PURPOSE: To determine the optimum numberof specimens to obtain at stereotaxic core breast biopsy.

MATERIALS AND METHODS:
biopsies were performed in 145 mammographically evident le- sions by using a dedicated
table with patients in the prone position. Samples were obtained with an automated gun and
a 14-gauge needle. Indications for biopsy were calcifications (n = 53) and masses (n =
92). Three to 11 (mean, 5.4) core biopsy specimens were obtained per lesion and were
analyzed separately.

RESULTS: Diagnostic material was obtained
in the first specimen in 102 (70%) of the 145 lesions. Obtaining two, three, four, five,
and six core specimens yielded a diagnosis in 117 (81%),129 (89%),132 (91%),137 (94%), and
140 (97%) of the 145 lesions, re- spectively. Obtaining five specimens yielded a diagnosis
in 46 (87%) of the 53 calcifications and 91 (99%) of the 92 masses. Obtaining six
specimens resulted in a diagnosis in 49 (92%) lesions evident as calcifications but did
not improve the yield on masses.

CONCLUSION:Stereotaxic 14-gauge core biopsy achieved a 99% diagnostic yield with five specimens for masses. Additional specimens may be necessary to diagnose some calcified lesions.

Stereotaxic core biopsies were performed with a dedicated table (Lorad DSM, Danbury,Conn) with patients in the prone position. A 14-gauge needle (Bard Urological, Covington,Ga) was used with an automated gun with either 22 mml throw (Pro-Mag 2.2 gun; Manan Medical Products, Northbrook, 111) or 23 mm throw (Biopty gun; Bard Urological). A total of 785 core specimens were obtained from 145 mammographically evident lesions; all lesions were considered indeterminate or suspicious for carcinoma and measured a minimum of 5 mm at the longest dimension.


Indications for biopsy included the presence of an uncalcified mass (n = 83; 57%), mass with calcification (n = 9; 6%), and calcifications without mass (n = 53; 37%). Three to I 1 (mean, 5.4) core biopsy samples were obtained from each lesion; five samples were obtained from each of 124 lesions. More than five samples per lesion were obtained from 18 lesions, which included two masses for which it was thought that patient movement during the acquisition of the initial core specimens decreased the likelihood of there being diagnostic information in those cores and 16 calcified lesions for which radiography of
the initial core biopsy specimens revealed little or no evidence of calcification. Fewer than five samples per lesion were obtained in three lesions (two masses and one lesion evident as calcifica- tions).

The pattern of obtaining core biopsy samples depended on the morphologic
characteristics of the lesion. For masses, the initial core biopsy sample was obtained from the central point of the lesion; images obtained before and after the needle was inserted documented accurate needle placement. Subsequent passes were made from the 12:00, 6:00, 3:00, and 9:00 positions of the lesion without obtaining additional images, and the variable distances from each subsequent pass to the central point of the lesion depended on lesion size. For calcifications, there was more variability in needle placement and in the number of images obtained. If the calcifications were heterogeneous in mor- phology, effort was made to target the most suspicious area for needle biopsy. If the calcifications were tightly clustered, the initial pass was obtained either by tar- geting the center of the cluster or by selecting a particularly distinctive calcification that could be reliably discerned on the two stereotaxic images; needle placement for subsequent passes was determned according to the geographic distribution of calcium in the breast parenchyma.


Radiography of the specimens was performed as previously described (7) on all lesions evident as calcifications without mass. If minimal or no evidence of calcification was seen on radiographs of the first five specimens, additional stereotaxic images were often obtained to retarget an area of calcification. Additional specimens were then obtained and imaged to maximize the likelihood that diagnostic material would be obtained.


Core biopsy specimens were placed separately in numbered containers of 10% neutral buffered formahn. Histopathologic analysis was performed of each specimen. A diagnosis was considered to have been made if a specific histopathologic entity concordant with the mammographic appearance was identified at pathologic examination. For lesions evident as calcifications without mass, calcifications had to be identified by means of radiography
of the specimens or histopathologic analysis. Hemorrhage was reported at histopathologic analysis if the volume of blood clot exceeded the volume of tissue in the specimen.

Histopathologic analysis revealed carcinoma in 45 (319'o) of the 145 lesions, atypical hyperplasia in 12 (8%), and specific benign entities in 83 (57%). In five cases (3%) the material obtained from biopsy was insufficient for diagnosis. In the 140 lesions for which diagnoses were made with stereotaxic biopsy, diagnostic material was present in a single specimen in 17 (12,7o) and in two or more specimens in 123 (88%). Histopathologic diagnoses were made in 524 (67%) of Ihe 785 core biopsy specimens.

Diagnostic material was present in the first specimen in 102 (70%) of the 145 lesions. Acquisition of two, three, four, five, and six specimens resulted in diagnoses of 117 (8lYo), 129 (89'7o), 132 (91%),137 (94%), and 140 (97%) of the 145 lesions, respectively (Figure, part a).

The first specimen yielded diag- nostic material in 77 (849o) of the 92 masses;
acquisition of two, three, four, and five specimens resulted in diagnosis of 84 (91%), 90 (98,7o), 90 (98%), and 91 (99%) of the 92 masses, respectively (Figure, part b). Additional core biopsy samples were obtained in two cases and did not provide diagnoses not obtained with prior specimens. In one case of a new 6-mm-diameter speculated mass detected at mammography, five stereotaxic core biopsy specimens yielded benign fibrofatty tissue. Because this was not consistent with the mammographic appearance, surgical excision was recommended and revealed a 10-nun-diameter invasive ductal carcinoma of the tubular subtype.

Of 53 lesions evident as calcifictions without mass, the first core biopsy specimen yielded diagnostic material in 25 (479o); acquisition of two, three, four, five, and six specimens resulted in diagnoses in 33 (621/o), 39 (74-7.), 42 (799,,), 46 (87%), and 49 (92%) of the lesions evident as calcifications, respectively (Figure, part c). Sixty-three additional specimens were obtained from 16 calcified lesions but did not yield any new diagnoses. In four calcified lesions that measured 8-22 mm (mean, 13 mm) at the longest dimension, stereotaxic core biopsy yielded no calcifications at histopathologic analysis.
Six to 10 (mean, seven) samples were obtained from each of these four lesions. Surgical excision was recommended in all four cases and was performed in two. In- vasive ductal carcinoma with calcification was diagnosed in one case and fibrocystic change with calcification was diagnosed in the other. Two patients declined surgery and will undergo follow-up mammography.

Radiography of the specimens, which was performed in all 53 lesions evident as calcifications without mass, revealed calcium in at least one specimen in 48 (989o) of the 49 lesions for which diagnostic material was obtained. Calcium was not observed at specimen radiography in any of the four calcified lesions for which material was inadequate to yield diagnosis. Radiography of the specimen revealed calcium in the first sample in 31 (589o) of the 53 lesions. Obtaining two, three, four, five, six, seven, and eight or more samples revealed radiozraiphic evidence of calcium in at least one biopsy specimen in 39 (74%), 41 (77%), 43 (81%), 45 (85%), 47 (89%), 48 (91%), and 48 (91%) of the 53 lesions, respectively.

Histologic evidence of hemorrhage was reported in 63 (8%) of the 785 core specimens. Hemorrhage was present in the first, second, third, fourth, and fifth core specimens in four (3%) of 145 lesions, six (4%) of 145 lesions, nine (6%) of 145 lesions, 14 (10%) of 143 lesions, and 16 (11%) of 142 lesions, respectively; overall, hemorrhage was detected in 49 (7%) of the 720 first through fifth core specimens and in 14 (22%) of the 65 sixth or higher core specimens. Diagnostic material was present in only eight (13%) of the 63
specimens in which hemorrhage was noted at histologic analysis. Clinically significant hemorrhage was not seen in any patient. No infectious complications were identified in any case.

The utility of stereotaxic core biopsy for obtaining tissue diagnosis of
mammographicaby evident suspicious lesions has been demonstrated in numerous studies (1-5). However, the optimum number of passes to be performed has not yet been deter- mined. It is difficult to glean this in- formation from previous studies in the literature because of the variety of needle sizes, automated guns, and stereotaxic equipment used, as well as the variability in the number of passes made.

In pioneering work in the field in 1990, Parker and colleagues (1) found agreement between the histologic results of surgical and stereotaxic biopsy in 89 (87%) of 102 lesions, with three to four passes per lesion. However, they used a variety of needle sizes: 18-gauge needles were used in 65 cases, 16-gauge in nine, and 14-gauge in 29. Two patients underwent biopsy with a short (11.5-mm) rather than a long (23-mm) excursion gun; 30 (29%) of 103 patients underwent biopsy with add-on equipment while in an upright position rather than with dedicated equipment while in a prone position. In 1991, Parker et al (2) performed a study with a 14-gauge needle, a long excursion gun, dedicated equipment with patients prone, and three to four passes per lesion to demonstrate
histologic agreement between stereotaxic and surgical biopsy results in 98 (96%) of 102 cases. In 1991, Dowlatshahi and colleagues (3) performed 250 stereotaxic core biopsies with dedicated equipment and patients in the prone position. They used a 20-gauge needle and two to three passes per lesion with a short excursion gun in 120 cases and a long excursion gun in 130. Concordance between the results of core and surgical biopsy was seen in 167-173 (67%- 69%) cases, with the range depending ,on whether the lesions that were determined to be benign at surgery and for which the core biopsy specimen was interpreted
as atypical or suspicious were scored as concordant or discordant.

Dronkers (4) found histologic agreement between the results of stereotaxic and surgical biopsy in 48 (90%) of 53 cases when he obtained two core biopsy specimens per lesion. He used an upright stereotaxic unit, an 18-gauge needle, and a short (10- mm) excursion gun. Elvecrog and colleagues (5) used equipment similar to that used in Parker et al's second series (2) and found histologic agree- ment in 96 (96%) of 100 cases; they obtained "at least five core specimens"; in all but one case, although the exact number of specimens obtained for each lesion and the number of lesions for which diagnoses were made with fewer than the maximal number of core specimens is not known. The variability of results in these different series may be related at least in part to variability in the
equipment used, as well as to the number of specimens obtained.

In our series, in which a 14-gauge needle, dedicated equipment with patients in the prone position, and a long excursion gun were used, five core samples were adequate to obtain tissue diagnoses in 99% of the masses. For some lesions evident as calcifications without mass, five core samples were not adequate. However, one can increase the likelihood of obtaining diagnostic material in these cases by performing radiography of the specimen during the procedure (7). As has been previously demonstrated, the likelihood that a specific histopathologic diagnosis will be obtained in a particular specimen is 81 To if calcium is presentand 387o if calcium is absent on radiographs of the specimen (7).
In this series, all cases for which radiog- raphy of the specimen showed calcium in at least one core sample yielded diagnostic material.

Stereotaxic biopsy of calcified lesions may be particularly challenging. The diagnostic yield of the first core biopsy specimen, for example, was 84% for masses but only 47,7o for lesions evident as calcifications. The lower yield for calcified lesions likely reflects the high degree of precision required to target microcalcifications, as well as the difficulties encountered in trying to identify the same microcalcification on two stereotaxic images.

There are few published reports documenting experience with obtaining six or more core biopsy specimens per lesion. Our data suggest that increasing hemorrhage in successive core biopsies, although not of clinical significance, may limit the diagnostic utility of obtaining a large number of specimens in a small area. Additional work with six or more core samples per lesion is necessary to further clarify these issues.

Although we did not assess the procedure time requirements prospectively, we estimate that the average time required for a radiologist to perform these biopsies was approxi- mately 20 minutes. This is similar to Parker's (8) estimate of an average of 15 minutes to make at least five passes per lesion with a digital stereotaxic unit. The use of digital equipment, which was previously reported to diminish the time requirements of needle localization procedures by almost 50% (9), is also helpful in minimizing the time required for stereotaxic biopsy.

In light of our experience, we have adopted the following protocol for 14-gauge
stereotaxic core breast biopsy. Five samples are obtained for all masses. For calcified lesions, a minimum of five samples are obtained and specimen radiography is performed; if no calcifications are identified, additional samples are obtained and imaged until calcium
is identified on radiographs or until specimens appear to be composed predominantly of hemorrhagic material (usually within 11 specimens), to increase the likelihood of making a histopathologic diagnosis.

1.From the Breast Imaging Section, Depart- ment of Radiology (L.L., D.D.D.,
A.F.A., B.M.D., L.E.H.) and the Department of Pathology (P.P.R.), Memorial Sloan-Kettering
Cancer Center, 1275 York Ave, New York, NY 10021. Received February 7,1994; revision
requested March 8; revision received April 8; accepted April 25. Address reprint
requests to LL.
2 Current address: The Jacqueline M. Wilentz Comprehensive Breast Center, Monmouth Medical
Center, Long Branch, NJ. ©RSNA, 1994

Acknowledgments: The authors thank David C. Perlman, MD, and Ann Barrow, BA, for in- valuable assistance and Lynda Tiangco, Dolores Brown-Moody, Youngduk Paik, Regina Sleavin, and Mary De Vito for outstanding technical support.

References

1. Parker SH, Lovin JD, Jobe WE, et al. Ste- reotaxic breait biopsy with a biopsy gun.
Radiology 1990; 176:741-747.

2. Parker SH, Lovin JD, Jobe WE, Burke Bj, Hopper KD, Yakes WF, Nonpalpable breast
lesions: stereotactic automated large- core biopsies. Radiology 1991; 180:403-407.
3. Dowlatshahi K, Yaremko ML, Kluskens LF, Jokich PM. Nonpalpable breast lesions: findings
of stereotaxic needle-core biopsy and fine-needle aspiration cytology. Radiol- ogy 1991;
181:745-750.

4. Dronkers DJ. Stereotaxic core biopsy of breast lesions. Radiology 1992; 183:631-634.

5. Eivecrog EL, Lechner MC, Nelson MT. Nonpalpable breast lesions: correlation of
stereotaxic large-core needle biopsy and sur- gical biopsy results. Radiology 1993; 188:
453-435.

6. Jackson VP, Reynolds HE. Stereotaxic needle-core biopsy and fine-needle aspira- tion
cytologic evaluation of nonpalpable breast lesions. Radiology 1991; 181:633-634.

7. Liberman L, Evans WP, Dershaw DD, et al. Radiography of microcalcifications in stereo-
taxic mammary core biopsy specimens. Ra- diology 1994; 190:223-225.

8. Parker SH. Stereotactic large-core breast biopsy. In: Parker SH, Jobe WE, eds. Percu-
taneous breast biopsy. New York, NY: Raven,1993;61-79.
9. Dershaw DD, Fleishman RC, Liberman L, Deutch B, Abramson AF, Hann L. Use of digital
mammography in needle localization procedures. AJR 1993; 161:559-562.