Introduction
Percutaneous sampling of non-palpable breast lesions can be done with the guidance of ultrasound or stereotaxis. This syllabus chapter will discuss these procedures and their indications as well as contrast the appropriateness of the various modalities.

Ultrasound-Guidance
The procedures that are suitable for ultrasound guidance are:

cyst aspiration fine-needle aspiration biopsy (FNAB) large-core needle biopsy (LCNB) preoperative fine wire localization

The guidance technique is the same for all of these procedures. The method of ultrasound guidance will be discussed first, followed by descriptions and details of the individual procedures.

It is critical that the entire imaging investigation be complete prior to performance of any procedure. These techniques all produce some degree of hematoma. It is preferable to obtain magnification mammographic views prior to intervention, so the true marginal characteristics can be documented without confusion. Irregular increased soft tissue density and indistinct borders due to hematoma can cause a benign mass to appear suspicious.

Bleeding diatheses and anticoagulant therapy should be dealt with prior to large needle procedures. Antibiotic prophylaxis is not usually used.

The Technique
The basis of any ultrasound- guided procedure is the use of real-time scanning to precisely monitor the passage of the needle through the skin and breast tissue to the target lesion. This requires that the lesion be visible with ultrasound, and that the lesion can be approached in such a manner that there is no risk of the needle entering the pleural space, or in some cases, an underlying prosthesis. The size of the lesion is immaterial. With skill, any mass large enough to perceive can be accurately targeted. Whether the breast is primarily fatty or dense is also unimportant, as long as the lesion can be demonstrated.

Equipment
Any high-frequency hand-held transducer may be used. Currently, linear array probes are more commonly used for both diagnosis and intervention. To see the needle with a linear array probe requires that the needle be almost perpendicular to the beam, i.e. parallel to the surface of the probe. This requirement may necessitate the needle being inserted from a point distant to the edge of the probe, especially for deeper lesions. This makes targeting more challenging because it increases the opportunity for deflection of the needle out of the scan plane, especially with small bore needles and in more fibrous tissue. These issues are less of a problem for more superficial masses, when the needle can be inserted closer to the probe, traverse the breast tissue obliquely, and yet still be relatively perpendicular to the beam.

Execution of the Procedure
Many options exist for executing the free-hand technique. Some use a two-person approach. It is much simpler for the sonologist to scan with one hand and insert the needle with the other. With practice, intuitive hand-eye coordination develops. This allows instantaneous minor adjustments of the scanning angle and needle angle to ensure continuous maintenance of the needle and lesion within the scan plane.

Any needle procedure should conform to clean (antiseptic) principles. The sonologist should wear gloves, more for their own protection than for the patient's. The skin and probe should be prepped with antiseptic. Ideally, sterile coupling gel should be used, but the gel does not come in contact with the needle insertion site. The tray equipment used for each procedure will vary.

With any kind of lesion, but especially for small solid masses, it is helpful to confirm as accurately as possible, that the sampling portion of the needle (the tip for FNAB, and the notch for LCNB) is within the lesion and to avoid volume averaging. It is theoretically possible that the needle will appear to be in the lesion when it actually lies immediately adjacent to the mass. To check needle position, there are maneuvers the radiologist can use, but pathologists need to be aware of this pitfall.

It is preferable to avoid the nipple-areolar complex for needle procedures because of its higher pain sensitivity. Surgeons prefer to make circumareolar incisions whenever possible because scars at the areolar border are potentially inconspicuous. If, for wire localizations, it is feasible to use the areolar border for needle insertion, both surgeon and patient will be appreciative.

If a lesion is visible both with mammography and ultrasonography, it is usually preferable to perform the procedure with sonographic guidance. This is faster, less costly and allows greater patient comfort. The patient is supine and no breast compression is required. Ultrasound-guidance allows more flexibility in choosing the needle entry site; and is useful in avoiding puncture of the nipple-areolar complex. It is also an advantage in conservative surgery for cancer because the shortest route from the skin to the lesion can be used.

For sampling, the ultrasound free-hand technique allows multidirectional sampling of the entire volume of the mass, rather than being limited to the perpendicular access provided by stereotaxis.

Cyst Aspiration
Aspiration is not medically indicated for lesions that fulfill the strict sonographic criteria for a simple cyst. However, if a simple cyst is painful/tender or if it is palpable, the patient may prefer aspiration for comfort or peace-of-mind. Any lesion with internal echoes and posterior enhancement may be a complex/complicated cyst. Aspiration of these masses may allow definitive diagnosis.

In my practice, if a complicated but asymptomatic cyst is seen, I give the patient the options of aspiration or short interval follow-up.

If a cyst is completely echo free on sonography, and if the fluid obtained is along the spectrum of white/yellow/green - then the fluid may be discarded. If there are any sonographic findings such as internal echoes, ill-defined margins or wall thickening, or if the fluid obtained is bloody or gelatinous, then the fluid is sent for cytological evaluation.

There are various methods for cyst aspiration once the needle is positioned. I prefer to attach the needle to a 10 cc syringe in a light-weight plastic aspiration gun prior to inserting the needle. Once the needle is in place, flexing the finger on the "trigger" pulls back the plunger to create a vacuum. Others prefer to use a plain syringe without a gun. This becomes more awkward because the needle must be immobilized while the plunger is pulled back. Another option is to interpose connecting tubing between the needle and syringe, and to have an assistant operate the syringe. Finally, the use of a vacutainer to create negative pressure has been described. [1]

Even if high fluid viscosity from some inspissated cysts precludes their complete evacuation, it is always possible to aspirate sufficient material from the lesion to allow cytological confirmation of its cystic nature.

Why Use Ultrasound For Palpable Cysts?
Palpable cysts are often successfully aspirated with only finger-guidance, and usually this suffices. Even palpable lesions can be missed though, and if aspiration yields no fluid, it is tempting for clinicians to conclude that the mass must be solid. Because of the far different implications of a solid mass, it is prudent for clinicians to refer patients for ultrasound imaging, with or without guided aspiration, prior to rendering a diagnosis of "solid". Often, the clinically attempted aspiration has failed because the cyst has been pushed away rather than penetrated by the needle. Sometimes, the cyst may feel superficial, but actually be much deeper than anticipated. I call this "The Princess and the Pea" phenomenon. These problems are simply dealt with using ultrasound.

Certainly, when a cyst is close to the chest wall, or in an augmented breast, the increased precision of needle positioning available with ultrasound guidance is critical.

Finally, although a relatively uncommon situation, intracystic masses both benign and malignant, may go undiagnosed if aspiration is done without ultrasound. It is feasible that a needle could be positioned within the fluid component of such a mass, fluid obtained and the palpable lump could disappear, without the solid component being sampled. Sonographic evaluation of such a mass would not overlook the solid component [2].

Fine Needle Aspiration Biopsy of Solid Masses
Tissue diagnosis is appropriate for new or enlarging solid masses. If a given mass appears benign using both mammographic and sonographic criteria then short interval follow-up without tissue diagnosis may be a reasonable alternative [3].

Incorporation of FNA in the diagnostic pathway is a team endeavor. The clinical impression will contribute in the case of a palpable mass. The mammographic and sonographic impressions are critical in the assessment of non-palpable lesions. The cytopathologist should be willing to make a diagnosis, given an adequate specimen. Finally the patient and her physician(s) must have an understanding of the strengths and weaknesses of the whole process.

In the case of a lesion that is probably benign by imaging criteria, the addition of benign cytology can increase diagnostic confidence and encourage the patient to choose follow-up rather than surgery.

More importantly, in the case of an otherwise benign-appearing mass, the unexpected finding of suspicious or malignant cytology can minimize the small but real possibility of a false negative imaging diagnosis. Cytology's main limitation is its inability to reliably distinguish in-situ from invasive cancer. For this reason, histologic biopsy is preferred to confirm invasive malignancy pre-operatively, to allow definitive surgery, including node dissection.



If a solid-appearing mass is somewhat likely to be a complicated cyst, FNA can be done first, and is preferable to core biopsy since accurate diagnosis and complete evacuation can be achieved with a much smaller and cheaper needle. Furthermore, the confident diagnosis of cyst after FNA eliminates the need for short-interval follow-up.

Reliable FNA requires accurate targeting of the lesion, aggressive sampling to obtain adequate tissue for diagnosis, and expert cytopathologic interpretation. Furthermore, because false negatives do occur, the radiologist must be prepared in a small percentage of cases, to recommend histologic biopsy (either percutaneous or surgical) based on mammographic and/or sonographic findings, even if cytology is benign. For this reason it is preferable that the sonologist interpreting the scan and performing the FNA procedure is also skilled in mammographic diagnosis, so that a final opinion encompassing input from all the modalities can be offered. Ideally, the radiologist should review the imaging when the cytology is available to ensure concordance between the two, and then offer recommendations regarding surveillance or other management.

The Technique
There is no single correct way to perform a fine needle biopsy. I prefer to use a 20 gauge needle with a 10 cc syringe and aspiration gun. Three separate needle passes, each with multiple to-and-fro excursions provide sufficient cellular material in the vast majority of cases. In my experience, the incidence of inadequate specimens is less than 2%. The most reasonable explanation for this occurrence is that some lesions are too densely fibrotic to shed cells into the needle.

Others use different needle sizes and a variety of methods. Aspirating guns are available in many styles. An assistant can also generate negative pressure if connecting tubing is positioned between the needle and syringe. Finally, some experts use a "no aspiration" biopsy technique; the needle hub is left open while the excursions are made. No negative pressure is used and the material enters the needle shaft by capillary action [4, 5]

Once the material is in the needle shaft, there may be a small amount of pinkish material visible in the hub of the needle. Rarely (more often with cancers) will bloody fluid be voluminous enough to enter the syringe.

Handling of the specimen will be dictated by the preference of the cytopathologist. The material can either be expressed onto slides and either air-dried or fixed in alcohol, or be rinsed into saline solution - later to be spun down and filtered for examination.

After each needle pass and on completion of the procedure, the patient should be instructed to firmly compress the area with clean gauze, including the needle insertion site, the needle path and the mass itself. She should be told to anticipate a visible bruise. The vast majority of patients can return to their normal activities and require no analgesics. It is prudent to ask them to avoid ASA, and to use acetaminophen if necessary for pain after the procedure.

How Radiologists Use Cytologic Interpretation
Radiologists expect that the cytology will be reported as definitively malignant, suspicious for malignancy, benign but nonspecific, definitively benign, or inadequate. The definitive benign category includes fibroadenomas, cysts, abscesses and lymph nodes. Lesions diagnosed as malignant or suspicious should go on to histologic biopsy (percutaneous or surgical) regardless of their imaging characteristics. Lesions that are highly suspicious on imaging should logically be biopsied initially with large-core techniques, since in these circumstances, a benign cytology would be sufficiently reassuring to allow surveillance.

If a mass appears benign on imaging and yields definitive or nonspecific benign cytology, then short-interval follow-up is a safe alternative to surgery.

However, if a mass appears indeterminate on imaging - a nonspecific benign cytology result should prompt a histologic biopsy. If an indeterminate mass on imaging yields a definitive cytological diagnosis of fibroadenoma, then surgery can be avoided, provided the patient will comply with imaging follow-up. Clearly, the threshold for recommending surgery will vary by institution and region, by medicolegal climate, by acceptance of the technique by patients and their referring doctors and by the experience of the radiologist and pathologist. It is highly recommended that accurate records of surgical results be obtained to allow cytologic-histologic correlation for patients having surgery. Patients not having surgery require documentation of stability on follow-up imaging. Currently, I recommend repeat measurements at 6, 12, 24 and 36 months. This is based on Sickles protocol [6], which is for lesions that are "probably benign" on mammography, but have not been subjected to percutaneous sampling.

For the small number of specimens that are hypocellular, I am comfortable to regard them as nonspecific benign, as long as I am certain that targeting was accurate. By doing so, this would not delay surgery for an otherwise suspicious or indeterminate mass, but would not prompt surgery for an otherwise benign appearing mass.

Limitations of Cytology
Although cytology obtained with ultrasound guidance has proven highly accurate and is potentially tremendously helpful in the diagnostic sequence, its limitations must be acknowledged and understood to avoid errors.

The concept of masses that do not shed cells into the needle has been introduced above. This is a reasonable explanation for hypocellular specimens and can contribute to nonspecific benign cytology. The vast majority of these lesions will be relatively fibrotic and the most common is the hyalinizing fibroadenoma. In contrast, most cancers shed their cells readily and are easily and accurately diagnosed with FNA. The notable exception to this generalization is lobular cancer. It is for this reason that nonspecific benign cytology should not take priority over the imaging impression when masses appear suspicious or indeterminate at imaging, and why masses that are borderline or suspicious on imaging should be initially sampled with LCNB.

Cytology cannot reliably distinguish fibroadenomas from phylloides tumors. The diagnosis of phylloides can certainly be suggested if there is a predominance of the stromal component, but this is not always present. The distinction is relevant because 10-15% of phylloides tumors are malignant. Thus, even cytology definitive for fibroadenoma should not delay surgery for a mass otherwise suggestive of a phylloides tumor. There are no definitive criteria and considerable overlap can occur. Sonographic criteria include size greater than 3 cm in greatest dimension, prominent cysts within a solid mass, or documented growth of greater than 20% in each of three dimensions in a six month period. [7] The latter criterion is conservative since fibroadenomas can fluctuate within this range during a menstrual cycle.

A further limitation of cytology is the inability to definitively diagnose hamartomas. Because these adenolipomas are composed of normal breast tissue elements, it is not surprising that partial sampling without architecture could be interpreted as nonspecific benign or even fibroadenoma. Hamartomas should not require sampling if they display the pathognomonic mammographic finding of fat within the mass. Not all do, however, so some will present for needle biopsy. Even those with intrinsic fat on mammography may demonstrate growth. In these cases they may be referred for tissue diagnosis simply to exclude malignant findings.

Finally, in the unlikely coincidence of a cancer that appears benign both mammographically and sonographically and yields benign cytology, there would presumably be significant interval growth demonstrable on the six month follow-up examination, allowing a delayed, albeit minimally so, diagnosis.

The sensitivity and specificity of FNA reported in the literature vary, but are as high as 99% and 100% respectively. Each radiologist must determine their own accuracy by meticulous practice audits. My own data have shown 95%sensitivity and 93% accuracy [8].

Unfortunately, it is well known that there is a paucity of trained cytopathologists in many parts of North America. This is part of the reason for more widespread use of core needle sampling instead of the less invasive FNA in the U.S. relative to Europe and parts of Canada. The difference in cost is also not trivial. The cost of core needle biopsy in the U.S. ranges up to $2000.00 [9, 10] , and in a managed-care environment there should be encouragement to train more cytopathologists. Although not representative of U.S. costs, combined FNA and cytology in British Columbia costs the equivalent of $63.00 U.S.

Ultrasound-Guided Large Core Needle Biopsy
If expert cytopathologic consultation is not easily obtained, then larger core needles can provide specimens interpretable by non-specialized pathologists. Needle sizes up to 11 gauge are commonly used. Hematoma is a more likely complication than with smaller bore needles. Needle track seeding has also been reported [11]

There are two styles of large core needle in widespread use: the spring-activated device (16, 14, 12G), and the directed vacuum assisted device, or DVAB (11G). The cost differential is considerable: spring activated needles cast $20- 30, and DVAB needles cost $250. The advantage of the larger needle is a lower incidence of undersampling. There is a higher risk of bleeding.

The technique for core needle biopsy is similar to that described above. Local anaesthetic should be used generously, Because of the large needle size, a scalpel nick must be made in the skin. For spring-activated needles, either a short (15mm) or long (22mm) throw can be chosen. One must ensure that the trajectory of the needle will be parallel to the chest wall, and that when in the post-fire position, that the tip will not injure an adjacent structure, or cause pneumothorax. Ultrasound guidance is used to position the needle tip proximal to the mass. The ideal position of the needle tip must be determined according to the size of the lesion, the length of throw of the needle, and the size of the sampling notch. The number of cores required should be determined in conjunction with the pathologist, and after careful audit of the institutions previous experience. Commonly, five or more cores are obtained, but one can be sufficient.

For DVAB, the needle is placed posterior to the mass, to allow the lesion to be sucked into the chamber when suction is applied.

When a lesion is small, there is a possibility that it will be inconspicuous after biopsy, or that the biopsy might remove the entire lesion. In the case of cancers, it is important to mark the biopsy site. Commercially available clip devices have long been available for 11-gauge devices. Recently, clip marking devices have become available for 14-gauge needles. One can also use home-made solutions; I use pediatric titanium vascular clips. [12]

Stereotactic Biopsy
Stereotaxis is required for percutaneous biopsy of calcifications, since these are not usually adequately seen with ultrasound. It is also required for masses that are not visible on ultrasound, or for masses that are visible, but would be more safely sampled stereotactically than sonographically. Stereotaxis uses the principle of parallax to determine the depth of needle placement when the needle is inserted directly over a lesion.

The stereotactic machine can be a dedicated (not used for anything else) table where the patient lies prone with her breast pendant through an opening, and the compression device and needle holder beneath. Alternatively, a stereotactic device can "add-on" to a mammography machine, allowing the patient to be sitting or decubitus during the procedure. This is also more economical because the machine can be used for regular mammography when not in use for biopsies.

Stereotactic units can be film-based or digital. Film-based units increase the time required for the procedure. Any images that are obtained must be developed in the processor, which typically takes 90 seconds. Digital stereotactic units use digital image receptors instead of x-ray film, so whenever an x-ray picture is made during the exam, the image can be viewed on the monitor in the room within seconds.

After the patient is placed in the machine, a scout view is obtained to confirm that the patient is properly positioned, and that the lesion is visible within the small area included on the stereo image. Next, two x-ray images are taken at approximately 30 degrees apart, 15 degrees on either side of neutral. These are displayed side-by-side on the monitor as the "stereo pair." The radiologist places a cursor on the center of the lesion on both views, and this determines the location of the lesion in the "X" and "Y" planes. By calculating the shift of the lesion on the two angled views, the machine determines the depth to which the needle must be inserted (the "Z" axis). The computer is programmed for the type of needle to be used, and this adjusts the position of the needle holder. That is, for a spring activated LCNB, the needle would be placed in the "pre-fire" position in the breast. The radiologist then "fires" the needle, so after deployment, the collecting notch will be at the level of the lesion. For DVAB of a mass, the needle can be placed alongside the lesion, so that when the collecting chamber is exposed and the vacuum activated, the lesion will be sucked into the notch, and cut by the cutting edge. For DVAB of an area of calcification, the needle can be centered within the area, or along side the group. The position of the collecting chamber can be rotated around the clock face, as desired. Specifically, it can be positioned away from an artery or implant, if close to the biopsy site, to minimize trauma during the procedure.

Pre-Operative Fine Wire Localization
If a definite diagnosis cannot be made percutaneously, then fine-wire localization and surgical excision will be necessary. This can occur when a lesion doesn't meet the imaging criteria for surveillance, yet is not amenable to percutaneous biopsy. Some lesions are sufficiently conspicuous on regular mammograms to allow fine wire placement, yet not sufficiently conspicuous on a stereo pair to allow stereotaxic biopsy. For a lesion to be targeted

stereotactically, it must be conspicuous on both images of a "stereo pair." A faint mass detected at mammography might be visible on one of the stereo pair, but obscured on the other by superimposed dense tissue. Similarly, extremely fine or diffuse calcification may not be suitable for stereotactic targeting.

Also, lesions that have shown atypical ductal hyperplasia or DCIS by core biopsy require excision, because these may be upstaged at surgery: DCIS to invasive cancer (33%) [13], and ADH to cancer (31%) [14]. Underestimation is less prevalent when more tissue is sampled, either by removing a greater number of specimens, or by vacuum-assisted needles.

As previously discussed, ultrasound guidance for localization procedures is preferable whenever feasible. This necessitates that the mass be detectable with ultrasound, and so microcalcifications usually cannot be localized this way. With ultrasound, the shortest route from the skin to the lesion can be used, and, there is an almost infinite choice of puncture sites circumferentially. As with other interventional procedures, US guidance for wire localization is particularly appropriate for masses at the periphery of the breast or in patients with prostheses.

I find that with creative patient positioning, most masses are within 2 cm of the skin, which means that ultrasound guidance can usually be used, and in virtually all cases, the masses can be skewered on the wire. In general, it should not be necessary to obtain mammographic images of the breast after US guided wire localization. Most surgeons will become comfortable with a well-labeled diagram indicating the location of the mass relative to the wire-insertion site, and the depth of the lesion. Some will prefer to have at least one conventional global image of the breast to help them with orientation. Presumably, this requirement is part of the learning curve for the surgeon, and will diminish with experience.

Specimen examination can be performed after US guided localization with either mammography or sonography [15, 16] .

Summary
Ultrasound and stereotactic guidance for interventional procedures in the breast are an important addition to the radiologist's armamentarium. The choice of imaging modality, and the sampling needle used should be based on the visibility of the lesion, the anticipated diagnosis of the lesion, the availability of cytopathologic expertise, and financial resources.

References

1. Fornage BD. Fine-needle aspiration biopsy with a vacuum test tube. Radiology 1988;169:553-554.
2. Mendelson EB, Tobin CE. Critical pathways in using breast US. RadioGraphics 1995;15:935-945
3. Sickles EA. Management of probably benign lesions. RSNA Categorical Course in Breast Imaging 1995;pp133-138.
4. Hopper KD, Abendroth CS, Sturtz KW, Matthews YL, Shirk SJ. Fine-needle aspiration biopsy for cytologic analysis: Utility of syringe handles, automated guns, and the nonsuction method. Radiology 1992;185:819-824.
5. Savage CA, Hopper KD, Abendroth CS, Hartzel JS, TenHave TR. Fine-needle aspiration biopsy versus fine-needle capillary (nonaspiration) biopsy: in vivo comparison. Radiology 1995;195:815-819.
6. Sickles EA. Periodic mammographic follow-up of probably benign lesions: results in 3,184 consecutive cases. Radiology 1991; 179:463-468.
7. Gordon PB, Gagnon FA, Lanzkowsky L. Solid Breast Masses Diagnosed as Fibroadenoma at Fine-Needle Aspiration Biopsy: Acceptable Rates of Growth at Long-term Follow-up. Radiology 2003; 229:233-238.
8. Gordon PB, Goldenberg SL, Chan N. Solid Breast Lesions: diagnosis with US-guided fine-needle aspiration biopsy. Radiology 1993;189:573-580.
9. Jackson V. The status of mammographically guided fine-needle aspiration biopsy of nonpalpable breast lesions. In: Breast imaging: current status and future directions. Radiol Clin North Am 1992;30:155-166
10. Lindfors K, Rosenquist C. Needle core biopsy guided with mammography: a study of cost-effectiveness. Radiology 1994;190:217-222.
11. Harter LP, Curtis JS, Ponto G, Craig PH. Malignant seeding of the needle track during stereotactic core biopsy. Radiology 1992;185:713-714.
12. Gordon PB. Technical Innovation: An alternative clip-marking method for use after 14-gauge large core needle biopsy of the breast; an opportunity for huge cost savings. JCAR, accepted for publication.
13. Reynolds HE, Jackson VP, Gin FM, et al. Large-Gauge Core Needle Biopsy of the Breast. The Breast Journal. Vol 2, No. 6, 1996:370-373.
14. Nguyen M, McCombs MM, Ghandehari S, et al. An Update on Core Needle Biopsy for Radiologically Detected Breast Lesions. Cancer 1996; 78:2340-2345.
15. Frenna TH, Meyer JE, Sonnenfeld MR. US of breast biopsy specimens. Radiology 1994;190:573
16. Fornage BD, Ross MI, Singletary SE, Paulus DD. Localization of impalpable breast masses: Value of sonography in the operating room and scanning of excised specimens. AJR 1994;163:569-573.