Abstract
Background/Aim: An accurate evaluation of resection margin (RM) is critical in breast-conserving surgery (BCS) as negative RM status is critical for successful local control. We compared gross and microscopic methods for RM evaluation and analyzed their concordances. Patients and Methods: Gross evaluation (GE), frozen section analysis (FSA), and permanent section diagnosis (PSD) were compared for specimens from 725 breast cancer patients. Results: The RM was grossly involved in 74 cases (10.2%). The sensitivity and specificity of GE were 22.9% and 96.1%, respectively. FSA revealed positive RM in 290 cases (40.0%), with high sensitivity (86.7%) and specificity (83.1%). With PSD, 240 cases (33.1%) showed RM involvement. Discordant results between gross and microscopic methods were observed in 104 cases (14.3%). Conclusion: Our observations of the low sensitivity of GE, high discordance rate between gross and microscopic methods, and high sensitivity and specificity of FSA support the necessity of intraoperative FSA for assessing RM status during BCS.
- Breast cancer
- breast-conserving surgery
- gross evaluation
- frozen section analysis
- permanent section diagnosis
Breast-conserving surgery (BCS) is an effective therapeutic procedure for breast cancer, resulting in better quality of life for younger patients compared to radical mastectomy (1-3). BCS with adjuvant radiation therapy is the preferred option for patients with stage I-II breast cancer. However, patients treated with BCS have higher rates of residual tumor on the surgical resection margin (RM) compared to those who underwent mastectomy (4, 5). Because RM involvement increases local recurrence rate by at least two folds, negative RM status is critical for successful local control (6). In this regard, an accurate evaluation of RM is important during BCS. The Society of Surgical Oncology-American Society of Radiation Oncology (SSO-ASTRO) Consensus Guideline recommends that ‘no ink on the tumor’ constitutes an adequate surgical RM for BCS (7). RM involvement has resulted in variable but often high re-excision rates ranging between 20-70%. Re-excision increases healthcare costs and results in inconvenience and delays adjuvant therapies (8-16).
The intraoperative RM assessment is a useful tool to avoid re-excision in breast cancer patients. Some techniques for assessing RM are available intraoperatively, including gross evaluation (GE) and frozen section analysis (FSA). GE includes direct inspection and palpation of the specimen immediately after surgical excision, without histological examination (17-20) while FSA involves rapid freezing of RM tissue samples, followed by hematoxylin and eosin (H&E) staining and histological examination (21-24). FSA has been reported to significantly reduce the re-excision rate (23). However, it requires highly skilled pathologists and technicians, and places additional costs. GE is a cost-effective and easy technique, but the available data regarding its adequacy are limited. To the best of our knowledge, there has been no large-scale studies investigating the concordances among the results of GE, FSA, and final RM status determined by permanent section diagnosis (PSD). In this study, we compared the outcomes of the three RM assessment methods (GE, FSA, and PSD), and analyzed their concordances in patients with breast cancer undergoing BCS.
Patients and Methods
Patient selection and data collection. This study (2019-05-051) was approved by the Institutional Review Board of Kangbuk Samsung Hospital. The electronic medical records and pathology reports were reviewed to identify breast cancer patients who underwent BCS for primary treatment between January 2015 and December 2020. We included 725 patients who were diagnosed with invasive ductal carcinoma (IDC) with or without ductal carcinoma in situ (DCIS). We excluded 67 patients whose tumors showed a complete or near-complete response to neoadjuvant chemotherapy, resulting in no grossly unidentifiable lesions. The following clinicopathological information was collected: patients’ age; tumor size (including IDC and DCIS); pathological tumor (pT) stage; histological grade; the expression status of estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor receptor 2 (HER2); lymph node metastasis; pathological nodal stage (pN); and the results of FSA and PSD. The expression status of hormone receptors was determined by immunostaining. ER or PR positivity was defined by Allred scores of 3-8 (25). Positive HER2 expression was established based on either a score of 3+ or HER2 amplification confirmed by silver in situ hybridization when the score was 2+ (26).
Intraoperative GE and FSA. All specimens were grossly and microscopically examined to assess RM status intraoperatively. The surgeon oriented the specimen using black silk sutures (Figure 1A). The superior and medial RMs were tagged by long and short sutures, respectively. When suture tagging was absent, the pathologist communicated to the surgeon to clarify the orientation. Six RMs, including the superior, inferior, medial, lateral, superficial, and deep RMs, were inked with blue, red, green, yellow, orange, and black, respectively. The specimen was sectioned at 3- or 4-mm intervals, perpendicular to the longest axis (Figure 1B). The safety distance between the tumor and each of the RMs was documented on the pathology report, and the closest RM to the tumor was indicated. Positive RM on GE (i.e., grossly involved RM) was defined as the abutment of the tumor to the RM (i.e., no distance between the tumor and RM).
Intraoperative frozen section analysis (FSA) during breast-conserving surgery for breast cancer. (A) The specimen is oriented using black silk sutures and six different color inks. The superior and medial resection margins (RMs) are tagged with long and short sutures, respectively. (B) The specimen is serially sectioned at 3- or 4-mm intervals, perpendicular to the longest axis. (C) RM tissue samples are obtained from six directions labeled with different color inks and submitted for microscopic examination. The superior, inferior, medial, lateral, superficial, and deep resection margins belong to different block numbers (A to F, respectively). The main tumor belongs to block number G. If necessary, additional tumors or lesions to be examined by FSA are designated as H, I, and so on.
Following GE, tissue samples for FSA were obtained from the RMs (Figure 1C). When the tumor was located sufficiently away (safety distance >5 mm), the RM was sampled by shaving the surface of the specimen. When the safety distance was ≤5 mm, both the tumor and RM were sampled with a perpendicular section to the surface. The FSA result was recorded according to the 2014 SSO-ASTRO Consensus Guideline (27). ‘No ink on the tumor’ was reported as negative RM regardless of the distance between the tumor and RM.
Comparisons among GE, FSA, and PSD. In addition to the original FSA slides (Figure 2A), we obtained additional PSD slides from the corresponding blocks used for FSA (Figure 2B). We determined whether the results of GE and FSA are concordant with PSD. In addition, discordance between gross and microscopic RM status was defined as the discordance between the direction of microscopically involved RM and that of the closest distance between the tumor and RM.
Representative photomicrographs of frozen and permanent sections. (A) Frozen section obtained from the RM tissue sample embedded in optimally cutting temperature compound reveals invasive carcinoma. (B) The matched permanent section is obtained from the same tissue sample that is subsequently formalin- and paraffin-embedded after frozen sectioning. It shows concordant results between frozen and permanent sections.
Statistical analysis. The chi-square test, Fisher’s exact test, and linear-by-linear association test were used to examine whether the results of the GE and FSA were significantly different from PSD and whether the discordance between gross and microscopic RM status and microscopically confirmed RM involvement were significantly associated with clinicopathological characteristics. All statistical analyses were performed using IBM SPSS Statistics for Windows, Version 23.0 (Armonk, NY, USA). Statistical significance was defined as a p-value <0.05.
Results
Table I summarizes patients’ clinicopathological characteristics. The patients’ ages ranged between 34 and 77 years (mean=55.3 years). Four hundred (5.2%) patients were older than 55 years. pT was distributed as follows: pT1 (including microinvasive carcinoma) in 76.9% (558/725); pT2 in 22.6% (164/725); and pT3 in 0.4% (3/725). The tumor size was 2.0 cm or larger in 229 (31.6%) cases. Two hundred (27.6%), 380 (52.4), and 145 (2.0) cases were graded as grade 1, grade 2, and grade 3, respectively. One hundred and twenty-nine patients (17.8%) developed lymph node metastasis. The positive rates of ER, PR, and HER2 expression were 79.7%, 68.6% and 18.6%, respectively. Triple-negative breast cancer was diagnosed in 102 patients (14.1%). DCIS was present in 520 patients (69.1%).
Clinicopathological characteristics.
The mean specimen size was 6.01 cm, and the mean closest safety distance between the tumor and the RM was 0.39 cm (Table II). Approximately 10% (74/725; 10.2%) of the cases had grossly involved RM. Two hundred and ninety (40.0%) and 240 (33.1%) cases were found to have positive RM on FSA and PSD, respectively. Despite high specificity of GE (96.1%), only 55 of the 240 cases with positive RM on PSD had grossly involved RM, indicating low GE sensitivity (22.9%; Table III). In contrast, FSA showed high sensitivity (86.7%) and specificity (83.1%; Table III).
Specimen size and the safety distance.
Concordance of gross evaluation and frozen section with permanent section.
Discordance between gross and microscopic RM status was observed in 104 cases (14.3%). No clinicopathological characteristics of breast cancer patients were significantly associated with the discordance (Table IV).
Clinicopathological significance of concordance between gross and microscopic resection margin status.
Table V summarizes the association between microscopically confirmed RM involvement and clinicopathological characteristics of breast cancer patients. Younger age (p=0.001), smaller specimen size (p=0.001), higher pT (p=0.010), and larger tumor size (p=0.004) were significantly associated with positive RM on PSD. The presence of DCIS was marginally associated with microscopically confirmed RM involvement (p=0.084).
Clinicopathological significance of resection margin involvement.
Discussion
Since the SSO-ASTRO Consensus Guideline was introduced, attention to the RM status in BCS has increased, particularly regarding the intraoperative methods for assessing RM involvement. The use of intraoperative FSA during BCS led to an extremely low re-excision rate of 2% (27). According to previous data obtained from a total of 13,081 breast cancer patients, the relative risk of recurrence for positive versus negative RM was 2.44 (27, 28). The risk of ipsilateral breast tumor recurrence in RM-positive breast cancer patients was at least two-fold that in RM-negative patients (6, 29). Despite the importance of histological examination using FSA in BCS, there have been few published studies comparing the results of GE, FSA, and PSD. In this study, GE results were in overall agreement with PSD when both methods revealed negative RM (i.e., high specificity of GE). However, when the results of GE and FSA were compared with PSD, the sensitivity of GE (22.9%) was much lower than that of FSA (86.7%).
GE is a fundamental diagnostic process providing guidance on sampling for histological examination and taking an important role in intraoperative decision-making to prevent residual disease. A previous study involving intraoperative macroscopic pathological assessment found a strong association between positive RMs and the presence of residual disease (5). In addition, GE is the most effective technique in identifying the boundaries of well-circumscribed solid masses. However, fibrous tissue extending from or surrounding the mass poses considerable challenges because it causes difficulty in distinguishing the tumor from the adjacent benign breast tissue. When a small and poorly circumscribed mass is associated with dense fibrosis and calcifications, GE alone is insufficient to differentiate the neoplastic lesion from benign tissue (29). Nunez et al. (20) documented that the 80% of accuracy of GE with an 8.6% false-negative rate seemed acceptable for routine clinical use and suggested that in cases where the grossly measured distance is 4 mm or more, re-excision can safely be avoided to optimize cosmetic outcome. However, the possibility of discrepancy in RM status between GE and histological examination cannot be disregarded. GE resulted in re-excision in 61% of the cases. Given their overall performance matrices, including 49% sensitivity, 86% specificity, and a high rate of discordance between cases with multifocal tumors, one should be cautious when applying GE alone in routine practice as a substitute for FSA. Balch et al. (17) also examined the accuracy of GE in 254 consecutive patients who underwent BCS. Re-excision for margin clearance had to be undertaken in 25% of the patients, indicating that GE did not accurately assess RM status in these patients. They concluded that other techniques for RM assessment should be considered to reduce the need for re-excision of the involved RM. Hoekstra et al. (30) observed that GE conducted by the surgeon resulted in unnecessary excision of additional normal breast tissue, even though it seemed effective in reducing re-excision rate. Nowikiewicz et al. (31) demonstrated that the number of re-operation cases was significantly reduced in the FSA group compared to the non-FSA group despite a significantly longer operative time in the former.
In this study, positive RM was significantly associated with younger age, smaller specimen size, higher pT, and larger tumor size. The higher rate of positive RM in younger patients can be attributed to the surgeons’ preference to preserve normal breast tissue as much as possible. This assumption is supported by the finding that the size of specimens with positive RM was smaller than those with negative RM. Previous studies have reported that RM involvement in BCS specimens was significantly associated with the following parameters: extensive intraductal component, multifocality, tumor size, adjacent DCIS, lobular histology, HER2 positivity, pN, caudal location, and microcalcifications on mammography (32-34). We observed a marginally significant association between the presence of DCIS and RM involvement (p=0.084). This finding was in line with a previous study by Vos et al. (35) showing that the presence of adjacent DCIS was associated with positive RM and a study by Cabioglu et al. (36) reporting that the re-operation rate in DCIS patients was twice as high (14%) as that in IDC patients (7%).
We found that 14.3% (104/725) of the cases showed discordance in RM status between gross and microscopic methods, highlighting the limited efficacy of GE as an RM assessment method for BCS. Achieving negative RM at initial operation is the primary goal for patients undergoing BCS for breast cancer. In a study on the economic implications of FSA by Boughey et al. (37), not only was the widespread use of intraoperative FSA beneficial to clinicians by providing appropriate guidance to perform surgery and to patients by minimizing the need for a re-operation, but it was also financially beneficial for both health care payers and providers.
Despite the advantage of a large sample size, this study has some limitations, including the lack of quantification of the extent of RM involvement, the lack of detailed information about the residual tumor in re-excision specimens, and the lack of long-term follow-up data on local recurrence, distant metastasis, and patients’ survival. In addition, our FSA protocol for six-directional RM evaluation by pathologists specialized in breast pathology may not be applicable at other institutions, thus our results may not be reproducible consistently.
In conclusion, our observations on the low sensitivity of GE, considerably high rate of discordance between gross and microscopic methods, and high sensitivity and specificity of FSA support the necessity of intraoperative FSA for assessing RM status during BCS. Further investigations are required to examine the overall performance of GE for RM assessment and the cost and benefit of routine FSA in other types of breast cancer cohorts. Confirmatory trials might be warranted to determine whether the FSA method we used is reproducible in other institutions.
Acknowledgements
This work was supported by the Chungnam National University Hospital Research Fund, 2021.
Footnotes
↵* These Authors contributed equally to this study.
Authors’ Contributions
All Authors made substantial contributions to the conceptualization and design of the study; the acquisition, analysis, interpretation, visualization, and validation of the data; drafting of the article; critical revision of the article for important intellectual content; and the final approval of the version to be published.
Conflicts of Interest
None of the Authors have any conflicts of interest or financial ties to declare regarding this study.
- Received July 15, 2022.
- Revision received July 28, 2022.
- Accepted July 29, 2022.
- Copyright © 2022 International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved.








