Intraoperative imprint cytology of sentinel lymph nodes in breast cancer patients: comparation with frozen section

Background/Aim. Sentinel lymph node (SLN) biopsy has been established as the standard of care for axillary staging in patients with invasive breast carcinoma and clinically negative lymph nodes. Intraoperative assessment of sentinel lymph nodes might be done by frozen section (FS), touch imprint cytology (TIC) and one step nucleic acid amplification. The aim of this study was to review our institution's results with SLN biopsy using TIC and FS technique as intraoperative diagnostic tool for breast cancer patients. Methods. SLNs from 101 patients were examined intraop-eratively by frozen hematoxylin-eosin (H&E) stain and by touch imprint cytology. Results of TIC were compared with FS and permanent histology sections. Results. The total number of dissected SLNs was 163 with a mean of 1.6 (1–4) per patient. The permanent H&E identified 19 (19%) patients with a sentinel lymph node metastasis and 82 (81%) patients with tumor-free sentinel nodes. The sensitiv-ity/specificity rates were 73.7%/99.3%, respectively for TIC and 84.2%/100%, respectively for FS. Relevant posi-tive/negative predictive values were 93.3%/96.6%, respectively for TIC and 100%/97.9%, respectively for FS. Conclusion. Our experience with TIC and FS for the intraoperative evaluation of SLNs is similar to the findings from previously reported studies. We detected the high specificity for both methods, but TIC technique appeared to be less sensitive than FS in detecting SLN metastases in breast cancer patients. TIC could be recommended as reasonable al-ternative to frozen section due to its simplicity and low cost.

patients with tumor-free sentinel nodes. The sensitivity/specificity rates were 73.7%/99.3%, respectively for TIC and 84.2%/100%, respectively for FS. Relevant positive/negative predictive values were 93.3%/96.6%, respectively for TIC and 100%/97.9%, respectively for FS. Conclusion. Our experience with TIC and FS for the intraoperative evaluation of SLNs is similar to the findings from previously reported studies. We detected the high specificity for both methods, but TIC technique appeared to be less sensitive than FS in detecting SLN metastases in breast cancer patients. TIC could be recommended as reasonable alternative to frozen section due to its simplicity and low cost.
Key words: sentinel lymph node biopsy; breast neoplasms; intraoperative period; diagnostic techniques and procedures.

Introduction
Axillary lymph node (ALN) status is an important prognostic factor and determinant of treatment for patients with breast carcinoma. Sentinel lymph node (SLN) biopsy has been established as the standard of care in assessing the axilla in patients with invasive breast carcinoma and clinically negative lymph nodes 1 . It is a minimally invasive procedure that accurately evaluates the status of the axilla and can obtain the same prognostic information derived from axillary lymph node dissection (ALND) with significantly less morbidity 1, 2 . SLN is the first node receiving lymphatic drainage directly from the primary tumor. Thus, it is the node most likely to be the site of initial lymphatic metastasis. Currently, there is level 1 of evidence that documents that SLN biopsy is as accurate as ALND for breast cancer staging. If SLN is negative, it is predicted that the rest of the ALNs will also be negative 3 . Conversely, if SLN is positive the rest of the ALNs might also contain metastatic tumor deposits 1, 4-6 .
SLN biopsies are performed in highly equipped institutions by lympho-scintigraphy scan or blue dye mapping of SLN or combined technique 1,4 .
Intraoperative evaluation of SLNs is enabled by various techniques such as frozen section (FS), touch imprint cytology (TIC) and one step nucleic acid amplification (OSNA) 4 . FS examination is the most common method for intraoperative diagnostics of SLN, but disadvantages are loss of valuable tissue for definitive histological examination, considerable time consumption, technical difficulty in preparation of fatty tissue, specific instrumentation and costs [5][6][7] . TIC is rapid, inexpensive, easy, widely available method with maximum tissue preservation that allows clear cytological details, but it requires well educated pathologist in terms of breast cytology. Immunohistochemical (IHC) staining for cytokeratines is not routinely used in intraoperative evaluation. ICH is very accurate technique, but is a time consuming and expensive method that requires a special equipment 1, 6-9 .
All intraoperative diagnostic techniques are followed by routine examination of paraffin-embedded and hematoxylineosin (H&E) stained sections, which is a reference standard, after which the definite staging of axillary lymph nodes is performed 1,3,8,10 .
Both FS and intraoperative cytology imprints have a wide variety in sensitivity rates 4,10,11 .
The purpose of this study was to determine our institution's experience using both FS and TIC techniques for intraoperative detection of metastases in SLN biopsy and comparation with standard permanent section examination.

Methods
Prospective study was performed at the Oncology Institute of Vojvodina, Sremska Kamenica, Serbia during 2014, 2015 and 2016. Study included 101 patients with histologically confirmed breast cancer and clinically negative axillary lymph nodes treated operatively with SLN biopsy. Neoadjuvant chemotherapy was an excluding factor for the study. Detection of SLN was performed by combined method: preoperative application of 1 mL methylene blue dye and 1 mL (1mCi) of radioactive isotope (Tc99 nanocolloid).
SLN were identified successfully in all patients and were sent immediately for pathological examination. Fresh lymph nodes larger than 3 mm were bisected along long axis and each surface was touched on glass slide. The imprint samples were air-dried and fixed in 95% alcohol, than stained by May-Grünwald-Giemsa Quick-stain (Bio-Optica, Italy) and analyzed under microscope. Fresh cut lymph nodes were then frozen, cut at 5 μm, 3-5 sections per slide, stained in standard H&E stain and microscopically analyzed. Slides were analyzed by experienced pathologist and reported to surgeon intraoperatively.
The decision of performing ALND was made based on results of FS of SLN. The SLNs specimens were then placed in cassette, fixed in 10% buffered formalin for routine processing and standard pathological examination.
Breast cancer tumor staging was performed based on tumor size, status of axillary lymph nodes and metastases, determining pathological (p)TNM -tumor-nodus-metastasis status of breast cancer according to the American Joint Committee on Cancer (AJCC) 12 . Assessment of breast cancer differentiation was performed by modified Bloom-Richardson score 13 .
The results of FS and TIC were compared with definitive postoperative histopathology results of SLNs and analyzed using Statistical Package for Social Sciences (SPSS), version 18 (SPSS Inc. Chicago, USA).

Results
Breast surgery with SLN biopsy was made in 101 female patients with breast cancer and clinically negative axillary lymph nodes. Sentinel lymph node was successfully obtained in all patients (100%) and 163 lymph nodes were recieved for pathological analysis. The patients ranged in age from 29 to 82 (mean age 58.2). None of the 10 patients with non invasive extensive high-grade in situ carcinoma had SLN metastases. Invasive breast tumors were classified as pTis (n = 10; 9.9%), pT1 (n = 58; 58.4%) and pT2 (n = 33; 32.7%). Most frequent type of the tumor in analyzed group was ductal invasive carcinoma (n = 69; 68.3%). Primary tumor grading using Bloom-Richardson Grading System found that 28.7% patients had grade 1, 43.6% had grade 2 and 27.7% grade 3 of breast carcinoma. Tumor size varied from 3 mm up to 40 mm and 4 cases were multifocal carcinomas. Patients and tumor characteristics are summarized in Table 1.  Average number of lymph nodes per patient was 1.6 (1 to 4). A total number of 163 SLN was examined by TIC, FS and permanent histopathological section methods. TIC was positive for SLN metastases in 15 cases ( Figure 1) and negative for metastases in 148 cases. FS detected metastases in 16 SLN ( Figure 2) and 147 were negative.
Metastatic deposits > 2 mm were marked as micrometastases and ≤ 2 mm as macrometastases. Permanent histology sections, considered a gold standard in diagnostics of metastatic deposits, showed metastases in 19 SLN. Macrometastases were present in 17 SLN and micrometastases in 2 examined lymph nodes.

Fig. 1 -Clusters of tumor cells in touch imprint cytology smears in lymph node with breast cancer metastasis (May-Grünwald-Giemsa, ×400).
There was discordance between TIC and histopathology reports in 6 SLN. Five cases of negative TIC turned out positive for metastases in histopathology, and one of the positive TIC was found negative in histopathology. Two cases of false negative in TIC were micrometastases and other 3 were macrometastases. In total 163 SLN examined, 14 were positive for metastases on both TIC and permanent sections, and 143 were negative after analyzing with both methods.

Fig. 2 -Metastasis of breast carcinoma in sentinel lymph node on frozen section (hematoxylin-eosin, ×10).
Intraoperatively, FS technique found 16 positive cases and 147 negative cases for nodal metastases. Permanent sections and FS showed discordance in 3 false negative cases, two for micrometastases and one for macrometastases ( Table 2).
Based on examination of 163 SLN acquired from 101 patients in our study, the sensitivity for metastases detected by TIC was (14/19)

Discussion
Sentinel lymph node biopsy is a worldwide accepted concept for patients with breast carcinoma. Therefore, intraoperative detection of SLN is an imperative. However, because of the lack of equipment such as special infrastructure for preparation, storage and handling of radioactive technetium 99-label colloid or ineffective purchase of methylene blue dye, many facilities still use ALND 6 .

FS -frozen section; TIC -touch imprint cytology.
Frozen section may provide information on the size of metastasis, but it causes loss of tissue for permanent sections, it is time consuming and expensive technique requiring a cryostat as well as skilled professionals. Touch imprint cytology requires less effort, it is faster, saves tissue for permanent sections, but pathologist needs to be trained for reporting cytology samples. Intraoperative cytology provides rapid results with minimal artifacts. However, number of examined cells in cytology samples is smaller 4,11 .
Numerous studies comparing FS and TIC in intraoperative evaluation of SNLs have demonstrated significant variation in sensitivity of 44-100% for FS and 34-95% for TIC 4,10 . However, the variations of the methodology involved in the intraoperative as well as permanent section histopathologic evaluation make it very difficult to reliably compare different studies.
Tew et al. 10 reviewed 31 studies comparing TIC and FS in the literature and overall sensitivity of TIC was 63%, with a pool sensitivity of 81% for macrometastases and 22% for micrometastases. A similar meta-analysis reporting on FS examination found an overall sensitivity of 78%, with 94% for macro-and 40% for micrometastases 11 .
In comparison of TIC and FS, although there was higher sensitivity of FS, no statistically significant difference between these two methods was found in the most of the studies 7,11,[13][14][15][16] . The lower sensitivity of TIC is usually caused by inadequate sampling, and might be overcome when the number of slides during TIC is increased. This can improve sensitivity of the method without losing tissue for permanent histological examination 7,15,17 .
High specificity for both FS and intraoperative cytology approach, indicates that the false positive rates of these techniques are close to zero 10,18 . Higher false negative rates for both methods of intraoperative examination of SLNs are seen in low nuclear grade metastatic tumors and particularly lobular carcinomas, since these tumor cells are small and poorly cohesive 11 .
Our study showed 73.7% sensitivity and 99.3% specificity for TIC. The case of false positive imprint was due to misinterpretation of epithelioid histiocytes. Germinal center lymphocytes or activated endothelial cells could also rarely be mistaken for tumor cells 15 . The omission of micrometastases is the major cause of false negative intraoperative diagnoses. In accordance to literature data, our study showed two false negative imprints as the result of micrometastases. Reasons for false negative result is smaller number of examined cells comparing to FS and unrecognized individual tumors cells in well differentiated carcinomas 11,18 .
The use of intraoperative immunohistochemistry with cytokeratins could minimize the intraoperative false negative rates. Such protocols are now available for using on either FS or cytology imprints. However, turnaround time for such protocols is 16-20 minutes, and it prolongs the time of the surgery and costs of the diagnostics and thus it is not a standardized procedure 14,15,17 . Recently, intraoperative ultrarapid IHC has been investigated for its feasibility, validity, and effectiveness in comparison with FS. Ultrarapid cytokeratin IHC significantly enhanced intraoperative detection of metastasis in SLNs without increased time for assessment. This technique is currently not widely available and requires specialized expertise 7,9 . Immunohistochemistry is a standard procedure if there is a suspicious presence of metastatic cells during permanent section examination.
The clinical prognostic significance of micrometastases in SLN remains controversial, and some authors consider micrometastases to behave similarly to macrometastases 1,17,18 . Several studies have questioned the clinical and pathologic significance of finding micrometastases in SLN, particularly in intraoperative consultations 19,20 .
Currently, the standard practice has been to offer completion axillary lymph node dissection in patients who are found to have positive SLN for metastatic carcinoma either during the primary surgical procedure or in permanent histopathology report [20][21][22] . However, the recently reported results from the American College of Surgeons Clinical Oncology Group (ACOSOG) Z0011 trial found that there was no statistically significant benefit from ALND for women who had clinically negative axilla but the SLN was positive 3 . Recommendations from recent studies advise that axillary lymph node dissection can be omitted in patients with one or two positive sentinel nodes when conventionally whole-breast radiation therapy is planned 21,22 . Therefore, the role of intraoperative assessment of SLN in breast cancer seems to be in evolution.

Conclusion
Our experience with TIC and FS for the intraoperative evaluation of SLNs is similar to the findings from previously reported studies. We detected the high specificity for both methods, but TIC technique appeared to be less sensitive than FS in detecting SLN metastases. TIC results can be obtained with reasonable accuracy within a short time frame, permitting intraoperative decisions regarding management of the axilla in the breast can-cer patients. Therefore TIC could be recommended as an alternative to FS in view of its simplicity and low cost.