graman HF, Haas DK, Jacob KM, Kally KM, Killebrew LK, Lechner MC, Perlman SJ, Smid AP, Tabar L, Taber FE, Wynn RT. Percutaneous large core breast biopsy: a multi-institutional study. Radiology 1994; 193: Weismann CF, Forstner R, Prokop E, Rettenbacher T. Three dimensional targeting: a new three-dimensional ultrasound technique to evaluate needle position during breast biopsy. Ultrasound Obstet Gynecol 2000; 16: Weismann CF, Forstner R, Prokop E, Rettenbacher T. Threedimensional targeting technique: can the number of core needle biopsies of a breast lesion be reduced without diagnostic loss? Radiology 2000; 217(Suppl.): Parker SH, Dennis MA, Stavros AT, Johnson KK. Ultrasound guided mammotomy. A new breast biopsy technique. JDMS 1996; 12: Venta LA. Image-guided biopsy of non-palpable breast lesion. In: Diseases of the Breast (2nd edn), Harris JR, Lippman ME, Morrow M, Osborne CK (eds). Lippincott Williams & Wilkins: Philadelphia, PA, 2000; Reynolds HE. Core needle biopsy of challenging benign breast conditions. A comprehensive literature review. AJR Am J Roentgenol 2000; 174: Liberman L, Dershaw DD, Rosen PP, Abramson AF, Deutch BM, Hann LE. Stereotaxic 14-gauge breast biopsy: how many core biopsy specimens are needed? Radiology 1994; 192: Burbank F, Parker SH, Fogarty TJ. Stereotactic breast biopsy: improved tissue harvesting with the Mammotome. Am Surg 1996; 62: Burbank F. Stereotactic breast biopsy of atypical ductal hyperplasia and ductal carcinoma in situ lesions: improved accuracy with directional vacuum-assisted biopsy. Radiology 1997; 202: Philipotts LE, Shaheen NA, Carter D, Lange R, Lee CH. Comparison of rebiopsy rates after stereotactic core needle biopsy of the breast with 11-gauge vacuum suction probe versus 14- gauge needle automated gun. AJR Am J Roentgenol 1999; 172: Jackman RJ, Nowels KW, Rodriguez-Soto J, Marzoni FA Jr, Finkelstein SI, Shepard MJ. Stereotactic, automated large-coreneedle biopsy of nonpalpable breast lesions: false-negative and histologic underestimation rates after long-term follow-up. Radiology 1999; 210: Nguyen M, McCombs MM, Ghandehari S, Kim A, Wang H, Barsley SH, Love S, Bassett LW. An update on core-needle biopsy for radiologically detected breast lesions. Cancer 1996; 78: Jackman RJ, Nowel KW, Shepard MJ, Finkelstein SI, Marzoni MA Jr. Stereotaxic large core-needle biopsy of 450 nonpalpable breast lesions with surgical correlation in lesions with cancer or atypical hyperplasia. Radiology 1994; 193: Parker SH, Jobe WE, Dennis MA, Stavros AT, Johnson KK, Yates WF, Truell JE, Price JG, Kortz AB, Clark DG. USguided automated large core biopsy. Radiology 1993; 187: Berg WA. When is core breast biopsy or fine needle aspiration not enough? Radiology 1996; 198: Diaz LK, Wiley EL, Venta LA. Are malignant cells displaced by large-gauge needle core biopsy of the breast. AJR Am J Roentgenol 1999; 173: Chao C, Torosian MH, Boraas MC, Sigurdson ER, Hoffman JP, Eisenberg B, Fowble B. Local recurrence of breast cancer in the stereotactic core needle biopsy site: case reports and review of the literature. Breast J 2001; 2: Stoiler A, Skinner J, Levine EA. A prospective study of seeding of the skin after core biopsy of the breast. Am J Surg 2000; 180: