- Email: [email protected]
Tissue compression is not necessary for needle-localized lesion identification
The American Journal of Surgery 190 (2005) 580 –582
Presentation
Tissue compression is not necessary for needle-localized lesion identification Jane E. Méndez, M.D.a, Deborah ter Meulen, M.D.b, James Padussis, B.A.c, Deborah Richard-Kowalski, R.T.(R)(M)b, Manuel Raagas, M.D.b, Sidney Pollack, M.D.b, Daniel O’Connor, M.D.b, Michael D. Stone, M.D.a,* a
Department of Surgery, Surgical Oncology, Boston Medical Center, Boston University School of Medicine, 88 E. Newton St., Boston, MA 02118, USA b Department of Radiology, Boston Medical Center, Boston University School of Medicine, Boston, MA 02118, USA c Medical Student, Boston University School of Medicine, Boston, MA 02118, USA Manuscript received June 14, 2005; revised manuscript June 16, 2005 Presented at the Sixth Annual Meeting of the American Society of Breast Surgeons, Los Angeles, California, March 16 –20, 2005
Abstract Background: Tissue compression to enhance lesion visibility on radiography of needle-localized breast biopsy specimens is widely used. We hypothesized that compression is not necessary for detection of lesions on specimen radiography. Methods: Forty-nine consecutive patients underwent needle-localization biopsies of 59 mammographic targets. All specimens were radiographed without and with compression. The films were later independently reviewed and compared with preoperative mammograms by 2 surgeons and a breast-imaging radiologist. The primary end point was identification of mammographic targets in noncompressed specimen radiographs. Results: Twenty-nine targets were masses, 36 contained calcifications, and 14 contained previously placed clips. All mammographically localized lesions were identified on noncompressed views. Overall concordance for the 2 images was 100% for all 3 reviewers and 98% among reviewers. Conclusions: Tissue compression before specimen radiography is not routinely necessary for target lesion identification. © 2005 Excerpta Medica Inc. All rights reserved. Keywords: Needle-localized breast lesion; Nonpalpable breast cancer; Specimen radiography; Tissue compression
Complete removal of nonpalpable breast cancers with negative margins is required before consideration of radiation therapy. Positive or close surgical margins are found in 20% to 60% of excision specimens, resulting in re-excision procedures. Tissue compression to enhance lesion visibility on specimen radiography is widely used, but it may increase the likelihood of positive margins by flattening a spherical specimen to a disk (“pancake effect”) [2]. We hypothesized that compression is not necessary for lesion detection.
* Corresponding author. Tel.: ⫹1-617-638-8654; fax: ⫹1-617-6388081. E-mail address: [email protected]
Methods Forty-nine consecutive patients underwent needle-localization biopsies of 59 mammographic targets from May to December 2004. Mammographic wire localization was performed for 57 targets, and sonographic guidance was used for 2 targets. All specimens were radiographed first without and then with compression. The TransSpec (E-Z-E-M, Long Island, New York) specimen device was used for compression (Fig. 1). Each surgeon reviewed the specimen radiographs intraoperatively to assure target retrieval. The films were later independently reviewed and compared with the preoperative mammograms by 2 surgeons and a breastimaging radiologist. Concordance for image was defined as target lesion seen or not seen by each reviewer on both compressed and noncompressed radiographs. Concordance
0002-9610/05/$ – see front matter © 2005 Excerpta Medica Inc. All rights reserved. doi:10.1016/j.amjsurg.2005.06.016
J.E. Méndez et al. / The American Journal of Surgery 190 (2005) 580 –582
581
Fig. 3. Identification of mass in compressed (A) and noncompressed (B) specimen radiographs. Fig. 1. TranSpec specimen radiography device.
Comments between physicians was defined as agreement among all reviewers.
Results Of the 59 specimen radiograph targets, 29 were masses, 36 contained calcifications, and 14 contained previously placed clips. Twenty-nine targets in 23 patients were either invasive or in situ cancer. Overall, 57 of 59 targets were seen on both views by 2 reviewers and 56 of 59 targets were seen on both views by the third. Overall concordance for the 2 images was 100% for all 3 reviewers and 98% among reviewers. Clips were identified in all 14 specimens and calcifications in 35 of 36 specimens (Figs. 2 through 4). In 1 patient, calcifications were not seen on either image, but these were seen in an additional, compressed specimen (considered concordant for images and reviewers). Fifteen patients had masses without associated calcifications or clips. One reviewer identified 13 of 15 masses, and the other 2 reviewers identified 1 additional mass target. In both cases, ultrasound-guided excision had been performed (Table 1).
Fig. 2. Identification of surgical clip in compressed (A) and noncompressed (B) specimen radiographs.
Nonpalpable breast cancers account for approximately 20% of all clinically diagnosed breast cancer [1]. The majority of breast cancers are detected by mammography [3]. Image-detected breast cancer requires accurate lesion identification and margin assessment. The International Breast Cancer Consensus Conference stated that radiography of all surgically excised microcalcifications and masses should be performed. The need for tissue compression and its effects on specimen identification, size, and margin status are not well understood. Our study showed that tissue compression before specimen radiography is not routinely necessary for lesion identification. On independent review of compressed and noncompressed specimens, there was 100% image concordance and 98% concordance among reviewers. Specimen radiographs of sonographically localized masses are more difficult to interpret. For these patients, comparison with the preoperative mammogram was helpful. Clingan et al [4] showed that tissue compression produced a significant change in specimen dimensions. They postulated that this change could alter the pathologic margin and alter treatment decisions. Their survey revealed that 71% of hospitals reported using firm but varying degrees of compression. Graham et al [2] reported a decrease in specimen volume and height associated with five factors: patient age, breast tissue density, mammographic lesion type, spec-
Fig. 4. Identification of microcalcifications in compressed (A) and noncompressed (B) specimen radiographs.
582
J.E. Méndez et al. / The American Journal of Surgery 190 (2005) 580 –582
Table 1 Lesion identification on the noncompressed view Rater
Surgeon A Surgeon B Breast imaging radiologist
Masses (n ⫽ 29)
Calcifications (n ⫽ 36)
Clips (n ⫽ 14)
Seen
Not seen
Seen
Not seen
Seen
Not seen
28 27 28
1* 2* 1*
35 35 35
1† 1† 1†
14 14 14
0 0 0
* Ultrasound-guided performed for mass not seen well on preoperative mammography. † Lesion found on additional specimen but not seen on the compressed or noncompressed views.
imen size, and use of compression during specimen radiography. The only variable that independently contributed to flattening was the use of compression during specimen radiography. Compression increased the degree of flattening from 41% to 54%. They concluded that the pancake phenomenon has important implications for the accuracy of margin analysis. Specimen radiography remains the standard of care. Despite increased operative time and some flattening of specimens, specimen radiography decreases the frequency of missed lesions and provides confirmation of the suspicious target to the surgeon. Tissue compression before specimen radiography is not routinely necessary for target lesion identification. Failure to find the lesion on noncompressed views should be followed by tissue compression, repeat specimen radiography, and operative search for the lesion within the breast. Further
studies will be necessary to determine whether this approach results in fewer positive margins.
References [1] Singletary SE. Surgical margins in patients with early stage breast cancer treated with breast conservation therapy. Am J Surg 2002;184: 383–93. [2] Graham R, Homer MJ, Katz J, et al. The pancake phenomenon contributes to the inaccuracy of margin assessment in patients with breast cancer. Am J Surg 2002;184:89 –93. [3] International Breast Conference Consensus Conference. Image detected breast cancer: state of the art diagnosis and treatment. J Am Coll Surg 2001;193:297–303. [4] Clingan R, Griffin M, Phillips J, et al. Potential margin distortion in breast tissue by specimen mammography. Arch Surg 2003;138: 1371– 4.
Presentation
Tissue compression is not necessary for needle-localized lesion identification Jane E. Méndez, M.D.a, Deborah ter Meulen, M.D.b, James Padussis, B.A.c, Deborah Richard-Kowalski, R.T.(R)(M)b, Manuel Raagas, M.D.b, Sidney Pollack, M.D.b, Daniel O’Connor, M.D.b, Michael D. Stone, M.D.a,* a
Department of Surgery, Surgical Oncology, Boston Medical Center, Boston University School of Medicine, 88 E. Newton St., Boston, MA 02118, USA b Department of Radiology, Boston Medical Center, Boston University School of Medicine, Boston, MA 02118, USA c Medical Student, Boston University School of Medicine, Boston, MA 02118, USA Manuscript received June 14, 2005; revised manuscript June 16, 2005 Presented at the Sixth Annual Meeting of the American Society of Breast Surgeons, Los Angeles, California, March 16 –20, 2005
Abstract Background: Tissue compression to enhance lesion visibility on radiography of needle-localized breast biopsy specimens is widely used. We hypothesized that compression is not necessary for detection of lesions on specimen radiography. Methods: Forty-nine consecutive patients underwent needle-localization biopsies of 59 mammographic targets. All specimens were radiographed without and with compression. The films were later independently reviewed and compared with preoperative mammograms by 2 surgeons and a breast-imaging radiologist. The primary end point was identification of mammographic targets in noncompressed specimen radiographs. Results: Twenty-nine targets were masses, 36 contained calcifications, and 14 contained previously placed clips. All mammographically localized lesions were identified on noncompressed views. Overall concordance for the 2 images was 100% for all 3 reviewers and 98% among reviewers. Conclusions: Tissue compression before specimen radiography is not routinely necessary for target lesion identification. © 2005 Excerpta Medica Inc. All rights reserved. Keywords: Needle-localized breast lesion; Nonpalpable breast cancer; Specimen radiography; Tissue compression
Complete removal of nonpalpable breast cancers with negative margins is required before consideration of radiation therapy. Positive or close surgical margins are found in 20% to 60% of excision specimens, resulting in re-excision procedures. Tissue compression to enhance lesion visibility on specimen radiography is widely used, but it may increase the likelihood of positive margins by flattening a spherical specimen to a disk (“pancake effect”) [2]. We hypothesized that compression is not necessary for lesion detection.
* Corresponding author. Tel.: ⫹1-617-638-8654; fax: ⫹1-617-6388081. E-mail address: [email protected]
Methods Forty-nine consecutive patients underwent needle-localization biopsies of 59 mammographic targets from May to December 2004. Mammographic wire localization was performed for 57 targets, and sonographic guidance was used for 2 targets. All specimens were radiographed first without and then with compression. The TransSpec (E-Z-E-M, Long Island, New York) specimen device was used for compression (Fig. 1). Each surgeon reviewed the specimen radiographs intraoperatively to assure target retrieval. The films were later independently reviewed and compared with the preoperative mammograms by 2 surgeons and a breastimaging radiologist. Concordance for image was defined as target lesion seen or not seen by each reviewer on both compressed and noncompressed radiographs. Concordance
0002-9610/05/$ – see front matter © 2005 Excerpta Medica Inc. All rights reserved. doi:10.1016/j.amjsurg.2005.06.016
J.E. Méndez et al. / The American Journal of Surgery 190 (2005) 580 –582
581
Fig. 3. Identification of mass in compressed (A) and noncompressed (B) specimen radiographs. Fig. 1. TranSpec specimen radiography device.
Comments between physicians was defined as agreement among all reviewers.
Results Of the 59 specimen radiograph targets, 29 were masses, 36 contained calcifications, and 14 contained previously placed clips. Twenty-nine targets in 23 patients were either invasive or in situ cancer. Overall, 57 of 59 targets were seen on both views by 2 reviewers and 56 of 59 targets were seen on both views by the third. Overall concordance for the 2 images was 100% for all 3 reviewers and 98% among reviewers. Clips were identified in all 14 specimens and calcifications in 35 of 36 specimens (Figs. 2 through 4). In 1 patient, calcifications were not seen on either image, but these were seen in an additional, compressed specimen (considered concordant for images and reviewers). Fifteen patients had masses without associated calcifications or clips. One reviewer identified 13 of 15 masses, and the other 2 reviewers identified 1 additional mass target. In both cases, ultrasound-guided excision had been performed (Table 1).
Fig. 2. Identification of surgical clip in compressed (A) and noncompressed (B) specimen radiographs.
Nonpalpable breast cancers account for approximately 20% of all clinically diagnosed breast cancer [1]. The majority of breast cancers are detected by mammography [3]. Image-detected breast cancer requires accurate lesion identification and margin assessment. The International Breast Cancer Consensus Conference stated that radiography of all surgically excised microcalcifications and masses should be performed. The need for tissue compression and its effects on specimen identification, size, and margin status are not well understood. Our study showed that tissue compression before specimen radiography is not routinely necessary for lesion identification. On independent review of compressed and noncompressed specimens, there was 100% image concordance and 98% concordance among reviewers. Specimen radiographs of sonographically localized masses are more difficult to interpret. For these patients, comparison with the preoperative mammogram was helpful. Clingan et al [4] showed that tissue compression produced a significant change in specimen dimensions. They postulated that this change could alter the pathologic margin and alter treatment decisions. Their survey revealed that 71% of hospitals reported using firm but varying degrees of compression. Graham et al [2] reported a decrease in specimen volume and height associated with five factors: patient age, breast tissue density, mammographic lesion type, spec-
Fig. 4. Identification of microcalcifications in compressed (A) and noncompressed (B) specimen radiographs.
582
J.E. Méndez et al. / The American Journal of Surgery 190 (2005) 580 –582
Table 1 Lesion identification on the noncompressed view Rater
Surgeon A Surgeon B Breast imaging radiologist
Masses (n ⫽ 29)
Calcifications (n ⫽ 36)
Clips (n ⫽ 14)
Seen
Not seen
Seen
Not seen
Seen
Not seen
28 27 28
1* 2* 1*
35 35 35
1† 1† 1†
14 14 14
0 0 0
* Ultrasound-guided performed for mass not seen well on preoperative mammography. † Lesion found on additional specimen but not seen on the compressed or noncompressed views.
imen size, and use of compression during specimen radiography. The only variable that independently contributed to flattening was the use of compression during specimen radiography. Compression increased the degree of flattening from 41% to 54%. They concluded that the pancake phenomenon has important implications for the accuracy of margin analysis. Specimen radiography remains the standard of care. Despite increased operative time and some flattening of specimens, specimen radiography decreases the frequency of missed lesions and provides confirmation of the suspicious target to the surgeon. Tissue compression before specimen radiography is not routinely necessary for target lesion identification. Failure to find the lesion on noncompressed views should be followed by tissue compression, repeat specimen radiography, and operative search for the lesion within the breast. Further
studies will be necessary to determine whether this approach results in fewer positive margins.
References [1] Singletary SE. Surgical margins in patients with early stage breast cancer treated with breast conservation therapy. Am J Surg 2002;184: 383–93. [2] Graham R, Homer MJ, Katz J, et al. The pancake phenomenon contributes to the inaccuracy of margin assessment in patients with breast cancer. Am J Surg 2002;184:89 –93. [3] International Breast Conference Consensus Conference. Image detected breast cancer: state of the art diagnosis and treatment. J Am Coll Surg 2001;193:297–303. [4] Clingan R, Griffin M, Phillips J, et al. Potential margin distortion in breast tissue by specimen mammography. Arch Surg 2003;138: 1371– 4.