Japanese Journal of Clinical Oncology Advance Access published October 15, 2014 Japanese Journal of Clinical Oncology, 2014, 1–7 doi: 10.1093/jjco/hyu158 Original Article
Original Article
Utility of intraoperative frozen section examinations of surgical margins: implication of margin-exposed tumor component features on further surgical treatment† Downloaded from http://jjco.oxfordjournals.org/ at Kainan University on April 21, 2015
Mizuho Kikuyama1,2,3, Sadako Akashi-Tanaka2,4, Takashi Hojo2, Takayuki Kinoshita2, Toshihisa Ogawa1,5, Yasuyuki Seto1, and Hitoshi Tsuda6,7,* 1
Department of Metabolic Care and Endocrine Surgery, Graduate School of Medicine, University of Tokyo, Tokyo, Department of Breast Surgery, National Cancer Center Hospital, Tokyo, 3Department of Breast Surgery, Kanto Central Hospital of the Mutual Aid Association of Public School Teachers, Tokyo, 4Breast Surgical Oncology Department, Showa University, School of Medicine, Tokyo, 5Department of Breast Surgery, Dokkyo Medical University, Saitama, 6 Department of Pathology and Clinical Laboratories, National Cancer Center Hospital, Tokyo, and 7Department of Basic Pathology, National Defense Medical College, Saitama, Japan 2
*For reprints and all correspondence: Hitoshi Tsuda, 3-2 Namiki, Tokorozawa, Saitama, 359-8513, Japan. E-mail: [email protected] †
Presented at the 32nd Annual San Antonio Breast Cancer Symposium, San Antonio, December 9th–13th, 2009.
Received 21 May 2014; Accepted 19 September 2014
Abstract Objective: In patients who underwent breast-conserving surgery, we attempted to identify the histological characteristics of margin-exposed tumor components on intraoperative frozen section examinations that were predictive of residual tumor components in additionally resected specimens. Methods: Of 1835 patients who underwent breast-conserving surgery, we identified 220 patients who had positive surgical margins determined by intraoperative frozen section examinations and who had undergone immediate additional resections. Two observers (M.K., H.T.) reviewed the slides of frozen sections and confirmed the presence of tumor components. Results: In additionally resected specimens, residual tumors were detected in 115 cases (52.3%) but not in 105 cases (47.7%). The primary tumor characteristics of extensive intraductal component (+), younger age, invasive lobular carcinoma and pathological T3 classification were significantly associated with the residual tumor components. The margin-exposed tumor components of the maximum diameter, number of positive margins and histological type were correlated with the residual tumors. Multivariate analysis showed that the maximum tumor diameter was an independent risk factor for residual tumors. Conclusions: Diagnosis of positive margins by intraoperative frozen section examinations was useful for predicting residual tumors, and three histological properties of the margin-exposed tumor components were correlated with the status of residual tumor components. Although it was impossible to clearly identify the single main factor for predicting patients for whom additional resections were not necessary, it may be possible to consider stratification of additional surgical therapy according to the characteristics of margin-exposed tumor components on intraoperative frozen section examinations. Key words: breast cancer, breast-conserving surgery, residual tumor
© The Author 2014. Published by Oxford University Press. All rights reserved. For Permissions, please email: [email protected]
1
2
Intraoperative examinations of margins
Introduction
Figure 1. Schematic representation of intraoperative frozen section examination for partially resected breast specimens. (A) A case showing carcinoma in the surgical margin at the initial resection, additional resection 1, additional resection 2; in the additional resection 3 specimen, no tumor cells were seen. (B) A case showing carcinoma in a surgical margin at the initial resection and additional resection 1, but not in the additional resection 2. (C) A case that showed carcinoma in a surgical margin at the initial resection; in the additional resection 1 specimen, no tumor cells were seen.
when the margin-exposed components were atypical cells or atypical ductal hyperplasia, we excised additional sections in principle and used permanent formalin-fixed, paraffin-embedded sections to assess these margins. When we repeated additional resections 2–3 times and the margins remained positive or preservation of esthetics did not seem possible after further resection, we switched to mastectomy. In a small number of cases, one or more additional IFEs were not performed despite a positive margin after the first IFE because the margins were located at the edge of the breast.
Patients and Methods Patients and samples From the database of the Department of Pathology, National Cancer Center Hospital (NCCH), Tokyo, we identified 1835 patients, including 4 with synchronous bilateral lesions, who underwent BCS at the Department of Breast Surgery, NCCH, Tokyo between October 1999 and July 2008. In these patients, we identified 220 patients who had positive surgical margins determined by IFE and who underwent immediate additional resection.
BCS We examined mammography and ultrasound with/without computed tomography (CT) images before BCS. We performed BCS in accordance with the guideline for BCS in Japan and the following criteria (11): (i) tumor size ≤3.0 cm, (ii) the absence of extensive intraductal component, (iii) the absence of multiple lesions, (iv) possible radiotherapy and (v) the patient’s decision. A resection margin was set between 1.5 and 2.0 cm from the edge of the tumor in our hospital. Before tumor resection, we performed indigo carmine marking with a 1.5–2.0 cm margin around the area of the tumor extension, as guided by ultrasound or CT images. Additional resections were performed when the margins were deemed positive on IFE and repeated the resections until clear margins were obtained (Fig. 1). Furthermore,
IFE The surgeons used surgical silk to mark the one- to four-way margins, and the pathologists prepared frozen sections from the margined tissue of resected specimens. From 1999 to 2006, margin assessment for BCS was often performed on the nipple side only or a second IFE after additional resection was not always performed. Since the middle of 2006, IFE was performed in four directions and repeated until the margin was negative. An attending pathologist cut margined tissues parallel to four directions of the partially resected breast specimen, and frozen sections were made from the cut surfaces and stained with hematoxylin–eosin by the attending technologist. Subsequently, the attending pathologist histologically examined the sides flanking the cut surfaces and reported the results to the surgeon during surgery. The surfaces of the true surgical margins, which were usually heat degenerated, were not subjected to IFE to avoid unclear judgment (Fig. 1).
Pathological assessment The pathologists routinely evaluated the pathological factors of the resected specimens, according to the General Rules for Clinical and Pathological Recording of Breast Cancer (12). We scrutinized the slides of specimens to determine the presence of residual tumors in the additionally resected tissues. Two observers (M.K. and H.T.) reviewed the slides of frozen sections under a microscope and confirmed
Downloaded from http://jjco.oxfordjournals.org/ at Kainan University on April 21, 2015
Assessment of surgical margins is important in breast-conserving surgery (BCS) because positive margin status is a risk factor for local recurrence. Additional resections are required for complete surgical removal after diagnosis of a positive margin. Residual tumors after BCS have been found in 37.5–76% (1,2), and their risk factors have been investigated in several studies. Close or positive margins of the initial excision (3,4), extensive intraductal component (EIC) (5–8), young patient age (4,7), tumor size, multifocal and non-palpable tumors (9), the absence of estrogen receptor expression, and lymph node metastasis have been shown to be associated with residual tumors. However, these factors were assessed by permanent histopathological diagnosis. On the other hand, a discrepancy between margin status and presence or absence of residual tumors has been reported (10). In fact, we often encounter patients who underwent additional resections after positive results on intraoperative frozen section examinations (IFE), with no tumors detected in the permanent sections. In BCS, we usually evaluate the surgical margins by IFE. We consider that the aim of IFE is to perform optimal BCS by microscopic confirmation of tumor extent. As to the utility of IFE, the assessment of tumor statuses during surgery could provide a useful indication for the change of operative procedure, e.g. additional resection or mastectomy. Furthermore, when negative surgical margins are obtained by an immediate additional resection after diagnosis of positive margins on IFE, the second surgery for additional resections could be avoided. Few reports are available on IFE and risk factors for residual tumors. If the absence of residual tumor components on IFE can be predicted, it may be possible to avoid additional partial resection or mastectomy. In the present study, we assessed the utility of IFE and tried to identify histological characteristics of margin-exposed tumor components on IFE as predictive factors for residual tumor components in additionally resected specimens in patients who underwent BCS.
Jpn J Clin Oncol 2014 the original diagnosis. When the residual tumors could not be accurately evaluated from the description of the pathology report, the two observers judged the status. In addition, we retrospectively evaluated the following pathological factors of margin-exposed components: the maximum diameter, number of positive margins and histological type of the margin components. Among 287 cases with positive margins on IFE, the additionally resected specimens in 67 cases could not be evaluated because of the absence of mapping figures in the pathological reports or degenerative changes in the specimens. Finally, 220 cases were evaluable and included in the analyses.
Statistical analysis
Results
carcinomas and in 85 (48.9%) of 174 invasive ductal carcinomas (P = 0.008). The pathological T3 ( pT3) group showed a significantly higher rate of residual tumor components (8 of 9, 88.9%) than the combined pathological Tis, T1 and T2 groups (107 of 211, 50.7%) (P = 0.025). Residual tumors were present in 73 (67.0%) of the 109 cases in the EIC (+) group and 41 (37.3%) of the 110 cases in the EIC (−) group (P < 0.0001). In addition, the group of tumors with moderate-to-extensive lymphatic invasion (ly) showed residual tumors more frequently (22 of 32, 68.8%) than the group of tumors with ly (1+) or without (93 of 186, 50.0%) (P = 0.05). Neoadjuvant chemotherapy (NAC), grade, vascular invasion, mastopathy, estrogen receptor, progesterone receptor and human epidermal growth factor receptor type 2 were not significantly associated with the presence of residual tumors in the additionally resected specimens (Table 2). A multivariate logistic analysis was performed, which included the five parameters that were significant or nearly significant in univariate analyses; the results are shown in Table 3. Tumors classified as pT3 (OR = 11.04; 95% CI, 1.13–107.89; P = 0.039), invasive lobular carcinoma (OR = 5.53; 95% CI, 1.74–17.56; P = 0.0037), EIC (+) (OR = 4.02; 95% CI, 2.16–7.46; P < 0.0001), ly (2+) (OR = 2.62; 95% CI, 1.06–6.48; P = 0.037), and younger patient age (OR = 0.94; 95% CI, 0.92–0.97; P = 0.0001) were independent risk factors for residual tumors.
Incidence of residual tumors Of the additionally resected specimens, residual tumors were identified in 115 (52.3%) of 220 cases, whereas no tumors were present in 105 (47.7%) cases. Histological analysis of the residual tumors showed non-invasive carcinoma components in 83 (72.2%) of 115 cases and invasive carcinoma components in 32 (27.8%) cases (Table 1).
Accuracy of IEF sections For 220 patients, 763 surgical margins were subjected to IFE. On IFE, 305 and 458 margins were diagnosed as positive and negative for tumor cells, respectively. By permanent section diagnosis of the corresponding formalin-fixed specimens, 305 and 458 margins were diagnosed as positive and negative for tumor cells, respectively. In total, 18 margins were negative on IFE but positive after permanent section diagnosis, whereas another group of 18 margins was positive on IFE but negative after permanent section diagnosis. The sensitivity, specificity and accuracy were 94.1, 96.1 and 95.3%, respectively. The falsenegative and -positive rates were 5.9 and 3.9%, respectively.
Clinicopathological factors of primary tumors The mean patient age in the residual tumors group was 48.8 years, whereas that in the no residual tumor group was 54.0 years (P = 0.0002). Histological type was obtained from the final pathology reports of the surgically resected specimens. Although the rate of residual tumors was not significantly different between the invasive ductal/lobular carcinoma and non-invasive ductal carcinoma types, residual tumors were present in 18 (78.3%) of 23 invasive lobular
Table 1. Status of residual carcinoma in additionally resected specimens Number of cases (%) Positive for carcinoma Non-invasive carcinoma component Invasive carcinoma component Negative for carcinoma
115 (52.3) 83 (72.2) 32 (27.8) 105 (47.7)
Characteristics of margin-exposed tumor components As the number of directions assessed on IFE changed in the middle of 2006 in principle, we divided the present cohort into the two groups: early period (before May 2006) and late period (after June 2006). In the group of the early period, only 1 and 2 directions were assessed in 7 (5.1%) and 14 (10.3%) of the 136 cases, respectively. On the other hand, almost all cases received the IFE for 3 or more directions (92.4%). There was a significant difference in the number of directions of IFE between these two periods (P = 0.0057), whereas a majority of cases in both periods had >3 assessed directions (Table 4). We investigated whether or not the number of assessed directions could reduce the residual tumors. The results showed that there is no significant difference between the number of the assessed directions and the presence of residual tumor (Table 5). Because the difference in periods or the number of assessed directions had no significant effect on the residual tumors, we analyzed the margin statuses without separating the study period. The number of positive margins, maximum diameter and histological type of the margin-exposed tumor components were evaluated (Table 6). We set a cutoff value of 6.0 mm, which was the median value, for the maximum diameter of the margin-exposed tumor components. Residual tumors were detected in 38 (36.5%) of 104, 63 (61.8%) of 102 and 14 (100%) of 14 of the tumor groups with
Original Article
Utility of intraoperative frozen section examinations of surgical margins: implication of margin-exposed tumor component features on further surgical treatment† Downloaded from http://jjco.oxfordjournals.org/ at Kainan University on April 21, 2015
Mizuho Kikuyama1,2,3, Sadako Akashi-Tanaka2,4, Takashi Hojo2, Takayuki Kinoshita2, Toshihisa Ogawa1,5, Yasuyuki Seto1, and Hitoshi Tsuda6,7,* 1
Department of Metabolic Care and Endocrine Surgery, Graduate School of Medicine, University of Tokyo, Tokyo, Department of Breast Surgery, National Cancer Center Hospital, Tokyo, 3Department of Breast Surgery, Kanto Central Hospital of the Mutual Aid Association of Public School Teachers, Tokyo, 4Breast Surgical Oncology Department, Showa University, School of Medicine, Tokyo, 5Department of Breast Surgery, Dokkyo Medical University, Saitama, 6 Department of Pathology and Clinical Laboratories, National Cancer Center Hospital, Tokyo, and 7Department of Basic Pathology, National Defense Medical College, Saitama, Japan 2
*For reprints and all correspondence: Hitoshi Tsuda, 3-2 Namiki, Tokorozawa, Saitama, 359-8513, Japan. E-mail: [email protected] †
Presented at the 32nd Annual San Antonio Breast Cancer Symposium, San Antonio, December 9th–13th, 2009.
Received 21 May 2014; Accepted 19 September 2014
Abstract Objective: In patients who underwent breast-conserving surgery, we attempted to identify the histological characteristics of margin-exposed tumor components on intraoperative frozen section examinations that were predictive of residual tumor components in additionally resected specimens. Methods: Of 1835 patients who underwent breast-conserving surgery, we identified 220 patients who had positive surgical margins determined by intraoperative frozen section examinations and who had undergone immediate additional resections. Two observers (M.K., H.T.) reviewed the slides of frozen sections and confirmed the presence of tumor components. Results: In additionally resected specimens, residual tumors were detected in 115 cases (52.3%) but not in 105 cases (47.7%). The primary tumor characteristics of extensive intraductal component (+), younger age, invasive lobular carcinoma and pathological T3 classification were significantly associated with the residual tumor components. The margin-exposed tumor components of the maximum diameter, number of positive margins and histological type were correlated with the residual tumors. Multivariate analysis showed that the maximum tumor diameter was an independent risk factor for residual tumors. Conclusions: Diagnosis of positive margins by intraoperative frozen section examinations was useful for predicting residual tumors, and three histological properties of the margin-exposed tumor components were correlated with the status of residual tumor components. Although it was impossible to clearly identify the single main factor for predicting patients for whom additional resections were not necessary, it may be possible to consider stratification of additional surgical therapy according to the characteristics of margin-exposed tumor components on intraoperative frozen section examinations. Key words: breast cancer, breast-conserving surgery, residual tumor
© The Author 2014. Published by Oxford University Press. All rights reserved. For Permissions, please email: [email protected]
1
2
Intraoperative examinations of margins
Introduction
Figure 1. Schematic representation of intraoperative frozen section examination for partially resected breast specimens. (A) A case showing carcinoma in the surgical margin at the initial resection, additional resection 1, additional resection 2; in the additional resection 3 specimen, no tumor cells were seen. (B) A case showing carcinoma in a surgical margin at the initial resection and additional resection 1, but not in the additional resection 2. (C) A case that showed carcinoma in a surgical margin at the initial resection; in the additional resection 1 specimen, no tumor cells were seen.
when the margin-exposed components were atypical cells or atypical ductal hyperplasia, we excised additional sections in principle and used permanent formalin-fixed, paraffin-embedded sections to assess these margins. When we repeated additional resections 2–3 times and the margins remained positive or preservation of esthetics did not seem possible after further resection, we switched to mastectomy. In a small number of cases, one or more additional IFEs were not performed despite a positive margin after the first IFE because the margins were located at the edge of the breast.
Patients and Methods Patients and samples From the database of the Department of Pathology, National Cancer Center Hospital (NCCH), Tokyo, we identified 1835 patients, including 4 with synchronous bilateral lesions, who underwent BCS at the Department of Breast Surgery, NCCH, Tokyo between October 1999 and July 2008. In these patients, we identified 220 patients who had positive surgical margins determined by IFE and who underwent immediate additional resection.
BCS We examined mammography and ultrasound with/without computed tomography (CT) images before BCS. We performed BCS in accordance with the guideline for BCS in Japan and the following criteria (11): (i) tumor size ≤3.0 cm, (ii) the absence of extensive intraductal component, (iii) the absence of multiple lesions, (iv) possible radiotherapy and (v) the patient’s decision. A resection margin was set between 1.5 and 2.0 cm from the edge of the tumor in our hospital. Before tumor resection, we performed indigo carmine marking with a 1.5–2.0 cm margin around the area of the tumor extension, as guided by ultrasound or CT images. Additional resections were performed when the margins were deemed positive on IFE and repeated the resections until clear margins were obtained (Fig. 1). Furthermore,
IFE The surgeons used surgical silk to mark the one- to four-way margins, and the pathologists prepared frozen sections from the margined tissue of resected specimens. From 1999 to 2006, margin assessment for BCS was often performed on the nipple side only or a second IFE after additional resection was not always performed. Since the middle of 2006, IFE was performed in four directions and repeated until the margin was negative. An attending pathologist cut margined tissues parallel to four directions of the partially resected breast specimen, and frozen sections were made from the cut surfaces and stained with hematoxylin–eosin by the attending technologist. Subsequently, the attending pathologist histologically examined the sides flanking the cut surfaces and reported the results to the surgeon during surgery. The surfaces of the true surgical margins, which were usually heat degenerated, were not subjected to IFE to avoid unclear judgment (Fig. 1).
Pathological assessment The pathologists routinely evaluated the pathological factors of the resected specimens, according to the General Rules for Clinical and Pathological Recording of Breast Cancer (12). We scrutinized the slides of specimens to determine the presence of residual tumors in the additionally resected tissues. Two observers (M.K. and H.T.) reviewed the slides of frozen sections under a microscope and confirmed
Downloaded from http://jjco.oxfordjournals.org/ at Kainan University on April 21, 2015
Assessment of surgical margins is important in breast-conserving surgery (BCS) because positive margin status is a risk factor for local recurrence. Additional resections are required for complete surgical removal after diagnosis of a positive margin. Residual tumors after BCS have been found in 37.5–76% (1,2), and their risk factors have been investigated in several studies. Close or positive margins of the initial excision (3,4), extensive intraductal component (EIC) (5–8), young patient age (4,7), tumor size, multifocal and non-palpable tumors (9), the absence of estrogen receptor expression, and lymph node metastasis have been shown to be associated with residual tumors. However, these factors were assessed by permanent histopathological diagnosis. On the other hand, a discrepancy between margin status and presence or absence of residual tumors has been reported (10). In fact, we often encounter patients who underwent additional resections after positive results on intraoperative frozen section examinations (IFE), with no tumors detected in the permanent sections. In BCS, we usually evaluate the surgical margins by IFE. We consider that the aim of IFE is to perform optimal BCS by microscopic confirmation of tumor extent. As to the utility of IFE, the assessment of tumor statuses during surgery could provide a useful indication for the change of operative procedure, e.g. additional resection or mastectomy. Furthermore, when negative surgical margins are obtained by an immediate additional resection after diagnosis of positive margins on IFE, the second surgery for additional resections could be avoided. Few reports are available on IFE and risk factors for residual tumors. If the absence of residual tumor components on IFE can be predicted, it may be possible to avoid additional partial resection or mastectomy. In the present study, we assessed the utility of IFE and tried to identify histological characteristics of margin-exposed tumor components on IFE as predictive factors for residual tumor components in additionally resected specimens in patients who underwent BCS.
Jpn J Clin Oncol 2014 the original diagnosis. When the residual tumors could not be accurately evaluated from the description of the pathology report, the two observers judged the status. In addition, we retrospectively evaluated the following pathological factors of margin-exposed components: the maximum diameter, number of positive margins and histological type of the margin components. Among 287 cases with positive margins on IFE, the additionally resected specimens in 67 cases could not be evaluated because of the absence of mapping figures in the pathological reports or degenerative changes in the specimens. Finally, 220 cases were evaluable and included in the analyses.
Statistical analysis
Results
carcinomas and in 85 (48.9%) of 174 invasive ductal carcinomas (P = 0.008). The pathological T3 ( pT3) group showed a significantly higher rate of residual tumor components (8 of 9, 88.9%) than the combined pathological Tis, T1 and T2 groups (107 of 211, 50.7%) (P = 0.025). Residual tumors were present in 73 (67.0%) of the 109 cases in the EIC (+) group and 41 (37.3%) of the 110 cases in the EIC (−) group (P < 0.0001). In addition, the group of tumors with moderate-to-extensive lymphatic invasion (ly) showed residual tumors more frequently (22 of 32, 68.8%) than the group of tumors with ly (1+) or without (93 of 186, 50.0%) (P = 0.05). Neoadjuvant chemotherapy (NAC), grade, vascular invasion, mastopathy, estrogen receptor, progesterone receptor and human epidermal growth factor receptor type 2 were not significantly associated with the presence of residual tumors in the additionally resected specimens (Table 2). A multivariate logistic analysis was performed, which included the five parameters that were significant or nearly significant in univariate analyses; the results are shown in Table 3. Tumors classified as pT3 (OR = 11.04; 95% CI, 1.13–107.89; P = 0.039), invasive lobular carcinoma (OR = 5.53; 95% CI, 1.74–17.56; P = 0.0037), EIC (+) (OR = 4.02; 95% CI, 2.16–7.46; P < 0.0001), ly (2+) (OR = 2.62; 95% CI, 1.06–6.48; P = 0.037), and younger patient age (OR = 0.94; 95% CI, 0.92–0.97; P = 0.0001) were independent risk factors for residual tumors.
Incidence of residual tumors Of the additionally resected specimens, residual tumors were identified in 115 (52.3%) of 220 cases, whereas no tumors were present in 105 (47.7%) cases. Histological analysis of the residual tumors showed non-invasive carcinoma components in 83 (72.2%) of 115 cases and invasive carcinoma components in 32 (27.8%) cases (Table 1).
Accuracy of IEF sections For 220 patients, 763 surgical margins were subjected to IFE. On IFE, 305 and 458 margins were diagnosed as positive and negative for tumor cells, respectively. By permanent section diagnosis of the corresponding formalin-fixed specimens, 305 and 458 margins were diagnosed as positive and negative for tumor cells, respectively. In total, 18 margins were negative on IFE but positive after permanent section diagnosis, whereas another group of 18 margins was positive on IFE but negative after permanent section diagnosis. The sensitivity, specificity and accuracy were 94.1, 96.1 and 95.3%, respectively. The falsenegative and -positive rates were 5.9 and 3.9%, respectively.
Clinicopathological factors of primary tumors The mean patient age in the residual tumors group was 48.8 years, whereas that in the no residual tumor group was 54.0 years (P = 0.0002). Histological type was obtained from the final pathology reports of the surgically resected specimens. Although the rate of residual tumors was not significantly different between the invasive ductal/lobular carcinoma and non-invasive ductal carcinoma types, residual tumors were present in 18 (78.3%) of 23 invasive lobular
Table 1. Status of residual carcinoma in additionally resected specimens Number of cases (%) Positive for carcinoma Non-invasive carcinoma component Invasive carcinoma component Negative for carcinoma
115 (52.3) 83 (72.2) 32 (27.8) 105 (47.7)
Characteristics of margin-exposed tumor components As the number of directions assessed on IFE changed in the middle of 2006 in principle, we divided the present cohort into the two groups: early period (before May 2006) and late period (after June 2006). In the group of the early period, only 1 and 2 directions were assessed in 7 (5.1%) and 14 (10.3%) of the 136 cases, respectively. On the other hand, almost all cases received the IFE for 3 or more directions (92.4%). There was a significant difference in the number of directions of IFE between these two periods (P = 0.0057), whereas a majority of cases in both periods had >3 assessed directions (Table 4). We investigated whether or not the number of assessed directions could reduce the residual tumors. The results showed that there is no significant difference between the number of the assessed directions and the presence of residual tumor (Table 5). Because the difference in periods or the number of assessed directions had no significant effect on the residual tumors, we analyzed the margin statuses without separating the study period. The number of positive margins, maximum diameter and histological type of the margin-exposed tumor components were evaluated (Table 6). We set a cutoff value of 6.0 mm, which was the median value, for the maximum diameter of the margin-exposed tumor components. Residual tumors were detected in 38 (36.5%) of 104, 63 (61.8%) of 102 and 14 (100%) of 14 of the tumor groups with
