Accuracy of intraoperative imprint cytology for sentinel lymph node evaluation in the treatment of...

Accuracy of Intraoperative Imprint Cytology forSentinel Lymph Node Evaluation in the Treatmentof Breast CarcinomaA 6-Year Study

Charles Cox, M.D.1

Barbara Centeno, M.D.2

Dan Dickson, B.S.1

John Clark, B.S.1

Santo Nicosia, M.D.2

Elisabeth Dupont, M.D.1

Harvey Greenberg, M.D.1

Nicholas Stowell, B.S.1

Laura White, B.S.1

Jayesh Patel1

Ben Furman, M.D.1

Alan Cantor, Ph.D.3

Ardeshir Hakam, M.D.2

Nazeel Ahmad, M.D.2

Nils Diaz, M.D.2

Jeff King, B.S.1

1 Department of Surgery, Comprehensive BreastCancer Program, H. Lee Moffitt Cancer Center andResearch Institute, University of South Florida,Tampa, Florida.

2 Department of Pathology, Comprehensive BreastCancer Program, H. Lee Moffitt Cancer Center andResearch Institute, University of South Florida,Tampa, Florida.

3 Department of Biostatistics, ComprehensiveBreast Cancer Program, H. Lee Moffitt CancerCenter and Research Institute, University of SouthFlorida, Tampa, Florida.

Address for reprints: Charles E. Cox, M.D., Com-prehensive Breast Program, H. Lee Moffitt CancerCenter and Research Institute, 12902 MagnoliaDrive, Tampa, FL 33612; Fax: (813) 979-7287;E-mail: [email protected]

Received June 2, 2004; revision received July 26,2004; accepted September 8, 2004.

BACKGROUND. The current report provides results from a large retrospective anal-

ysis of intraoperative imprint cytology performed on axillary sentinel lymph nodes

(IICN) removed over the course of 2137 breast surgeries (4905 lymph nodes). It is

hoped that these results may serve as benchmarks for those interested in using this

technique.

METHODS. The current study included 2078 patients with T1–2 invasive breast

carcinoma who underwent sentinel lymph node biopsy (SLNB) and IICN. Lymph

nodes were bivalved, imprinted, stained with Diff-Quik (Baxter Diagnostics,

McGaw Park, IL), and reviewed by a cytopathologist. A positive intraoperative

diagnosis led to immediate complete axillary lymph node dissection (CALND). On

final pathology, lymph nodes found to be negative on hematoxylin and eosin

staining were submitted for cytokeratin staining.

RESULTS. Of the 2137 cases for which SLNB was performed, 673 were found to have

positive lymph node status on final pathology. Of these 673 cases, 359 were

identified by IICN, resulting in a sensitivity rate of 53.3%. The specificity and overall

accuracy rates for this technique were 99.5% and 85.0%, respectively. In IDC cases,

IICN had a sensitivity rate of 55.5%, compared with 38.7% in ILC cases. Based on

these results, the reoperative CALND rate was calculated to be approximately

14.7%, with 54.5% of these reoperative procedures being performed for cases in

which lymph nodes positive only for micrometastases were found. Macrometas-

tasis-positive lymph nodes that went undetected by IICN were present in only 154

of the 2137 cases examined (7.2%).

CONCLUSIONS. IICN accurately predicts final lymph node status in 85.0% of pa-

tients. Although the accuracy of this technique varies with tumor size and type,

IICN remains a time-efficient and cost-effective adjunct to SLNB. Cancer (Cancer

Cytopathol) 2005;105:13–20. © 2004 American Cancer Society.

KEYWORDS: breast carcinoma surgery, cytopathology, imprint cytology, micro-metastasis.

BACKGROUNDIntraoperative assessment of sentinel lymph nodes to ascertain lym-phatic metastases and the resulting need for axillary dissection isdifficult at best. Many strategies have been investigated, includingextensive evaluation of intraoperative frozen sections, intraoperativeimprint cytology (IICN), assessment of intraoperative imprints of mul-tiple lymph node sections, and intraoperative imprint cytology withthe adjunct of immediate immunohistochemical staining for cytoker-atin (CK).1–30 None of these techniques is satisfactory with regard tothe following three considerations: speed, reliability, and cost-effec-tiveness.

13CANCERCYTOPATHOLOGY

© 2004 American Cancer SocietyDOI 10.1002/cncr.20738Published online 16 December 2004 in Wiley InterScience (www.interscience.wiley.com).

In the current report, we provide results from arelatively large retrospective analysis of intraoperativeimprint cytology of axillary sentinel lymph nodes re-moved during breast surgery (2137 cases, for a com-bined total of 4905 lymph nodes), with the hope thatour findings may serve as a benchmark for those in-terested in using this technique.

MATERIALS AND METHODSThe medical charts of 3355 patients, comprising3460 cases (105 patients had procedures performed onboth breasts), were reviewed. All of the abovemen-tioned patients had IICN performed between July 1996and May 2003 at the H. Lee Moffitt Cancer Center andResearch Institute (University of South Florida,Tampa, FL) after having undergone sentinel lymphnode biopsy (SLNB) during lumpectomy or mastec-tomy.

Only patients with pathologically staged T1–2 in-vasive tumors, in which metastatic disease is moredifficult to detect, were included in the study; a total of94 T3– 4 tumors were excluded from the study. Despiteundergoing SLNB, patients with ductal carcinomain situ (DCIS) were excluded from the study, as wechose to focus on invasive tumors. At our institution,patients with DCIS receive SLNB as the standard ofcare, because approximately 13% of all DCIS cases areupstaged to invasive disease by the time of finalpathologic assessment. The total number of DCIScases excluded from the study totaled 565. Also ex-cluded from the study were 2 cases of metaplasticcarcinoma, 11 cases of Paget disease, and 16 cases oflobular carcinoma in situ. All such cases were ex-cluded so that the study could focus on the three mostprominent types of invasive breast carcinoma: inva-sive ductal carcinoma (IDC), invasive lobular carci-noma (ILC), and mixed IDC/ILC. Other cases wereexcluded from the study for the following reasons:negative findings on hematoxylin and eosin (H & E)staining with no subsequent CK staining (n � 8), theabsence of a request for IICN on the part of the sur-geon (n � 275), and the absence of IICN in benignprophylactic cases in which sentinel lymph node map-ping had been performed (n � 145). An additional 131patients who were also involved in the national Amer-ican College of Surgeons Oncology Group Z10 studyand who underwent IICN were excluded from thestudy because they did not undergo CK testing at ourinstitution, as mapping was used only to confirm thepresence of extraaxillary disease in such patients. Inaddition, 76 individuals with preoperatively clinicallypalpable lymph nodes were excluded, as lymph nodemetastases were already suspected in roughly 60% ofthis subgroup. Patients with invasive T1–2 tumors and

macroscopically positive lymph nodes were includedin the study.

After exclusions, 2074 women and 4 men, for atotal of 2078 patients and a combined total of2137 cases (59 patients underwent procedures on bothbreasts), qualified for the current study. An average of2.3 lymph nodes per case were subjected to IICN; thus,4905 lymph nodes were analyzed in the current studypopulation. (The number of lymph nodes excisedranged from 1 to 12 per case.) Patients ranged in agefrom 21 to 92 years, with a mean age of 58.6 years.

SLNB ProcedureSince April 1994, the detection of sentinel lymphnodes at our institution has been performed usingboth isosulfan blue dye and radioactive colloid. Inpatients undergoing lumpectomy, sentinel lymphnode dissection was performed at the beginning of theprocedure, and these lymph nodes were sent to cyto-pathologists before the removal of the tumor. In pa-tients undergoing mastectomy, SLNB generally wasperformed simultaneously to eliminate the need fortwo incisions. Details regarding the SLNB procedure atour institution have been reported previously.31–33

IICN results obtained from cytopathologists were usedby surgeons as the primary intraoperative tool fordetermining whether total axillary lymph node dissec-tion would be performed.

IICN ProcedureAfter removal, each sentinel lymph node was wrappedin a saline-moistened Telfa pad (Kendall Health CareProducts, Mansfield, MA) and placed inside a clean,capped container. Lymph nodes were immediatelysubmitted to the pathology laboratory, where an as-sistant, a fellow, or one of six cytopathologists per-formed IICN. The standard procedure was to bisect thelymph node along its longest axis.

Following sectioning, a prelabeled glass slide wasapplied gently but firmly over each lymph node sec-tion. The imprint was then rapidly air-dried andstained with Diff-Quik solution (Baxter Diagnostics,McGaw Park, IL) and inspected under a light micro-scope by one of six cytopathologists.

The attending cytopathologist assigned eachlymph node one of the following four classifications:positive, suspicious, atypical, or negative. An intraop-erative ‘positive’ classification typically led to totalaxillary lymph node dissection. Treatment of suspi-cious findings varied from surgeon to surgeon. Forcases with suspicious findings, further intraoperativeevaluation of frozen sections or intraoperative CKstaining was typically performed, either at the requestof the surgeon or at the discretion of the cytopatholo-gist. (Details regarding the intraoperative CK staining

14 CANCER (CANCER CYTOPATHOLOGY) February 25, 2005 / Volume 105 / Number 1

procedure are described elsewhere.34) Because themajority of breast surgeons at our institution did notconsider suspicious findings on IICN alone to begrounds for total axillary lymph node dissection, sus-picious findings were treated as negative findings inthe current study. Atypical results also were consid-ered negative by surgeons. Immediate total axillarylymph node dissection was not performed for patientswith negative findings on IICN; final pathology resultsfrom CK and H & E analyses of lymph node statuswere used to make final decisions regarding delayedcompletion axillary lymph node dissection (CALND)in such patients. Lymph nodes were placed in 10%neutral buffered formalin for embedding in paraffin,and permanent histologic findings made after IICN

were phoned to the treating surgeon.

Final Pathologic EvaluationSections of each sentinel lymph node were evaluatedby one of nine pathologists during the study period.All sentinel lymph nodes were evaluated at one levelor more with standard H & E staining. Immunohisto-chemical CK staining for micrometastases was per-formed on lymph nodes when final pathologic H & Estaining revealed no metastatic disease. In the currentstudy, micrometastases were defined as metastaticcell clusters measuring � 2 mm in greatest diameter.

Statistical MethodsEach lymph node and case was classified as beingtrue-positive (TP), true-negative (TN), false-negative(FN), or false-positive (FP). True-positive cases werethose that were found to contain carcinoma on IICN

and also on subsequent H & E and CK staining. Thefollowing formulas were used to calculate statisticalparameters:

Sensitivity � TP/(TP � FN)

Specificity � TN/(TN � FP)

Overall Accuracy � (TP � TN)/

(TP � FP � TN � FN)

RESULTSBecause 59 patients underwent procedures on bothbreasts, the 2078 patients included in the currentstudy accounted for 2137 cases. IDC accounted for1800 of these cases (84.2%), ILC accounted for 172(8.0%), and mixed IDC/ILC accounted for 165 (7.7%).Results on the case level can be found in the text andtables; results on the lymph node level are presentedin the tables only. IICN findings were compared withfinal pathology results (H & E and CK), which served as

the gold standard. The distribution of disease typesincluded in the study is presented in Table 1.

IICN Results for All Cases of Invasive T1–2 DiseaseIn the current study, macrometastasis-positive lymphnodes that went undetected on IICN were present inonly 154 of 2137 cases (7.2%). Based on final pathol-ogy, 673 cases (31.5%) had lymph nodes that werepositive for carcinoma. IICN detected carcinoma in359 of these 673 positive cases (sensitivity, 53.3%);thus, immediate CALND was performed for approxi-mately 16.8% of all cases. The specificity of IICN in theassessment of lymph nodes in all invasive T1–2 tumorswas 99.5%, and the overall accuracy rate was 85.0%(Tables 2, 3).

Of the 673 cases in which positive lymph nodeswere present, 171 (25.4%) had at least 1 sentinel lymphnode that was positive for micrometastases but had nomacrometastases (� 2 mm) in any other sentinellymph node. IICN identified 11 of these 171 cases, fora sensitivity of 6.4% in the detection of micrometas-tases in IDC, ILC, and mixed IDC/ILC.

Overall, 502 of the 673 total lymph node–positivecases (74.6%) had at least 1 sentinel lymph node thatwas positive for macrometastases. IICN identified 348of these 502 cases, resulting in a total sensitivity of69.3% in the detection of macrometastases.

Complete IICN results for IDC, ILC, and mixedIDC/ILC cases are reported in Tables 3– 6.

TABLE 1Distribution of Histologic Types among T1–2 Breast Carcinomas

Histologic type No. of cases (%) No. of lymph nodes (%)

IDC 1800 (84.2) 4139 (84.4)Mixed IDC/ILC 172 (8.0) 393 (8.0)ILC 165 (7.7) 373 (7.6)Total 2137 4905

IDC: invasive ductal carcinoma; ILC: invasive lobular carcinoma.

TABLE 2Results of Imprint Cytology of Lymph Nodes from Patients with AllBreast Carcinoma Types

Imprint cytologydiagnosis

Final histopathologic diagnosis

By case By lymph node

Positive (%) Negative (%) Positive (%) Negative (%)

Positive 359 (16.8) 7 (0.3) 464 (9.5) 14 (0.3)Negative 314 (14.7) 1457 (68.2) 457 (9.3) 3970 (80.9)Total 2137 4905

IIC for SLN Evaluation in Breast Carcinoma/Cox et al. 15

IICN Results for Cases of T1–2 IDCIDC accounted for 1800 of the 2137 cases included inthe current study (84.2%). Based on final pathologicfindings, lymph nodes that were positive for carci-noma were present in 557 of these 1800 cases (30.9%).IICN identified 309 of these 557 cases, for a sensitivityof 55.5%. The specificity of IICN in the assessment oflymph nodes in IDC cases was 99.5%, and the overallaccuracy rate was 85.9% (Tables 3, 4).

Of the 557 IDC cases in which positive lymphnodes were present, 138 (24.8%) had at least 1 sentinellymph node that was positive for micrometastases but

had no macrometastases in any other sentinel lymphnode. IICN identified 10 of these 138 cases, resulting ina sensitivity of 7.2% for this technique in the detectionof micrometastases in IDC.

At least 1 sentinel lymph node that was positivefor macrometastases was present in 419 of the 557lymph node–positive cases of IDC. IICN identified 299of these 419 cases, for a total sensitivity of 71.4% in thedetection of IDC macrometastases. Complete IICN re-sults for IDC cases are presented in Tables 3 and 4.

IICN Results for Cases of T1–2 ILCILC accounted for 172 of the 2137 cases included inthe current study (8.0%). Final pathology showed thatpositive lymph nodes were present in 62 of these172 cases (36.0%). IICN identified 24 of the 62 lymphnode–positive cases of ILC, for a sensitivity of 38.7%.The specificity of IICN in the assessment of lymphnodes in ILC cases was 100.0%, and the overall accu-racy rate was 77.9% (Tables 3, 5).

Of the 62 ILC cases in which positive lymph nodeswere present, 19 (30.7%) had at least 1 sentinel lymphnode that was positive for micrometastases but had nomacrometastases in any sentinel lymph node. IICN

identified 1 of these 19 cases, for a sensitivity of 5.3%in the detection of ILC micrometastases.

At least 1 sentinel lymph node that was positivefor macrometastases was present in 43 of the 62 lymphnode–positive cases of ILC (69.4%). IICN identified23 of these 43 cases, resulting in a total sensitivity of53.5% for this technique in the detection of ILCmacrometastases.

Complete IICN results on the lymph node level forILC cases are presented in Tables 3 and 5.

IICN Results for Cases of T1–2 Mixed IDC/ILCMixed IDC/ILC accounted for 165 of the 2137 casesincluded in the current study (7.7%). Final pathologyidentified at least 1 positive sentinel lymph node in54 of these 165 cases (32.7%). IICN identified 26 of the54 lymph node–positive cases of mixed IDC/ILC, for asensitivity of 48.1% (Table 3). The specificity of IICN in

TABLE 3Sensitivity and Overall Accuracy of Imprint Cytology of Lymph Nodeson the Case and Lymph Node Levels

Tumor typeNo. of cases/lymph nodes

Sensitivity(%)

Specificity(%)

Overallaccuracy (%)

By caseIDC 1800 55.5 99.5 85.9Mixed 165 48.1 99.1 82.4ILC 172 38.7 100.0 77.9Total 2137 53.3 99.5 85.0

By lymph nodeIDC 4139 52.7 99.7 91.1Mixed 373 44.4 99.3 87.4ILC 393 35.3 99.7 85.8Total 4905 50.4 99.6 90.4


TABLE 4Results of Imprint Cytology of Lymph Nodes for Patients withInvasive Ductal Carcinoma






TABLE 5Results of Imprint Cytology of Lymph Nodes for Patients withInvasive Lobular Carcinoma






TABLE 6Results of Imprint Cytology of Lymph Nodes for Patients with MixedInvasive Ductal Carcinoma/Invasive Lobular Carcinoma







the assessment of lymph nodes in mixed IDC/ILC caseswas 99.1%, and the overall accuracy rate was 82.4%.

Of the 54 mixed IDC/ILC cases in which positivelymph nodes were present, 14 (25.9%) had at least1 sentinel lymph node that was positive for microme-tastases but had no macrometastases in any othersentinel lymph node. IICN was unable to detect mi-crometastases in any of these 14 cases.

At least 1 sentinel lymph node that was positivefor macrometastases was present in 40 cases of mixedIDC/ILC. IICN identified 26 of these 46 cases, resultingin a total sensitivity of 65.0% for this technique in thedetection of macrometastases in mixed IDC/ILC.

Complete IICN results on the lymph node level formixed IDC/ILC cases are presented in Tables 3 and 6.

IICN Accuracy in Invasive T1 Tumors Compared withInvasive T2 TumorsOf the 2137 cases included in the current study, 1604were T1 tumors, and the remaining 533 were T2 le-sions. For invasive T1 tumors taken as a whole, IICN

had a sensitivity of 46.5% and a specificity of 99.6%; forinvasive T2 tumors, the corresponding rates were63.0% and 99.2%, respectively (Table 7). In invasive T1IDC, IICN had a sensitivity of 49.1% and a specificity of99.6%, compared with 65.9% and 99.0%, respectively,in invasive T2 IDC. With regard to invasive ILC, IICN

had a sensitivity of 32.3% and a specificity of 100.0% inT1 tumors, compared with 45.2% and 100.0%, respec-tively, in T2 tumors. Finally, in invasive T1 mixedIDC/ILC, IICN had a sensitivity of 41.4% and a speci-ficity of 98.8%, whereas this technique had a sensitiv-ity of 56.0% and a specificity of 100.0% in invasive T2mixed IDC/ILC (Figs. 1, 2; Table 7).

Time ConsiderationsCytopathologists at our institution typically renderedtheir intraoperative diagnoses within 13–18 minutes of

specimen receipt (depending on the number of lymphnodes submitted), for an average of 2.33 minutes perlymph node. Generally, IICN did not increase the totalsurgery time for patients undergoing lumpectomy orexcisional biopsy, as surgeons typically submittedlymph nodes before removing the breast lesion. Inaddition, IICN did not typically cause a significantincrease in total surgery time for patients undergoingmastectomy, as IICN was usually performed inconjunction with, and during, hemostasis and inci-sional closure.

DISCUSSIONOur institution’s extensive experience with IICN showsthat this technique is an accurate, practical, morbi-dity-reducing, and cost-efficient procedure. At theH. Lee Moffitt Cancer Center, IICN is preferred over fro-zen sections for the intraoperative assessment of lymphnodes, because it is faster and less prone to artifactualresults. Using H & E staining with CK immunohisto-chemistry as the gold standard (Figs. 3–6), the overallaccuracy of IICN in invasive T1–2 tumors assessed at ourinstitution between July 1996 and May 2003 was foundto be 85.0%, with a sensitivity rate of 53.3%. Despite itslimitations in identifying micrometastases, IICN elimi-nated the need for a second surgical procedure in ap-proximately 83.6% of all patients, resulting in a reopera-tive CALND rate of approximately 14.7% (314 of 2137),with 54.5% of these procedures (171 of 314) being per-formed for cases in which lymph nodes positive only for

TABLE 7Accuracy of Imprint Cytology of Lymph Nodes by Tumor Size

Tumor size No. of cases Sensitivity (%) Specificity (%)

T1IDC 1368 49.1 99.6ILC 121 32.3 100.0Mixed 115 41.4 98.8Total 1604 46.5 99.6

T2IDC 432 65.9 99.0ILC 51 45.2 100.0Mixed 50 56.0 100.0Total 533 63.0 99.2


FIGURE 1. Sensitivity of intraoperative imprint cytology in T1 tumors

compared with T2 tumors.

FIGURE 2. Specificity of intraoperative imprint cytology in T1 tumors

compared with T2 tumors.


micrometastases were found (Fig. 6). In addition, theimplementation of IICN is straightforward, requiring thesimple armamentarium of a trained and interested cy-topathologist who is willing to render immediate intra-operative analyses of imprinted lymph nodes.

To our knowledge, few institutions have pub-lished large studies regarding the accuracy of IICN.Furthermore, it is difficult to compare data acrossmuch of the current literature, due to large differ-ences in case types. Published IICN sensitivity ratesrange from 33% to 96%.1– 4,6 – 8,10,11,13,16,18,20 –22,29,35

Creager et al.13 analyzed the accuracy of IICN strictlyon the case level, which yields a higher sensitivitycompared with lymph node–level analyses.13 In ad-dition, Creager et al.,13 Lee et al.,6 Cserni,11 Ratan-awichitrasin et al.,22 and Karamlou et al.,7 did notstate whether patients with clinically positive lymphnodes were included in their studies. (The inclusionof such patients could affect the accuracy of IICN.)Finally, using H & E alone as their gold standard,Kane et al.35 reported a sensitivity of 41% on thecase level; unlike others, Kane and colleagues didnot include macroscopically positive lymph nodesin their sensitivity calculations.

Other discrepancies in the reported accuracy of IICN

can be attributed to variations in tumor size and tumortype distributions as well as differences in gold standardsused (H & E or CK). For instance, some studies includedonly invasive T1–2 tumors, whereas others included in-vasive tumors of all sizes. In addition, Lee et al.,6 Henry-Tillman et al.,8 Llatjos et al.,4 and Van Diest et al.1 all usedH & E alone as a gold standard, whereas most otherstudies used H & E in conjunction with CK immunohis-tochemistry.2,10,11,13,16,21,29

Two apparent outliers in terms of reported ac-curacy are the studies of Motomura et al.2 and Hen-ry-Tillman et al.8 Motomura and colleagues, whoexcluded clinically positive lymph nodes and ana-lyzed lymph nodes from patients with T1–2 tumorsusing H & E alone or H & E with CK immunostainingas a reference, reported a remarkable sensitivity rateof 92% on the lymph node level.2 Henry-Tillman andcolleagues, who used H & E alone as a gold standard,reported a sensitivity of 89.2% in T1–2 tumors, a ratethat is much higher than corresponding rates inother studies involving similar standards.8 Theseoutliers may be partially attributable to sectioningprotocols. Lymph nodes were sectioned at 2 mmthickness and 2–3 mm thickness during imprint cy-tology by Motomura et al.2 and Henry-Tillman etal.,8 respectively. In contrast, numerous other stud-ies, including the current one, typically bisectedlymph nodes or cut sections measuring � 2 mm inthickness.1,7,11,13,22,29,35 This difference could ex-plain the observed gap in sensitivity, although other

studies that cut lymph nodes into � 2 mm sec-tions4,6,16,21 did not achieve the sensitivities re-ported by Motumura et al.2 and Henry-Tillman etal.8 Cserni10 recently reported that increasing thesurface area sampled improves the accuracy of IICN;however, little has been done to confirm this find-ing. Further investigation and a more uniform ap-proach to IICN are needed to clarify the apparentsensitivity discrepancies among published studiesof this technique.

Outside of a limited number of institutions, IICN

does not appear to have been accepted by the medicalcommunity as a whole. At our institution, assessmentof frozen sections was found to be slower, more con-ducive to artifactual findings, and more expensivecompared with IICN.

In the current study, the accuracy of IICN in iden-tifying metastases did vary according to tumor type. InIDC cases, IICN had a sensitivity of 55.5%, and thecorresponding rate in mixed IDC/ILC cases was 48.1%.In contrast, the sensitivity of IICN in the detection oflymph node metastases in patients with lobular carci-noma was only 38.7% (Tables 6, 7). Nonetheless, de-spite the finding that ILC metastases are more difficultto detect compared with metastases from other breasttumor types,14 ILC metastases were identified by IICN

in 24 of 62 cases (38.7%) in the current study, thuseliminating the need for a second surgical procedure.In contrast to the findings reported by Creager et al.,14

the difference in IICN sensitivity between IDC and ILCproved to be statistically significant in the currentstudy (chi-square test: P � 0.012).

CONCLUSIONSOur institution has used IICN to evaluate sentinellymph nodes since 1997. The current study demon-strates that this technique is practical, efficient, andmorbidity-reducing, resulting in a reoperativeCALND rate of approximately 14.7% (314 of 2137),with 54.5% of these procedures (171 of 314) beingperformed for cases in which lymph nodes positiveonly for micrometastases were found. The sensitiv-ity of IICN in detecting metastatic tumor cells withinthe sentinel lymph nodes of T1–2 breast tumors isless than the sensitivity associated with exfoliativeand aspiration biopsy cytology, due in large part tosampling and interpretive errors. Nonetheless, theimplementation of IICN at our institution has suc-cessfully decreased rates of second surgery whilepreserving all available lymph node tissue for thesubsequent detection of micrometastases in perma-nent sections. The rapid evaluation of lymph nodestatus, although necessitating a return to the oper-ating room for further lymph node removal in 14.7%of all cases, allows the more efficient use of operat-


ing room time compared with frozen section evalu-ation. In the current study, macrometastasis-posi-tive lymph nodes that went undetected by IICN werepresent in only 154 of 2137 cases (7.2%).

Because IICN is not accurate in the detection ofmicrometastases, intraoperative immunohistochemi-cal staining and step-sectioning of lymph nodes withimprinting of all faces may increase the diagnosticability of pathologists and reduce the number of sec-ond procedures needed for patients with breast carci-noma. We and others20 –21 have used a modified andrapid intraoperative immuno-CK procedure to en-hance the diagnostic accuracy of IICN, particularly inpatients with ILC. (IICN was found to be less sensitivein patients with ILC than in patients with IDC.) Selec-

tive but uniform application of this intraoperative im-muno-CK technique and a better understanding of theprognostic relevance of micrometastases will furtherclarify the role of IICN immunocytochemistry in thedetection of metastatic disease.

RECOMMENDATIONSIICN should be performed for all lymph nodes inwhich there is clinical suspicion of metastatic disease.Selective rapid intraoperative immunohistochemicalanalysis of cytologic imprints can be useful, and suchanalysis is necessary for patients with ILC. IICN is notrequired for patients with DCIS or for patients receiv-ing prophylactic mastectomy unless there is clinicalsuspicion of lymph node metastases.

FIGURE 3. Malignant cells with minimal nuclear atypia from a low-grade

ductal carcinoma. This sentinel lymph node was interpreted as being positive

on intraoperative imprint cytology. Original magnification �60; Diff-Quik stain-

ing (Baxter Diagnostics, McGaw Park, IL).

FIGURE 4. High-grade ductal carcinoma cells showing nuclear membrane

convolutions. Original magnification �60; Diff-Quik staining (Baxter Diagnos-

tics, McGaw Park, IL).

FIGURE 5. Cells found on retrospective review of a touch imprint of a sentinel

lymph node intraoperatively categorized as being negative. Original magnifi-

cation �60; Diff-Quik staining (Baxter Diagnostics, McGaw Park, IL).

FIGURE 6. Rare cytokeratin-positive tumor cells. Original magnification �60;

avidin-biotin complex immunoperoxidase staining.


REFERENCES1. Van Diest PJ, Torrenga H, Borgstein PJ, et al. Reliability of

intraoperative frozen section and imprint cytological inves-tigation of sentinel lymph nodes in breast cancer. Histopa-thology. 1999;35:14 –18.

2. Motomura K, Inaji H, Komoike Y, et al. Intraoperative sentinellymph node examination by imprint cytology and frozen sec-tioning during breast surgery. Br J Surg. 2000;87:597–601.

3. Menes TS, Tartter PI, Mizrachi H, Smith SR, Estabrook A.Touch preparation or frozen section for intraoperative de-tection of sentinel lymph node metastases from breast can-cer. Ann Surg Oncol. 2003;10:1166 –1170.

4. Llatjos M, Castella E, Fraile M, et al. Intraoperative assess-ment of sentinel lymph nodes in patients with breast carci-noma: accuracy of rapid imprint cytology compared withdefinitive histologic workup. Cancer. 2002;96:150 –156.

5. Liang R, Craik J, Juhasz ES, Harman CR. Imprint cytologyversus frozen section: intraoperative analysis of sentinellymph nodes in breast cancer. ANZ J Surg. 2003;73:597–599.

6. Lee A, Krishnamurthy S, Sahin A, Symmans WF, Hunt K,Sneige N. Intraoperative touch imprint of sentinel lymphnodes in breast carcinoma patients. Cancer. 2002;96:225–231.

7. Karamlou T, Johnson NM, Chan B, Franzini D, Mahin D.Accuracy of intraoperative touch imprint cytologic analysisof sentinel lymph nodes in breast cancer. Am J Surg. 2003;185:425– 428.

8. Henry-Tillman RS, Korourian S, Rubio IT, et al. Intraopera-tive touch preparation for sentinel lymph node biopsy: a4-year experience. Ann Surg Oncol. 2002;9:333–339.

9. Dabbs DJ, Fung M, Johnson R. Intraoperative cytologic ex-amination of breast sentinel lymph nodes: test utility andpatient impact. Breast J. 2004;10:190 –194.

10. Cserni G. Effect of increasing the surface sampled by im-print cytology on the intraoperative assessment of axillarysentinel lymph nodes in breast cancer patients. Am Surg.2003;69:419 – 423.

11. Cserni G. The potential value of intraoperative imprint cy-tology of axillary sentinel lymph nodes in breast cancerpatients. Am Surg. 2001;67:86 –91.

12. Cserni G. Intraoperative sentinel lymph node examinationby imprint cytology and frozen sectioning during breastsurgery. Br J Surg. 2000;87:1596.

13. Creager AJ, Geisinger KR, Shiver SA, et al. Intraoperative eval-uation of sentinel lymph nodes for metastatic breast carci-noma by imprint cytology. Mod Pathol. 2002;15:1140–1147.

14. Creager AJ, Geisinger KR, Perrier ND, et al. Intraoperativeimprint cytologic evaluation of sentinel lymph nodes forlobular carcinoma of the breast. Ann Surg. 2004;239:61– 66.

15. Creager AJ, Geisinger KR. Intraoperative evaluation of sen-tinel lymph nodes for breast carcinoma: current methodol-ogies. Adv Anat Pathol. 2002;9:233–243.

16. Bochner MA, Farshid G, Dodd TJ, Kollias J, Gill PG. Intra-operative imprint cytologic assessment of the sentinel nodefor early breast cancer. World J Surg. 2003;27:430 – 432.

17. Beach RA, Lawson D, Waldrop SM, Cohen C. Rapid immuno-histochemistry for cytokeratin in the intraoperative evaluationof sentinel lymph nodes for metastatic breast carcinoma. ApplImmunohistochem Mol Morphol. 2003;11:45–50.

18. Barranger E, Antoine M, Grahek D, Callard P, Uzan S. Intra-operative imprint cytology of sentinel nodes in breast can-cer. J Surg Oncol. 2004;86:128 –133.

19. Baitchev G, Gortchev G, Todorova A. Intraoperative sentinel

lymph node examination by imprint cytology during breastsurgery. Curr Med Res Opin. 2002;18:185–187.

20. Aihara T, Munakata S, Morino H, Takatsuka Y. Touch im-print cytology and immunohistochemistry for the assess-ment of sentinel lymph nodes in patients with breast cancer.Eur J Surg Oncol. 2003;29:845– 848.

21. Mullenix PS, Carter PL, Martin MJ, et al. Predictive value ofintraoperative touch preparation analysis of sentinel lymphnodes for axillary metastasis in breast cancer. Am J Surg.2003;185:420 – 424.

22. Ratanawichitrasin A, Biscotti CV, Levy L, Crowe JP. Touchimprint cytological analysis of sentinel lymph nodes fordetecting axillary metastases in patients with breast cancer.Br J Surg. 1999;86:1346 –1348.

23. Reichert RA. Use of touch preps for intraoperative diagnosisof sentinel lymph node metastases in breast cancer. AnnSurg Oncol. 1999;6:513.

24. Rubio IT, Korourian S, Cowan C, Krag DN, Colvert M, Klim-berg VS. Use of touch preps for intraoperative diagnosis ofsentinel lymph node metastases in breast cancer. Ann SurgOncol. 1998;5:689 – 694.

25. Salem AA, Douglas-Jones AG, Sweetland HM, Mansel RE.Intraoperative evaluation of axillary sentinel lymph nodesusing touch imprint cytology and immunohistochemistry: I.Protocol of rapid immunostaining of touch imprints. EurJ Surg Oncol. 2003;29:25–28.

26. Liu Y, Silverman JF, Sturgis CD, Brown HG, Dabbs DJ, RaabSS. Utility of intraoperative consultation touch preparations.Diagn Cytopathol. 2002;26:329 –333.

27. Veronesi U, Zurrida S, Mazzarol G, Viale G. Extensive frozensection examination of axillary sentinel nodes to determineselective axillary dissection. World J Surg. 2001;25:806 – 808.

28. Viale G, Bosari S, Mazzarol G, et al. Intraoperative examina-tion of axillary sentinel lymph nodes in breast carcinomapatients. Cancer. 1999;85:2433–2438.

29. Shiver SA, Creager AJ, Geisinger K, Perrier ND, Shen P,Levine EA. Intraoperative analysis of sentinel lymph nodesby imprint cytology for cancer of the breast. Am J Surg.2002;184:424 – 427.

30. Salem AA, Douglas-Jones AG, Sweetland HM, NewcombeRG, Mansel RE. Evaluation of axillary lymph nodes usingtouch imprint cytology and immunohistochemistry. Br JSurg. 2002;89:1386 –1389.

31. Cox CE, Salud C, Whitehead GF, Reintgen DS. Sentinellymph node biopsy for breast cancer: combined dye-isotopetechnique. Breast Cancer. 2000;7:389 –397.

32. Cox CE, Bass SS, McCann CR, et al. Lymphatic mapping andsentinel lymph node biopsy in patients with breast cancer.Annu Rev Med. 2000;51:525–542.

33. Cox CE, Salud CJ, Harrinton MA. The role of selective sen-tinel lymph node dissection in breast cancer. Surg ClinNorth Am. 2000;80:1759 –1777.

34. Weinberg ES, Dickson D, White L, et al. Cytokeratin stainingfor intraoperative evaluation of sentinel lymph nodes inpatients with invasive lobular carcinoma. Am J Surg. 2004;188:419 – 422.

35. Kane JM III, Edge SB, Winston JS, Watroba N, Hurd TC.Intraoperative pathologic evaluation of a breast cancer sen-tinel lymph node biopsy as a determinant for synchronousaxillary lymph node dissection. Ann Surg Oncol. 2001;8:361–367.


Accuracy of intraoperative imprint cytology for sentinel lymph node evaluation in the treatment of...

Documents

Transcript of Accuracy of intraoperative imprint cytology for sentinel lymph node evaluation in the treatment of...