Accepted Manuscript Lumpectomy Specimen Margins Are Not Reliable in Predicting Residual Disease in Breast Conserving Surgery Rong Tang, MD, Suzanne B. Coopey, MD, Michelle C. Specht, MD, Lan Lei, MD, Michele A. Gadd, MD, Kevin S. Hughes, MD, Elena F. Brachtel, MD, Barbara L. Smith, MD, PhD PII:
S0002-9610(14)00649-7
DOI:
10.1016/j.amjsurg.2014.09.029
Reference:
AJS 11380
To appear in:
The American Journal of Surgery
Received Date: 28 February 2014 Revised Date:
6 September 2014
Accepted Date: 15 September 2014
Please cite this article as: Tang R, Coopey SB, Specht MC, Lei L, Gadd MA, Hughes KS, Brachtel EF, Smith BL, Lumpectomy Specimen Margins Are Not Reliable in Predicting Residual Disease in Breast Conserving Surgery, The American Journal of Surgery (2015), doi: 10.1016/j.amjsurg.2014.09.029. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Title: Lumpectomy Specimen Margins Are Not Reliable in Predicting Residual Disease in Breast Conserving Surgery
RI PT
Authors and Affiliations: *Rong Tang MD1,2 and *Suzanne B. Coopey MD1, Michelle C. Specht MD1, Lan Lei MD1, Michele A. Gadd MD1, Kevin S. Hughes MD1, Elena F. Brachtel MD3, Barbara L. Smith MD, PhD1 *Authors contributed equally to this work
M AN U
SC
1 Division of Surgical Oncology, Massachusetts General Hospital, Boston, MA 02114 2 Division of Breast Surgery, Hunan Provincial Tumor Hospital, The Affiliated Tumor Hospital of Xiangya Medical School of Central South University, China, 410013 3 Department of Pathology, Massachusetts General Hospital, Boston, MA 02114
EP
TE D
Corresponding Author Barbara L Smith MD, PhD MGH Breast Center Massachusetts General Hospital, Yawkey 9A 55 Fruit Street, Boston, MA 02114 Telephone: 617-724-1074 Fax: 617-724-1079 Email: [email protected]
AC C
Disclosures: The authors have no conflicts of interest to disclose. Summary: We compared tumor present at the lumpectomy specimen margins with tumor present in the lumpectomy cavity at the individual margin level. Lumpectomy specimen margin status predicted corresponding lumpectomy cavity status with only 50.9% sensitivity and 35% positive predictive value. Key words: breast cancer, lumpectomy margins, shaved cavity margins, residual disease
ACCEPTED MANUSCRIPT ABSTRACT: Background In breast conserving surgery, the concordance between lumpectomy margin (LM) status and the status of the corresponding lumpectomy cavity remains uncertain.
RI PT
Methods We analyzed pathology reports of lumpectomies from 2004-2006. We included those which contained both ink-directed LM and complete (≥4) separate corresponding shaved cavity margins (SCM). SCM pathology was used as a surrogate for lumpectomy cavity status, to determine the predictive value of LM for residual disease
M AN U
SC
Results Pathology from 1201 pairs of LM and SCM from 242 patients was compared. LM status predicted corresponding lumpectomy cavity status with 50.9% sensitivity, 69.5% specificity, 35% positive predictive value and 81.4% negative predictive value, giving an overall accuracy of 64.9%.
AC C
EP
TE D
Conclusions Oriented lumpectomy margins are not reliable for predicting lumpectomy cavity status, and therefore not reliable for directing re-excision. Taking complete, oriented shaved cavity margins at the time of lumpectomy may improve accuracy compared with traditional lumpectomy margin assessment.
ACCEPTED MANUSCRIPT ABSTRACT: Background In breast conserving surgery, the concordance between lumpectomy margin (LM) status and the status of the corresponding lumpectomy cavity remains uncertain.
RI PT
Methods We analyzed pathology reports of lumpectomies from 2004-2006. We included those which contained both ink-directed LM and complete (≥4) separate corresponding shaved cavity margins (SCM). SCM pathology was used as a surrogate for lumpectomy cavity status, to determine the predictive value of LM for residual disease
M AN U
SC
Results Pathology from 1201 pairs of LM and SCM from 242 patients was compared. LM status predicted corresponding lumpectomy cavity status with 50.9% sensitivity, 69.5% specificity, 35% positive predictive value and 81.4% negative predictive value, giving an overall accuracy of 64.9%.
AC C
EP
TE D
Conclusions Oriented lumpectomy margins are not reliable for predicting lumpectomy cavity status, and therefore not reliable for directing re-excision. Taking complete, oriented shaved cavity margins at the time of lumpectomy may improve accuracy compared with traditional lumpectomy margin assessment.
ACCEPTED MANUSCRIPT INTRODUCTION Breast conservation therapy (BCT) is equivalent to mastectomy in terms of overall and disease-free survival.1-3 The main disadvantage of BCT is the increased risk of local recurrence, which is 8–14% at 20 years of follow-up.2, 3
RI PT
Previous studies have reported that the strongest predictor of local recurrence is positive surgical margins.4, 5 The site of local recurrence is often close to the original tumor site with histological characteristics similar to the primary tumor, suggesting that local recurrences arise from residual tumor left at the time of lumpectomy.6-8
M AN U
SC
Although the optimum width of tumor-free lumpectomy margins has recently been revisited, 9 the ability to identify and accurately re-excise positive lumpectomy margins remains critical for safe breast conservation. Re-excision of inadequate margins achieves local control rates similar to those of initial lumpectomies with adequate margins. 10, 11 In order to ensure adequate tumor clearance, several methods have been developed for margin assessment and detection of residual disease within the lumpectomy cavity, including pathologic evaluation of lumpectomy margins, taking additional shaved cavity margins, and intraoperative ultrasound or radiographic examination of the lumpectomy specimen or cavity.12-15
TE D
Among these, spatial orientation of lumpectomy specimens with multi-color inking and ink-directed re-excision of involved lumpectomy margins (LM) is widely practiced. This technique can provide exact tumor-to-margin distance to determine if re-excision is required, and allows for targeted re-excision. Targeted re-excision versus whole cavity re-excision decreases the amount of tissue excised and improves cosmesis.16
AC C
EP
One shortcoming of multi-color inking and ink-directed re-excision is that no residual tumor is found at the time of re-excision in 35-49.5%17-20 and therefore, some patients undergo unnecessary re-excisions. In addition, negative margins do not always guarantee complete excision. The reported rate of residual carcinoma in re-excision or mastectomy specimens following initial negative LM >0.1 cm 8, 21 or >0.2 cm 19, 22 in width is as high as 43%. An alternative approach, the shaved cavity margins (SCM) technique, consists of taking thin strips of tissue from all aspects of the lumpectomy cavity at the time of lumpectomy in a tangential manner.23, 24 SCM represent the tissue just outside the standard lumpectomy perimeter—thus, it is reasonable to use SCM status as a surrogate for lumpectomy cavity status. Based on this assumption, our study evaluated the predictive value of oriented and inked LM for residual disease in the corresponding lumpectomy cavity. We wished to determine (1) how often additional tumor will be found in the lumpectomy cavity adjacent to an inked lumpectomy margin deemed negative by conventional pathology assessment, (2) how often there will be no additional tumor found in the lumpectomy cavity adjacent to an inked lumpectomy margin deemed positive by conventional pathology assessment and (3) to identify causes and possible solutions for these false negative and false positive margin assessments.
1
ACCEPTED MANUSCRIPT METHODS:
RI PT
An Institutional Review Board-approved retrospective analysis was performed to identify all consecutive patients undergoing BCT for primary invasive breast cancer or ductal carcinoma in situ (DCIS) between 1 January 2004 and 31 December 2006 at the Massachusetts General Hospital, Boston, MA. This time frame was selected because it represented a period of transition from lumpectomy alone to lumpectomy plus excision of SCM at our institution.
SC
All surgical procedures were performed by four dedicated breast surgeons (B.L.S., K.S.H., M.A.G., M.C.S.) The decision to resect one or more SCM and the thickness of the shaves were at the discretion of the surgeon. In general, most surgeons aimed for SCM thickness between 0.2 and 1.0 cm.
M AN U
This study included only patients who had: oriented lumpectomy specimens, all 6 LM (superior, inferior, medial, lateral, anterior, and posterior) inked and evaluated separately, and complete (≥4) SCM taken during the same procedure. Patients with unoriented lumpectomies, fewer than 4 SCM, or complete but not separate SCM (eg: shaves taken as hemispheres) were excluded from this study.
TE D
Lumpectomy specimens were sent to pathology with sutures that allowed for spatial orientation of the specimen (e.g. long suture for lateral, short suture for superior). Specimens were inked with multiple colors by the pathologist and sectioned. Representative tissue, including the closest margins, was submitted for histologic evaluation. Most SCM were entirely submitted for histologic evaluation. The presence and extent of tumor was recorded for each SCM.
AC C
EP
LM’s predictive value was analyzed at the individual margin level, considering each LM separately. Pathologic data of the LM and corresponding SCM were recorded in a paired pattern based on the presence or absence of tumor and tumor type. LM were considered clinically positive/involved if cancer cells were within 0.2cm of the inked margin’s surface because this generally triggers re-excision at our institution. The lumpectomy cavity was considered positive for residual disease when the SCM contained tumor, regardless of the distance of cancer cells to the SCM surface. LM-to-tumor distance, degree of margin involvement (none, one focus, several foci, broad front), and tumor histopathology of the primary tumor, including histology, size, grade, presence of extensive DCIS, lymphovascular invasion (LVI), multifocality, margin involvement, and estrogen receptor (ER) status, human epidermal growth factor receptor 2 (HER2) status, and lymph node status were determined. Pearson’s chi-square test was used to test for categorical variables (risk factors for positive lumpectomy margins). Statistical analyses were performed using STATA 12.0, and p-values ≤0.05 were considered statistically significant.
2
ACCEPTED MANUSCRIPT RESULTS:
RI PT
242 patients with 1452 LM and 1201 SCM were identified. Mean patient age was 53.3 years (range: 31-86 years). All patients were female. 123 (50.8%) had invasive ductal carcinoma (IDC) and DCIS, 66 (27.3%) had DCIS alone, 26 (10.7%) had IDC alone, 19 (7.9%) had invasive lobular cancer (ILC) alone, 7 (2.9%) had ILC and IDC with or without DCIS, and 1 (0.4%) had tubular carcinoma. The mean invasive tumor size was 1.8cm (range 0.1-10cm). The average lumpectomy volume, SCM volume and total resected volume was 54.5 (range 2.9-282.1) cm3, 20.7 (range 1-206) cm3, and 75.2 (range 5.6-391.7) cm3, respectively.
M AN U
SC
Histopathology of 1201/1452 LM that had corresponding SCM was compared with their corresponding SCM findings. There were 242 LM at each of 6 locations, the majority of superior (239, 98.8%), inferior (234, 96.7%), medial (237, 97.9%) and lateral (229, 94.6%) LM had corresponding SCM taken. However, many fewer anterior (101, 41.7%) and posterior (161, 66.5%) LM had corresponding SCM. This occurred when lumpectomies were performed either just beneath the skin, when lumpectomies reached pectoralis fascia, or when anterior margins were taken as part of other adjacent SCM specimens. This information was recorded in the operative report.
TE D
The predictive value of LM for residual cavity disease was evaluated using SCM as a surrogate for residual cavity disease. The sensitivity, specificity, positive predictive value, and negative predictive value were determined. At the individual margin level, overall 426 (35.5%) LM were involved (tumor < 2mm from ink), and 293 (24.4%) SCM contained tumor. When an individual LM was involved, the corresponding SCM was negative 65.0% of the time. When an individual LM was negative, the corresponding SCM was positive 18.6% of the time. The ability of the LM status to predict corresponding SCM status had 50.9% sensitivity, 69.5% specificity, 35% positive predictive value, and 81.4% negative predictive value, giving an overall accuracy of 64.9%. (Table 1).
AC C
EP
We found that there was a significant correlation between number of LM involved and residual disease. In patients with one LM involved, there was residual disease somewhere within the SCM tissue 40.3% of the time. In patients with 2-3 positive LM, and 4-6 positive LM, the cavity contained residual disease 59.8% of the time and 91.9 % of the time respectively (p< 0.001). Additionally, we studied the positive LM’s histological characteristics to determine if they were predictive of residual disease within the lumpectomy cavity. LM involved with only one focus of tumor were significantly less likely to have residual lumpectomy cavity involvement than LM involved with several tumor foci or with tumor over a broad front (19.2% vs. 37.8% and 39%, p
S0002-9610(14)00649-7
DOI:
10.1016/j.amjsurg.2014.09.029
Reference:
AJS 11380
To appear in:
The American Journal of Surgery
Received Date: 28 February 2014 Revised Date:
6 September 2014
Accepted Date: 15 September 2014
Please cite this article as: Tang R, Coopey SB, Specht MC, Lei L, Gadd MA, Hughes KS, Brachtel EF, Smith BL, Lumpectomy Specimen Margins Are Not Reliable in Predicting Residual Disease in Breast Conserving Surgery, The American Journal of Surgery (2015), doi: 10.1016/j.amjsurg.2014.09.029. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Title: Lumpectomy Specimen Margins Are Not Reliable in Predicting Residual Disease in Breast Conserving Surgery
RI PT
Authors and Affiliations: *Rong Tang MD1,2 and *Suzanne B. Coopey MD1, Michelle C. Specht MD1, Lan Lei MD1, Michele A. Gadd MD1, Kevin S. Hughes MD1, Elena F. Brachtel MD3, Barbara L. Smith MD, PhD1 *Authors contributed equally to this work
M AN U
SC
1 Division of Surgical Oncology, Massachusetts General Hospital, Boston, MA 02114 2 Division of Breast Surgery, Hunan Provincial Tumor Hospital, The Affiliated Tumor Hospital of Xiangya Medical School of Central South University, China, 410013 3 Department of Pathology, Massachusetts General Hospital, Boston, MA 02114
EP
TE D
Corresponding Author Barbara L Smith MD, PhD MGH Breast Center Massachusetts General Hospital, Yawkey 9A 55 Fruit Street, Boston, MA 02114 Telephone: 617-724-1074 Fax: 617-724-1079 Email: [email protected]
AC C
Disclosures: The authors have no conflicts of interest to disclose. Summary: We compared tumor present at the lumpectomy specimen margins with tumor present in the lumpectomy cavity at the individual margin level. Lumpectomy specimen margin status predicted corresponding lumpectomy cavity status with only 50.9% sensitivity and 35% positive predictive value. Key words: breast cancer, lumpectomy margins, shaved cavity margins, residual disease
ACCEPTED MANUSCRIPT ABSTRACT: Background In breast conserving surgery, the concordance between lumpectomy margin (LM) status and the status of the corresponding lumpectomy cavity remains uncertain.
RI PT
Methods We analyzed pathology reports of lumpectomies from 2004-2006. We included those which contained both ink-directed LM and complete (≥4) separate corresponding shaved cavity margins (SCM). SCM pathology was used as a surrogate for lumpectomy cavity status, to determine the predictive value of LM for residual disease
M AN U
SC
Results Pathology from 1201 pairs of LM and SCM from 242 patients was compared. LM status predicted corresponding lumpectomy cavity status with 50.9% sensitivity, 69.5% specificity, 35% positive predictive value and 81.4% negative predictive value, giving an overall accuracy of 64.9%.
AC C
EP
TE D
Conclusions Oriented lumpectomy margins are not reliable for predicting lumpectomy cavity status, and therefore not reliable for directing re-excision. Taking complete, oriented shaved cavity margins at the time of lumpectomy may improve accuracy compared with traditional lumpectomy margin assessment.
ACCEPTED MANUSCRIPT ABSTRACT: Background In breast conserving surgery, the concordance between lumpectomy margin (LM) status and the status of the corresponding lumpectomy cavity remains uncertain.
RI PT
Methods We analyzed pathology reports of lumpectomies from 2004-2006. We included those which contained both ink-directed LM and complete (≥4) separate corresponding shaved cavity margins (SCM). SCM pathology was used as a surrogate for lumpectomy cavity status, to determine the predictive value of LM for residual disease
M AN U
SC
Results Pathology from 1201 pairs of LM and SCM from 242 patients was compared. LM status predicted corresponding lumpectomy cavity status with 50.9% sensitivity, 69.5% specificity, 35% positive predictive value and 81.4% negative predictive value, giving an overall accuracy of 64.9%.
AC C
EP
TE D
Conclusions Oriented lumpectomy margins are not reliable for predicting lumpectomy cavity status, and therefore not reliable for directing re-excision. Taking complete, oriented shaved cavity margins at the time of lumpectomy may improve accuracy compared with traditional lumpectomy margin assessment.
ACCEPTED MANUSCRIPT INTRODUCTION Breast conservation therapy (BCT) is equivalent to mastectomy in terms of overall and disease-free survival.1-3 The main disadvantage of BCT is the increased risk of local recurrence, which is 8–14% at 20 years of follow-up.2, 3
RI PT
Previous studies have reported that the strongest predictor of local recurrence is positive surgical margins.4, 5 The site of local recurrence is often close to the original tumor site with histological characteristics similar to the primary tumor, suggesting that local recurrences arise from residual tumor left at the time of lumpectomy.6-8
M AN U
SC
Although the optimum width of tumor-free lumpectomy margins has recently been revisited, 9 the ability to identify and accurately re-excise positive lumpectomy margins remains critical for safe breast conservation. Re-excision of inadequate margins achieves local control rates similar to those of initial lumpectomies with adequate margins. 10, 11 In order to ensure adequate tumor clearance, several methods have been developed for margin assessment and detection of residual disease within the lumpectomy cavity, including pathologic evaluation of lumpectomy margins, taking additional shaved cavity margins, and intraoperative ultrasound or radiographic examination of the lumpectomy specimen or cavity.12-15
TE D
Among these, spatial orientation of lumpectomy specimens with multi-color inking and ink-directed re-excision of involved lumpectomy margins (LM) is widely practiced. This technique can provide exact tumor-to-margin distance to determine if re-excision is required, and allows for targeted re-excision. Targeted re-excision versus whole cavity re-excision decreases the amount of tissue excised and improves cosmesis.16
AC C
EP
One shortcoming of multi-color inking and ink-directed re-excision is that no residual tumor is found at the time of re-excision in 35-49.5%17-20 and therefore, some patients undergo unnecessary re-excisions. In addition, negative margins do not always guarantee complete excision. The reported rate of residual carcinoma in re-excision or mastectomy specimens following initial negative LM >0.1 cm 8, 21 or >0.2 cm 19, 22 in width is as high as 43%. An alternative approach, the shaved cavity margins (SCM) technique, consists of taking thin strips of tissue from all aspects of the lumpectomy cavity at the time of lumpectomy in a tangential manner.23, 24 SCM represent the tissue just outside the standard lumpectomy perimeter—thus, it is reasonable to use SCM status as a surrogate for lumpectomy cavity status. Based on this assumption, our study evaluated the predictive value of oriented and inked LM for residual disease in the corresponding lumpectomy cavity. We wished to determine (1) how often additional tumor will be found in the lumpectomy cavity adjacent to an inked lumpectomy margin deemed negative by conventional pathology assessment, (2) how often there will be no additional tumor found in the lumpectomy cavity adjacent to an inked lumpectomy margin deemed positive by conventional pathology assessment and (3) to identify causes and possible solutions for these false negative and false positive margin assessments.
1
ACCEPTED MANUSCRIPT METHODS:
RI PT
An Institutional Review Board-approved retrospective analysis was performed to identify all consecutive patients undergoing BCT for primary invasive breast cancer or ductal carcinoma in situ (DCIS) between 1 January 2004 and 31 December 2006 at the Massachusetts General Hospital, Boston, MA. This time frame was selected because it represented a period of transition from lumpectomy alone to lumpectomy plus excision of SCM at our institution.
SC
All surgical procedures were performed by four dedicated breast surgeons (B.L.S., K.S.H., M.A.G., M.C.S.) The decision to resect one or more SCM and the thickness of the shaves were at the discretion of the surgeon. In general, most surgeons aimed for SCM thickness between 0.2 and 1.0 cm.
M AN U
This study included only patients who had: oriented lumpectomy specimens, all 6 LM (superior, inferior, medial, lateral, anterior, and posterior) inked and evaluated separately, and complete (≥4) SCM taken during the same procedure. Patients with unoriented lumpectomies, fewer than 4 SCM, or complete but not separate SCM (eg: shaves taken as hemispheres) were excluded from this study.
TE D
Lumpectomy specimens were sent to pathology with sutures that allowed for spatial orientation of the specimen (e.g. long suture for lateral, short suture for superior). Specimens were inked with multiple colors by the pathologist and sectioned. Representative tissue, including the closest margins, was submitted for histologic evaluation. Most SCM were entirely submitted for histologic evaluation. The presence and extent of tumor was recorded for each SCM.
AC C
EP
LM’s predictive value was analyzed at the individual margin level, considering each LM separately. Pathologic data of the LM and corresponding SCM were recorded in a paired pattern based on the presence or absence of tumor and tumor type. LM were considered clinically positive/involved if cancer cells were within 0.2cm of the inked margin’s surface because this generally triggers re-excision at our institution. The lumpectomy cavity was considered positive for residual disease when the SCM contained tumor, regardless of the distance of cancer cells to the SCM surface. LM-to-tumor distance, degree of margin involvement (none, one focus, several foci, broad front), and tumor histopathology of the primary tumor, including histology, size, grade, presence of extensive DCIS, lymphovascular invasion (LVI), multifocality, margin involvement, and estrogen receptor (ER) status, human epidermal growth factor receptor 2 (HER2) status, and lymph node status were determined. Pearson’s chi-square test was used to test for categorical variables (risk factors for positive lumpectomy margins). Statistical analyses were performed using STATA 12.0, and p-values ≤0.05 were considered statistically significant.
2
ACCEPTED MANUSCRIPT RESULTS:
RI PT
242 patients with 1452 LM and 1201 SCM were identified. Mean patient age was 53.3 years (range: 31-86 years). All patients were female. 123 (50.8%) had invasive ductal carcinoma (IDC) and DCIS, 66 (27.3%) had DCIS alone, 26 (10.7%) had IDC alone, 19 (7.9%) had invasive lobular cancer (ILC) alone, 7 (2.9%) had ILC and IDC with or without DCIS, and 1 (0.4%) had tubular carcinoma. The mean invasive tumor size was 1.8cm (range 0.1-10cm). The average lumpectomy volume, SCM volume and total resected volume was 54.5 (range 2.9-282.1) cm3, 20.7 (range 1-206) cm3, and 75.2 (range 5.6-391.7) cm3, respectively.
M AN U
SC
Histopathology of 1201/1452 LM that had corresponding SCM was compared with their corresponding SCM findings. There were 242 LM at each of 6 locations, the majority of superior (239, 98.8%), inferior (234, 96.7%), medial (237, 97.9%) and lateral (229, 94.6%) LM had corresponding SCM taken. However, many fewer anterior (101, 41.7%) and posterior (161, 66.5%) LM had corresponding SCM. This occurred when lumpectomies were performed either just beneath the skin, when lumpectomies reached pectoralis fascia, or when anterior margins were taken as part of other adjacent SCM specimens. This information was recorded in the operative report.
TE D
The predictive value of LM for residual cavity disease was evaluated using SCM as a surrogate for residual cavity disease. The sensitivity, specificity, positive predictive value, and negative predictive value were determined. At the individual margin level, overall 426 (35.5%) LM were involved (tumor < 2mm from ink), and 293 (24.4%) SCM contained tumor. When an individual LM was involved, the corresponding SCM was negative 65.0% of the time. When an individual LM was negative, the corresponding SCM was positive 18.6% of the time. The ability of the LM status to predict corresponding SCM status had 50.9% sensitivity, 69.5% specificity, 35% positive predictive value, and 81.4% negative predictive value, giving an overall accuracy of 64.9%. (Table 1).
AC C
EP
We found that there was a significant correlation between number of LM involved and residual disease. In patients with one LM involved, there was residual disease somewhere within the SCM tissue 40.3% of the time. In patients with 2-3 positive LM, and 4-6 positive LM, the cavity contained residual disease 59.8% of the time and 91.9 % of the time respectively (p< 0.001). Additionally, we studied the positive LM’s histological characteristics to determine if they were predictive of residual disease within the lumpectomy cavity. LM involved with only one focus of tumor were significantly less likely to have residual lumpectomy cavity involvement than LM involved with several tumor foci or with tumor over a broad front (19.2% vs. 37.8% and 39%, p
