Breast biopsy state of the art PDF

Preview

Summary

articulo, procedimientos mama,...

Description

Clinical Radiology 65 (2010) 259–270

Contents lists available at ScienceDirect

Clinical Radiology jo u r n a l h o m e p a g e : w w w . e l s e v i e r h e a l t h . c o m / jo u r n a l s / c r a d

Review

Image-guided breast biopsy: state-of-the-art E.A.M. O’Flynn*, A.R.M. Wilson, M.J. Michell South East London Breast Screening Programme and National Breast Screening Training Centre, Kings College Hospital NHS Foundation Trust, London SE5 9RS, UK

art icle informat ion Article history: Received 27 June 2009 Received in revised form 20 December 2009 Accepted 4 January 2010

Percutaneous image-guided breast biopsy is widely practised to evaluate predominantly non-palpable breast lesions. There has been steady development in percutaneous biopsy techniques. Fine-needle aspiration cytology was the original method of sampling, followed in the early 1990s by large core needle biopsy. The accuracy of both has been improved by ultrasound and stereotactic guidance. Larger bore vacuum-assisted biopsy devices became available in the late 1990s and are now commonplace in most breast units. We review the different types of breast biopsy devices currently available together with various localization techniques used, focusing on their advantages, limitations and current controversial clinical management issues. Ó 2010 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.

Introduction Percutaneous image-guided breast biopsy has been developed largely to solve the diagnostic problem of increasing numbers of impalpable breast lesions found through breast cancer screening. The effectiveness and improvements in accuracy achieved over the last 20 years have meant that this technique is now used to assess both palpable and non-palpable lesions in both the screening and symptomatic settings. The high accuracy of non-operative diagnosis with needle biopsy has led to a steady decline in the number of diagnostic open surgical biopsies performed over the years. From the most recent published screening data, 2616 open biopsies were performed in the UK in 2008/9 of which 69% were benign and 31% malignant. The malignant open biopsy rate has fallen from 2.04 per 1000 women in 1996/7 to 0.40 per 1000 women in 2008/9, as the non-operative diagnosis rate for cancers has increased from 63% to a substantial 95%.1

* Guarantor and correspondent: E.A.M. O’Flynn, Department of Breast Radiology, Kings College Hospital, Denmark Hill, London, SE5 9RS, UK. Tel.: þ44 2032993875; fax: 02032994363. E-mail address: lizofl[email protected] (E.A.M. O’Flynn).

All women with breast abnormalities are assessed using the triple diagnostic method. Clinical and imaging assessment should be carried out prior to needle biopsy. Imaging abnormalities are categorized according to a standard breast imaging reporting system. In the UK, the Royal College of Radiologists Breast Group (RCRBG) breast imaging classification is used.2 This uses a five-point scale to indicate the level of suspicion for malignancy based on the imaging findings. The score can be incorporated into diagnostic protocols and provides a guide for the need for needle biopsy: Category 1 (normal) or 2 (benign findings) lesions do not require sampling. Category 3 (indeterminate/probably benign) lesions carry a low risk of malignancy 0.5–2%,3,4 and because of this small risk of malignancy, further investigation is indicated. The RCRBG classification system is similar to other systems (e.g. BI-RADS d Breast Imaging Reporting and Data System from the American College of Radiology), but in contrast to BI-RADS, it advocates needle biopsy for equivocal or probably benign (BI-RADS 3) findings to achieve a definitive diagnosis, rather than short-term follow-up, which has been shown to be associated with significant psychological morbidity.5 Percutaneous needle breast biopsy is routinely performed in centres for category 4 findings (suspicious of malignancy) and category 5 findings (highly suspicious of malignancy) lesions due to

0009-9260/$ – see front matter Ó 2010 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.crad.2010.01.008

260

E.A.M. O’Flynn et al. / Clinical Radiology 65 (2010) 259–270

their greater association with malignancy d 33–50% and 90%, respectively.4 In some cases, particularly very large cancers, needle biopsy is performed not simply to prove the presence of malignancy, but to characterize the lesion histologically and to obtain information that is important in planning overall oncological management. This includes histological type and grade, basal subtype, hormone and HER2 receptor status, and genetic profiling. Assessment of the axilla with ultrasound is also routinely carried out with axillary node sampling, if required, to determine nodal status preoperatively and reduce the need for re-operation if nodes are found to be positive. There has been steady development in percutaneous biopsy techniques since the early 1990s, driven by the large numbers of women undergoing assessment for nonpalpable lesions detected by mammographic screening. Fine-needle aspiration cytology (FNA) was the original method of sampling and has been used for diagnosis since the 1950s. This was followed in the 1990s by increasing use of core needle biopsy (CB). The accuracy of both has been improved by ultrasound and stereotactic guidance. Larger bore vacuum-assisted biopsy (VAB) devices then became available in the late 1990s and are now commonplace in most breast units. Studies of discomfort experienced during needle biopsy have shown no significant difference related to the type of biopsy device used.6 Informed consent is required for any procedure, but written consent is not essential for diagnostic breast needle sampling procedures. The policy regarding obtaining written consent for breast interventional procedures is decided locally in accordance with hospital policy. The main risks associated with breast biopsy are bleeding and haematoma formation, post-procedure pain or discomfort. Anticoagulants are a relative contraindication and patients are advised to stop them before biopsy is undertaken if this is clinically possible. Lignocaine is used as a local anaesthetic for core biopsies and may be offered for FNA. In many centres lignocaine combined with 1:200,000 adrenaline is used as this has been shown to reduce bleeding and bruising. Lignocaine without adrenaline may be preferred for the skin to minimize the risk of local skin necrosis, particularly in elderly patients. A longer-acting anaesthetic, such as bupivicaine in combination with lignocaine with or without adrenaline is used in some centres for VABs. This article reviews the different types of breast biopsy devices currently available together with various localization techniques used, focussing on their advantages, limitations, and current controversial clinical management issues.

Types of sampling Fine needle aspiration (FNA) FNA is the most basic and inexpensive form of sampling using 18, 20, or 22 G disposable needles. The needle is attached to a 20 ml plastic syringe (to obtain sufficient

suction) with or without extension tubing. The importance of aspiration for increasing the cell yield during diagnostic FNA is well known.7 Flushing the needle may further increase the yield, but compromise the quality of the specimen and is not routinely practiced in most centres.8 Once the sample has been obtained, it is smeared onto glass slides and given to the cytopathologist as a ‘‘dry’’ slide or a ‘‘wet’’ slide fixed in formalin. It is a simple technique, well tolerated by patients, and can be done in almost any setting, with or without image guidance (if the lesion is palpable). It offers the possibility of rapid diagnosis if immediate cytology reporting is available, and its best application is in differentiating between cystic and solid lesions. It is currently the technique of choice for sampling axillary lymph nodes because of the theoretical risks of using automated core biopsy in the axilla. However, core biopsy of the axilla is now advocated by some centres. FNA of symptomatic simple cysts is commonly performed and the aspirate is discarded if there are no atypical features on imaging or evidence of blood staining of the aspirate. Aspiration of abscesses is more effective using a larger bore needle (e.g. 16–12 G).9 The most significant limitations of diagnosis using FNA are related to oncological management and include unreliability of hormone receptor status measurement, inability to distinguish between invasive and non-invasive disease, as well as inability to reliably confirm the benign nature of calcification. There is a variable frequency of insufficient samples for image-guided FNA ranging from 8.5–46%. The rate of insufficient samples is higher for FNA under stereotactic (39.9%) than ultrasound guidance (8.5%) and is higher for calcifications (46.1%) than masses (26.6%). The potential diagnostic yield of FNA is markedly reduced when there is diffuse thickening in the breast. The absence of an on-site cytopathologist is associated with an insufficient sample in 31.2%, compared with 14.5% if a pathologist is present.10,11 As a result, FNA is no longer recommended as a routine technique for breast diagnosis.

Core biopsy (CB) Image-guided CB of the breast became widely practised following the publication by Parker et al. in 1993.12 A 14 G spring-loaded, automated needle with a long excursion is used allowing small cylinders of tissue to be cut and collected within the notch of the needle. Although it has been shown that diagnostic accuracy increases significantly with increasing needle size,13 in everyday practice 14 G needles generally give excellent results. A 16 G needle may occasionally be found helpful if the breast tissue is particularly dense causing difficulty in positioning the needle. In addition to the needle size, the biopsy device also influences the size and quality of the tissue specimen14 d a long throw excursion of 22 mm achieves the best results. Some devices allow the throw to be shortened to 15 mm. After acquisition of each core sample, the needle has to be removed from the breast to obtain the specimen and be re-inserted for further samples. These subsequent samples

261

E.A.M. O’Flynn et al. / Clinical Radiology 65 (2010) 259–270

are frequently composed of blood, as destruction of the breast architecture and focal haemorrhage occurs.15 It is a particular problem in stereotactic biopsy. One attempt to address this is with the coaxial system for breast biopsy.16 However, despite being used commonly in other fields of radiology, there has been limited uptake of this technique in the evaluation of breast lesions. The optimal number of specimens necessary to achieve a reliable histological diagnosis has been debated and varies according to the mammographic and ultrasound appearances, with fewer passes required for solid lesions compared with microcalcifications.17,18,19 For stereotactic CB, diagnostic sensitivity is improved by increasing the number of cores taken to six or more. For ultrasound-visible masses, some reports suggest that greatest accuracy is achieved with three CB specimens.20 However, if the radiologist is >90% confident that the lesion has been adequately sampled a single pass is usually sufficient for diagnosis. 21 The success rate is largely dependent on the type of lesion being sampled and latest European guidelines suggest not being too dogmatic with regards to number, realizing there will be variability between cases and operators.22 Specimen radiography is essential to demonstrate the removal of representative calcifications from suspicious microcalcification clusters.23

False-negative results illustrate one of the limitations of CB. False-negative rates for ultrasound-guided CB range form 016,24 to 9%,25,26 with an overall false-negative rate of 0.4% based on 3380 biopsies.25 This is comparable with the mean cancer miss rate for needle-localized open breast biopsy of 2% (range 0–8%).28 The most common cause of false-negative results is inaccurate tissue sampling. This is dependent on the lesion position, size, mobility, and type; the size and density of the breast; operator experience; and patient compliance. A further limitation is that the CB sample may provide incomplete characterization of the histological findings and lead to underestimation of the presence of disease. This occurs most commonly with lesions such as microcalcification or possible complex sclerosing lesions requiring stereotactic guidance. Reported underestimation rates for high-risk lesions range from 3.429 to 100%.27 The finding of atypical ductal hyperplasia (ADH) on CB is associated with the presence of ductal carcinoma in situ (DCIS) or invasive malignancy in up to 56%.29,30 DCIS is upgraded to invasive malignancy at open surgery in between 16% and 55.5%15,30 of cases as shown in Table 1. Underestimation has also been reported in radial scars, papillary lesions, lobular carcinoma in situ (LCIS), and phylloides tumours; surgical excision or repeat biopsy is recommended in these cases or

Table 1 Breast biopsy studies with surgical correlation [14 G ultrasound (US)-guided, US vacuum-assisted biopsy (VAB) and stereotatic-VAB] Reference

Method

No lesions (mal/hr/ben)

Helbich et al. 2004 15

Simon et al. 2000 40

14 G US CNB 14 G US CNB 14 G US CNB 14 G US CNB 14 G US CNB 14 G US CNB 14 G US CNB US-VAB

Cho et al. 2005 26

US-VAB

61 (NG/NG/NG) 151 (56/1/94) 399 (161/16/222) 128 (59/3/61) 1279 (769/NG/NG) 561 (114/19/428) 1352 (671/86/304) 71 (18/1/52) 416 (65/15/336) 52 (NG/52/NG) 406 (76/11/319) 223 (79/26/123) 318 (NG/NG/NG) 851 (73/135/643) 1223 828 395

Liberman et al. 1998 24 Memarsadhegi 2003

27

Pijnappel et al. 2004

25

Dillon et al. 2005 Cho et al. 2005

29

26

Schueller et al. 2008

Grady et al. 2005

30

41

US-VAB

Cassano et al. 2007 42

US-VAB

Zuiani et al. 2007 44

Stereo 11 G-VAB Stereo 11 G-VAB Stereo 11 G VAB Stereo 11 G-VAB 9 G-VAB

Pfarl et al. 2002 45 Burak et al. Jr. 2000 46 Lourenco et al. 2007

47

mal, malignant; hr, high risk; ben, benign; NG, not given.

Surgical verification 61 23 260 89 NG NG 1061 NG NG 47 NG 171 214 201 239 168 71

False negative rate (%) 0% (0/61) 0% (0/56) 3.1% (5/161) 9% (8/89) 1.7% (13/769) 3% (4/128) 1.6% (11/671) 5.2% (1/19) 1% (1/69) NG 0.6% (2/319) 1.3% (1/78) 3.3% (7/214) NG NG

Underestimation rate NG NG 100% (16/16) 33.3% (1/3) 3.4% (26/769) 55% (12/22) 31.4% (27/86) NG 36% (8/22) 0% (0/29) 2.6% (2/76) 17.2% (16/93) NG 11.8% (16/135) 28.6% (48/168) 25.4% (18/71)

262

E.A.M. O’Flynn et al. / Clinical Radiology 65 (2010) 259–270

Table 2 Comparison of sampling choice for specific lesion type Lesion type

Fine-needle aspiration

Core biopsy

Vacuum-assisted biopsy

Mass Microcalcification Asymmetric density Axillary lymph node

þ

þþþ þ þ þ

þ þþþ þþþ

þþþ

cases where there is discordance between imaging and histologic findings. Seeding of benign and malignant cells away from the target lesion has been reported in 37% of cases after 14 G large CB.31 Tissue seeding was seen more often if the time between needle biopsy and open biopsy was less than 15 days (42 versus 15% with a time interval of more than 28 days) suggesting that tumour cells do not survive displacement.32 In modern practice diagnostic surgical procedures are very rarely required for the evaluation of breast lesions because a non-operative diagnosis can be obtained resulting in a significant cost saving for management of patients with benign and malignant disease. Compared to surgical biopsy, it is a short procedure that does not deform the breast, is associated with minimal residual skin scarring, and importantly, no residual mammographic scarring. 33,34

Vacuum-assisted biopsy (VAB) VAB was developed to address the limitations of CB and FNA. It addresses the need for larger volumes of tissue for histological examination and allows biopsy of lesions that are difficult to sample (e.g., microcalcifications and distortions). Table 2 illustrates the difference in choice of sampling method for specific types of lesion. VAB is powered with suction and a rotating cutter, which obtains multiple samples from the lesion. The vacuum draws tissue into an aperture in the probe where it is separated from the surrounding breast tissue by the

rotating cutter and transported to the specimen port for collection without removing the needle from the biopsy site. This enables multiple samples to be taken by rotating the shaft of the needle after a single insertion. Six cores are normally taken on a routine 360  rotation and larger needle sizes, up to 7 G, allow removal of the lesion in a piecemeal fashion through a 2–3 mm skin incision that does not require surgical closure. Several devices are available, each with their own minor technical differences. They are adaptable to be used under either ultrasound, stereotactic, or magnetic resonance imaging (MRI) guidance.35 The MammotomeÔ (Ethicon Endo-Surgery, Inc. CN, Ohio, USA) was the first vacuumassisted device and has been marketed since 1995. Technological developments have seen the production of the VacoraÔ (Bard Biopsy Systems, Tempe, AZ, USA), AtecÔ (Hologic, Inc. Bedford, MA, USA), EvivaÔ (Hologic, Inc. Bedford, MA, USA) and EnCorÔ (SenoRx, Inc. Irvine, CA, USA; Fig. 1) as alternative devices. All work on the same basic principle described above with subtle individual differences outlined in Table 3. The most important feature of the VacoraÔ is that it needs to be removed after each sample has been taken, negating the single insertion advantage of the other devices. A particular advantage of the EnCorÔ and EvivaÔ devices is that the biopsy needle is set off-centre, which enables accurate localization and access to lesions close to the chest wall. VAB enables more accurate and larger volume tissue sampling than core-needle biopsy. Standard VAB is generally performed with a 10 or 11 G needle, however, 8 and 7 G needles are used for larger lesions and therapeutic excisions. The volume of tissue removed through the probes varies considerably and is dependent on the number of cores taken. Each core volume, however, is significantly higher than that obtained from a 14 G large CB as Table 4 shows. Figs. 2 and 3 show a comparison in size between sampling devices and core samples obtained. Relative to CB, VAB provides larger specimens, offers a higher calcification retrieval rate, is less sensitive to

Figure 1 EnCorÔ vacuum-assisted biopsy device.

263

E.A.M. O’Flynn et al. / Clinical Radiology 65 (2010) 259–270 Table 3 Comparison of the attributes of vacuum-assisted mammotomy systems Attribute

Ethicon MammotomeÔ

Bard VacoraÔ

Hologic Atec Ô

Hologic EvivaÔ

Senorx EnCorÔ

Drivers required Command unit Vacuum adjustment Manual vacuum control Needle gauge Multiple core retrieval Cutting method Needle sharpness Core sample size Volume of tissue per minute Open or closed tissue collection Smaller chamber size of choice Speed of tissue retrieval Needle rotation Lavage Programmable functions Biopsy site marker system Local anaesthetic function Probe offset

Separate for US, x-ray and MRI Same for all No Yes 11 and 8 Yes Rotating þþ þþ þþ Open

Same for all Self-contained No No 14 and 10 No Rotating þ þþ þ Open

Same for all Various options Different units Yes 12 and 9 Yes Rotating þþ þþ þþþ Closed

Same for all Various options Different units Yes 12 and 9 Yes Rotating þþ þþ þþþ Closed...