Clinical efficacy of intraoperative ultrasound for margin assessment in breast conserving surgery

Background: Surgical margin negativity is highly related to local recurrence of breast cancer. The authors performed this study to evaluate if specimen mammography or ultrasonography can replace the frozen section procedure for surgical margins. Methods: One-hundred fifty five patients with breast cancer were included in this study. After the surgery, the frozen biopsies were assessed in more than three different directions, and all specimens were analyzed with mammography and ultrasonography. The clinicopathologic characteristics of the patients were assessed, and closest tumor margin–resection margin distance (TM–RM distance) to the tumor was compared among specimen mammography, ultrasonography, and pathology. Results: On comparing initial cases of positive and negative margins, the mean closest TM–RM distance in specimen ultrasonography and final pathologic reports was statistically different between both groups (DCIS: p < 0.001, p = 0.006; IDC: p = 0.042, p = 0.022). Conclusion: When the closest TM–RM distance is less than 1.8 mm in specimen ultrasonography, the frozen section cannot be waived because of high risk of margin positivity. However, if the closest TM– RM distance is >4 mm in specimen ultrasonography, the frozen section can be omitted carefully because of the very low risk of margin positivity.

in the surgery room [6,[10][11][12]. Using IOUS, the surgeon can assess whether the breast cancer is well removed with sufficient surgical margins, whether the daughter nodule is removed together, and what is the closest tumor margin-resection margin (TM-RM) distance. Specimen mammography is also a useful diagnostic tool to detect non-palpable breast mass with microcalcification or complete removal of localized wire after resection [13,14].
Although there are many advantages of IOUS and specimen mammography in breast cancer surgery, it is still difficult to completely replace frozen sections in margin assessment because the ultrasound generally shows subjective results depending on the operator [15][16][17]. Therefore, the authors evaluated the incidence of negative or positive surgical margin with a closest TM-RM distance in specimen mammography, specimen ultrasonography and final pathologic reports.

Methods
From 2008 to 2014, 1040 patients with operable breast cancers underwent surgical management at the Kyungpook National University Hospital. Among them, 557 (53.56%) patients who underwent breast conserving surgery with intraoperative margin evaluation were reviewed for this study, and single mass-forming ductal carcinoma in situ (DCIS) and invasive ductal carcinoma (IDC) cases were included in this study. In addition, nine patients who had mixed type of invasive cancer were excluded. Finally, 155 eligible patients with breast cancer who underwent breast conserving surgery and intra-operative margin evaluation with frozen sections as well as specimen mammogram and ultrasound were included for this study.
All breast cancers were diagnosed by needle biopsy, and the size, location, number of tumors were evaluated with mammography, ultrasonography, and breast magnetic resonance imaging (MRI) before surgery. The treatment strategy of each patient was determined by a multidisciplinary team discussion which comprised breast and plastic surgeons, oncologists, radiologists, pathologists, and radiation oncologists. Based on the tumor stage and characteristics, adjuvant chemotherapy, radiotherapy, or endocrine treatment therapy was additionally applied. The clinicopathologic characteristics were assessed, including patient's age, history of contralateral breast cancer, clinical and pathologic tumor size, nodal status, stage and tumor characteristic including estrogen receptor (ER), progesterone receptor (PR), HER2 gene status.
All procedures performed in studies involving human participants were in accordance with the ethical standards of the Institutional Review Board of the Kyungpook National University Chilgok Hospital (KNUCH 2015-05-205).

Surgical management and margin evaluation with pathologic finding
The breast cancers were identified with preoperative ultrasonography, and the wire localization with skin marking was performed when the tumor was non-palpable or a daughter nodule existed. The breast cancer was removed with at least a 2-mm safety margin by palpation or blue-dye stained gel marking technique. Either sentinel lymph node biopsy or axillary lymph node dissection was performed according to the axillary lymph node status. For margin assessment, the parenchymal tissues were obtained from more than three different directions of the surgical cavity. The positive result of surgical margin was defined as atypical ductal hyperplasia, ductal carcinoma in situ, or invasive ductal carcinoma. Additional excision and secondary margin evaluation were performed when margin positivity was diagnosed in the intraoperative frozen section until negative surgical margins were secured.
The pathologists assessed the diameters between tumor edge and superior, inferior, lateral, and medial margins and described the same in their final pathologic findings. The closest TM-RM distance was also described in pathologic reports.

Margin evaluation in specimen mammography and ultrasonography
After the breast cancer was removed, the specimen was marked with a black silk tie 3 − 0 for identification of directions. A short tie with one metal clip was fixed in the superior direction and a long tie with two metal clips was fixed in the lateral direction. The specimen was then delivered to the mammography room for specimen mammography, which was performed using a dedicated mammography unit (Lorad Selenia, Hologic) with a single anteroposterior 1.8 × magnification. The specimen was compressed using a standard rigid plastic compression plate during mammography.
Then, it was delivered to the radiology unit to be subjected to ultrasonography. For the acoustic wave transmission, water was used instead of ultrasonic gel; water has similar acoustic impedance as that of the ultrasonic gel. The sample was immersed in water, and the radiologists obtained ultrasonographic images of the specimen with various directions of margin ( Fig. 1).
After mammographic and ultrasonographic images of specimen were linked to Picture Archiving and Communication System (PACS), the surgeon measured the closest TM-RM distance in the surgery room and determined where the frozen tissue should be evaluated further.

Statistical analysis
Statistical analyses were performed using SPSS ver. 25.0 (SPSS, Chicago, IL, USA). Categorical variables were analyzed using the χ 2 and Fisher's exact tests, and oncologic outcomes were assessed using Kaplan-Meier analysis to identify factors affecting locoregional recurrence or distant metastasis.
The p-value < 0.05 was regarded as statistically significant.

Results
There were 86 cases of DCIS and 69 cases of IDC. There was no significant difference in patient age, history of contralateral breast cancer, and type of breast surgery between the two groups. Axillary lymph node dissection and adjuvant chemotherapy were only performed in the IDC group.
Although most pathologic findings, including pathologic tumor size, did not show statistical difference between two groups, the HER2 gene was significantly highly expressed in the DCIS group. Positive results in the initial frozen section were identified in eight cases of DCIS and five cases of IDC (p = 0.645). There were three cases of atypical ductal hyperplasia and five cases of DCIS in the DCIS group, and five cases of DCIS in the IDC group. The incidence of locoregional recurrence and distant metastasis occurred only in the IDC group (p = 0.024, < 0.001) ( Table 1).  (Fig. 2, Table 2).   (Table 3).  [18][19][20], and the local control of breast cancer is an essential factor in oncologic outcome because it may prevent distant metastasis. Although the surgical margin can be confirmed on the final pathologic reports, a re-operation would be needed when if positive surgical margins are identified. If the surgical margin can be confirmed during surgery, only re-excision under general anesthesia is needed to secure negative results.
However, the frozen section for margin assessment requires time and effort on the part of the pathologist which prolongs the operation. Therefore, researchers have tried to predict the margin status using imaging modalities, including specimen mammography, tomosynthesis, and IOUS, to overcome these shortcomings [9, 12, 21-23]. However, most of these are 2D imaging modalities, and their results for the closest distance between tumor and surgical margin would be different compared to the true margin; this is particularly true for specimen mammography because of compression technique, but there are some differences depending on clinics. In addition, because the sonographic images cannot detect an existence of microscopic tumor, a pathologic margin assessment cannot be completely omitted. The authors assumed that a certain distance can be a standard point which has a high probability of positive or negative potential in margin assessment. Then, the number of frozen sections can be reduced, which may lead to reduced surgery duration and reduced burden to the pathologists.
In a recent study, the closest distance was shorter in specimen mammography than in specimen ultrasound and final pathologic reports in both DCIS and IDC groups. Clingan et al. reported that the TM-RM distance of specimen mammography could be distorted by compression of specimen [24].
Conversely, the closest margins of specimen ultrasound showed a similar pattern as those of the final pathologic report. In addition, although the possibility of margin positivity was higher when the closest margin was less than 1. to compensate for this disadvantage [25][26][27][28]. However, a specific device and agent are necessary to conduct such studies. Another limitation of this study is that the closest TM-RM distance of each technique could not be exactly matched in comparison, which is a common limitation in retrospective studies. More accurate data could have been produced if the tissue could be sampled or measured only in the same direction by staining or placing a clip. However, it is possible to assume that a similar position would have been measured without necessarily indicating the position.

Conclusion
For margin assessment in breast conserving surgery, it would be considerable to omit the frozen section directions wherein the closest TM-RM distance in specimen ultrasound is longer than 4 mm; however, frozen sections should be used in the directions that show the closest TM-RM distance in specimen ultrasound to be shorter than 1.8 mm.
According to our results, the burden on the operators and pathologists who perform frozen sections could be reduced in breast conserving surgery, and the IOUS could be considered a more essential diagnostic tool for breast conserving surgery.