Sentinel lymph nodes with technetium-99m colloidal rhenium sulfide in patients with esophageal carcinoma

Sentinel Lymph Nodes with Technetium-99m ColloidalRhenium Sulfide in Patients with EsophagealCarcinoma

Hiroyuki Kato, M.D., Ph.D.1

Tatsuya Miyazaki, M.D., Ph.D.1

Masanobu Nakajima, M.D.1

Junko Takita, B.A.1

Makoto Sohda, M.D.1

Yasuyuki Fukai, M.D.1

Norihiro Masuda, M.D.1

Minoru Fukuchi, M.D., Ph.D.1

Ryokuhei Manda, M.D., Ph.D.1

Hitoshi Ojima, M.D., Ph.D.1

Katsuhiko Tsukada, M.D., Ph.D.1

Takayuki Asao, M.D., Ph.D.1

Hiroyuki Kuwano, M.D., Ph.D.1

Noboru Oriuchi, M.D., Ph.D.2

Keigo Endo, M.D., Ph.D.2

1 Department of Surgery I, Gunma University Fac-ulty of Medicine, Maebashi, Japan.

2 Department of Nuclear Medicine, Gunma Univer-sity Faculty of Medicine, Maebashi, Japan.

The authors thank A. Nakabayashi, H. Emura, M.Ohnuma, T. Aoyagi, and Y. Saitoh for their excel-lent secretarial assistance and C. Yoshida for herassistance with the data management and biosta-tistical analysis.

Address for reprints: Hiroyuki Kato, M.D., Ph.D.,Department of Surgery I, Gunma University Facultyof Medicine, 3-39-22, Showa-machi, Maebashi,Gunma, 371-8511, Japan; Fax: (011) 81-27-220-8230; E-mail: [email protected]

Received January 31, 2003; revision received April9, 2003; accepted April 30, 2003.

BACKGROUND. The authors assessed the detection of sentinel lymph nodes in

patients with esophageal squamous cell carcinoma (SCC) using technetium-99m

colloidal rhenium sulfide. They studied whether an analysis of sentinel lymph

nodes using cytokeratin (CK) immunohistochemistry increased the accuracy of

staging.

METHODS. The authors observed 25 patients with thoracic esophageal carcinomas

who underwent radical esophagectomy. The day before surgery, technetium-99m

colloidal rhenium sulfide was injected into the submucosa at four sites around the

primary tumor. Lymphoscintigraphy was performed. Esophagectomy and regional

lymph node dissection were performed 17 hours after the technetium-99m injec-

tion. After surgery, the resected lymph nodes were evaluated by CK staining.

RESULTS. Lymphoscintigraphy detected sentinel lymph nodes in 92% of the pa-

tients (23 of 25 patients). The accuracy of sentinel lymph node was 91.3% (21 of 23

patients), the sensitivity was 86.7% (13 of 15 patients), and the false-negative rate

was 8.7% (2 of 23 patients). A comparison of the number of sentinel lymph nodes

and clinicopathologic factors showed that there was a significant association

between the number of sentinel lymph nodes and lymph node status (P � 0.01),

pathologic stage (P � 0.05), and the number of metastatic lymph nodes (P � 0.05).

Occult metastasis was detected by CK staining in 14 (56%) of the 25 patients and

in 23 (1.7%) of 1406 lymph nodes. Because the 2 false-negative (sentinel lymph

node–negative and nonsentinel lymph node–positive) patients who had occult

metastases in the sentinel lymph nodes, the accuracy of sentinel lymph node

evaluation using CK staining was 100% (23 of 23).

CONCLUSIONS. Lymphatic mapping with technetium-99m colloidal rhenium sul-

fide was used to identify the lymphatic basin and was feasible in patients with

esophageal SCC. An analysis of sentinel lymph nodes using CK immunohistochem-

istry increased the accuracy of sentinel lymph node. Cancer 2003;98:932–9.

© 2003 American Cancer Society.

KEYWORDS: sentinel lymph node, esophageal squamous cell carcinoma, occultmetastasis, cytokeratin immunostaining, rhenium sulfide.

The sentinel lymph node is defined as the first lymph node in aregional lymphatic basin that receives lymph flow from a primary

tumor.1,2 According to the sentinel lymph node hypothesis, the find-ing of a histologically negative sentinel lymph node supports thehistologic negativity of the remaining regional lymph nodes, thusavoiding unnecessary lymph node dissection. Feasibility studies ofsentinel lymph node biopsy have been reported for malignanciessuch as breast carcinoma, melanoma, and thyroid carcinoma.3– 6 Sen-tinel lymph node biopsy in breast carcinoma has been evaluated

932

© 2003 American Cancer SocietyDOI 10.1002/cncr.11559Published online 27 June 2003 in Wiley InterScience (www.interscience.wiley.com).

since the early 1990s. In 1992, Morton et al.3 reporteda dye technique for lymphatic mapping and sentinellymph node biopsy in clinically lymph node–negativepatients with melanoma. In 1993, Krag et al.4 reportedsentinel lymph node biopsy using a radiopharmaceu-tical agent in patients with breast carcinoma. In 1994,Giuliano et al.7 reported a similar study using isosulfanblue dye. In 1996, Albertini et al.1 reported a combinedprocedure using both of these agents. Since thesereports, there have been many studies of lymphaticmapping and sentinel lymph node biopsy in breastcarcinoma.8 –29 More detailed evaluation of sentinellymph nodes using serial sectioning or cytokeratin(CK) immunohistochemistry has improved the stagingaccuracy of sentinel lymph node biopsy.8,9,22,27–29 Im-munohistochemical staining of lymph nodes has in-creased the detection of occult lymph node metastasisand upstaged approximately 10% of the overall popu-lation undergoing lymphatic mapping.27,28

In patients with gastrointestinal carcinoma, sen-tinel lymph nodes identified by a radioguided tech-nique had an overall diagnostic accuracy of 96% forregional lymph node metastasis.30 However, therehave been few published clinical studies of sentinellymph nodes in patients with esophageal squamouscell carcinoma (SCC). A detailed evaluation of sentinellymph nodes using immunohistochemical staininghas not been reported.

We assessed the feasibility of using technetium-99m colloidal rhenium sulfide to detect sentinellymph nodes in patients with esophageal SCC. Westudied whether an analysis of sentinel lymph nodesby CK immunohistochemistry improved the stagingaccuracy of sentinel lymph nodes.

MATERIALS AND METHODSPatientsWe studied 25 consecutive patients (20 men and 5women) with thoracic esophageal SCCs who under-went radical esophagectomy between January 1999and December 2001at the Gunma University Facultyof Medicine (Maebashi, Japan). None of these patientshad received preoperative treatment. The median agewas 62.2 years (range, 49 –76 years). Tumor stage anddisease grade were determined using the TNM classi-fication system.

Resectability was determined by conventionalstaging, which included a computed tomographicscan of the neck, chest, and abdomen, a bone scan, anendoscopic ultrasound, and esophagography. Patientswith limited local metastasis who were considered bythoracic surgeons to have resectable disease were in-cluded in the current study. Informed consent was

obtained from all patients before they underwent theexamination.

Surgery and PathologyThe day before surgery, 0.5 mL (0.5 millicuries) oftechnetium-99m colloidal rhenium sulfide (NANOCIS,CIS Bio-international, Gif sur Yvette, France) was in-jected into the submucosa at 4 sites around the pri-mary tumor using esophagoscopy. Before the surgery,anterior and posterior lymphoscintigraphy from theneck to the upper abdomen was performed with alarge-field scintillation camera. Sentinel lymph nodeswere defined as hot lymph nodes identified by twosurgeons at the time of lymphoscintigraphy. Becauseesophageal SCC can show unidirectional or bidirec-tional lymphatic spread, the sentinel lymph node wasassigned to assess the distribution of spread. Theselymph nodes were defined as above (A) if the radio-colloid was localized at a site above the injection siteor below (B) if it was localized at a site below theinjection site.

After lymphoscintigraphy, all patients immedi-ately underwent esophagectomy and regional lymphnode dissection 17 hours after the technetium-99mcolloidal rhenium sulfide injection. Two procedureswere used. In some cases, a standard esophagectomywas performed using the McKeown method (right tho-racotomy followed by laparotomy and neck incisionwith a cervical anastomosis), as well as three-field(thoracoabdominal and cervical; n � 13) lymph nodedissection if indicated. In other cases, the Ivor Lewisesophagectomy (right thoracotomy and laparotomywith anastomosis in the chest) and two-field (thora-coabdominal; n � 12) lymph node dissection wereperformed. In all patients, a complete thoracic esoph-agectomy was performed, including the esophagogas-tric junction. At surgery, the scintigraphic hot spotswere detected in vivo with a hand-held gamma raydetector (Navigator, U.S. Surgical, Norwalk, CT).

After surgery, the lymph nodes were separatedfrom the resected esophagus and adjacent tissue. Theywere assigned specific numbers indicating their local-ization, according to the guidelines of the JapaneseSociety for Esophageal Diseases.31 The average num-ber of resected lymph nodes for patients with three-field and two-field lymph node dissection was 69.8and 41.6, respectively. The surgical specimens werefixed, embedded, stained with hematoxylin and eosin(HE), and evaluated microscopically by two patholo-gists. In the routine pathologic evaluation of dissectedlymph nodes, two pathologists evaluated one cross-section stained with HE. The extent of lymph nodemetastasis and the localization of the lymph nodeswere documented.

Sentinel Lymph Nodes in Esophageal Carcinoma/Kato et al. 933

Immunohistochemical Detection of CytokeratinsTwo extra cross-sections were cut in all dissectedlymph nodes. One section was stained with HE, andthe other was used for immunohistochemical detec-tion of CK. Sections that were 4 �m thick were cutfrom each paraffin block of the 1406 lymph nodes andthen deparaffinized and treated with 0.3% H2O2-meth-anol solution for 30 minutes at room temperature toblock endogenous peroxidase activity. After rehydra-tion through a graded ethanol series, we incubatedeach slide for 30 minutes in 0.1% trypsin solution inCaCl2/phosphate-buffered saline (PBS). After the slidehad been rinsed in 0.1 M PBS (pH 7.4), nonspecificbinding sites were blocked by incubation with 10%normal horse serum for 30 minutes. The sections werethen incubated with primary antibody overnightat 4 °C. The monoclonal antibody cocktail AE1/AE3(Boehringer-Mannheim, Mannheim, Germany) wasdiluted by a ratio of 1:1000. Immunohistochemistrywas performed using the ABC system (Vectastatin,Burlingame, CA). Incubation with the secondary anti-body, biotinylated anti-mouse IgG, was carried out for30 minutes. The chromogen used was 3,3�-diamino-benzidine tetrahydrochloride, applied as a 0.02% so-lution containing 0.005% H2O2 in 50 mM ammoniumacetate– citric acid buffer (pH 6.0). The sections werecounterstained lightly with hematoxylin. Negativecontrols were established by replacing the antibodywith PBS or normal mouse serum.

Statistical Analysis and Evaluation of CytokeratinStainingTwo pathologists who were blind to the results of theroutine pathology experiments evaluated immuno-staining for CK. All newly detected lymph node me-tastases were defined as occult metastases.

The relation between the number of sentinellymph nodes and other parameters was determinedby analysis of variance. Simple regression analysis wasused to determine the correlation between the num-ber of sentinel and metastatic lymph nodes. The sen-sitivity, specificity, and accuracy of sentinel lymphnode were calculated by the standard definitions.32

RESULTSDetection of Sentinel Lymph NodesThe results for the 25 patients in the current study aresummarized in Table 1. The overall sentinel lymphnode detection rate using lymphoscintigraphy was92% (23 of 25 patients; Fig. 1). There were 57 sentinellymph nodes in 23 patients, with a mean of 2.3 lymphnodes in each patient (range, 1–5 lymph nodes).Thirty-nine sentinel lymph nodes were classified as A

and 18 as B. Sentinel lymph nodes were classified as Ain 14 (61%) of 23 patients, B in 5 (22%) of 23 patients,and as both A and B in 4 (17%) of 23 patients. In twopatients (Patients 1 and 2), sentinel lymph nodes werenot detected by lymphoscintigraphy and routinepathologic evaluation did not reveal lymph node me-tastasis. In 8 (Patients 3–10) of the 23 patients in whomwe detected sentinel lymph nodes, lymph node me-tastasis was not detected by routine pathologic evalu-ation. The remaining 15 patients had lymph nodemetastases on routine pathologic evaluation. In 13(Patients 13–25) of these 15 patients, the sentinellymph node was positive, but the remaining 2 patients(Patients 11 and 12) had false-negative results (senti-nel lymph node negative and nonsentinel lymphnode–positive). These two patients had advanced (T3or T4) tumors. The sentinel lymph nodes classified asA were negative, whereas the nonsentinel lymphnodes classified as B were positive. Lymphoscintigra-phy detected these sentinel lymph nodes in the uppermediastinum. The lymph node metastases were lo-cated in the abdominal and paraesophageal regionsadjacent to the primary tumor. The accuracy of sen-tinel lymph nodes was 91.3% (21 of 23), the sensitivitywas 86.7% (13 of 15), and the false-negative rate was8.7% (2 of 23).

Sentinel Lymph Nodes and Clinicopathologic FactorsThe correlation between the number of sentinel lymphnodes and clinicopathologic factors is shown in Table 2.There was a significant association between the numberof sentinel lymph nodes and each of the following fac-tors: tumor location (P � 0.05), lymph node status (P� 0.01), and pathologic stage (P � 0.05). In patients withadvanced tumor status, the number of sentinel lymphnodes was higher. The number of sentinel lymph nodesdid not correlate with gender or histologic grade. A sig-nificant correlation was observed between the numberof sentinel lymph nodes and the number of metastaticlymph nodes (P � 0.05; Fig. 2).

Cytokeratin ImmunohistochemistryOf 1406 lymph nodes evaluated, 55 (3.9%) were posi-tive on routine pathologic evaluation with HE stain-ing. Occult metastasis was detected by CK staining in14 (56%) of 25 patients and in 23 (1.7%) of 1406 lymphnodes (Fig. 3). The results of lymphatic mapping andoccult metastasis are shown in Figure 4. Of the twopatients in whom no sentinel lymph nodes were de-tected, one (Patient 2) had occult metastases in theparaesophageal lymph nodes adjacent to the primarytumor.

Of the 10 patients without sentinel lymph nodemetastasis on routine pathologic evaluation, 5 (50%)

934 CANCER September 1, 2003 / Volume 98 / Number 5

had occult metastases in the sentinel lymph nodes(Table 1). The two patients (Patients 11 and 12) withfalse-negative results (sentinel lymph node–nega-tive and nonsentinel lymph node–positive) on rou-tine pathologic evaluation had occult metastases inthe sentinel lymph nodes. Of the remaining eightpatients (Patients 3–10), three (Patients 6, 7, and 10)had occult metastases in the sentinel lymph nodesand five did not have occult metastasis in the sen-tinel and nonsentinel lymph nodes. An analysis ofsentinel lymph nodes using CK staining did notyield false-negative results (sentinel lymph node–negative and nonsentinel lymph node–positive).The accuracy of sentinel lymph nodes using CKstaining was 100% (23 of 23).

DISCUSSION

Lymphatic mapping has been proposed as an accuratemeans of identifying the regional lymph nodes most

likely to contain tumor cells metastasizing from a pri-mary solid neoplasm via the lymphatics.3,7,8 It isemerging as a useful technique, but currently there isno standard identification method.

In the current study, we demonstrated that lym-phatic mapping with technetium-99m colloidal rhe-nium sulfide detected sentinel lymph nodes in esoph-ageal SCC. The sentinel lymph node detection ratewas 92% and the accuracy was 91.3% on routinepathologic evaluation. The two patients with false-negative results had advanced (T3 or T4) tumors, sug-gesting that patients with early-stage esophageal SCCmight be better suited for sentinel lymph node map-ping. The accuracy of the sentinel lymph node con-cept using CK staining was 100%, because the 2 false-negative patients had occult metastases in the sentinellymph nodes. Based on these findings, we believe thatsentinel lymph node mapping can be used to identifythe lymphatic basin and is a feasible technique in

TABLE 1Results of Sentinel Lymph Node, Lymph Node metastasis, and Cytokeratin Staining in 25 Patients

PatientNo.

Age(yrs) Gender

Location ofprimary tumor pT pN

No. ofSLN

Distributionof the SLNa

Routine pathologic evaluationImmunostaining for CKd

SLN statusbDistribution ofthe LNMc SLN status Non-SLN status

1 70 Male Lower pT1 pN0 0 �2 53 Male Lower pT3 pN0 0 �3 56 Male Lower pT1 pN0 1 A � � �4 63 Male Mid pT1 pN0 1 A � � �5 53 Male Mid pT1 pN0 1 B � � �6 76 Male Mid pT1 pN0 1 A � � �7 69 Male Lower pT1 pN0 1 B � � �8 73 Male Mid pT1 pN0 2 A � � �9 63 Female Mid pT1 pN0 2 A � � �10 59 Male Mid pT1 pN0 3 A, B � � �11 49 Female Lower pT3 pN1 2 A � B � �12 65 Male Mid pT4 pN1 3 A � B � �13 70 Male Mid pT3 pN1 2 A � A � �14 65 Male Upper pT3 pN1 2 A � A � �15 65 Male Lower pT2 pN1 2 B � B � �16 58 Male Mid pT3 pN1 3 A � A, B � �17 55 Male Mid pT1 pN1 5 A � A � �18 57 Female Mid pT2 pN1 2 A � A, B � �19 72 Male Mid pT1 pN1 2 B � B � �20 79 Male Lower pT2 pN1 2 B � B � �21 54 Female Mid pT3 pN1 3 A, B � A � �22 64 Male Mid pT3 pN1 5 A, B � A, B � �23 51 Male Upper pT3 pN1 2 A � A � �24 65 Male Mid pT4 pN1 5 A � A, B � �25 51 Female Mid pT3 pN1 5 A, B � A, B � �

SLN: sentinel lymph node; CK: cytokeratin; LNM: lymph node metastasis; A: above; B: below.a The localization of the sentinel lymph node was defined as above (A) or below (B) the injection site of the radionuclide.b Sentinel lymph node metastases were indicated pathologically as negative (�) or positive (�).c The localization of lymph node metastasis was defined as above (A) or below (B) the site of the primary tumor.d Lymph node metastases identified by pathology and cytokeratin immunostaining were indicated as negative (�) or positive (�).


patients with esophageal SCC. An analysis of sentinellymph nodes using CK staining increased the accuracyof sentinel lymph node staging. However, we suggestthat the sentinel lymph node concept might be moreadaptable to patients with early-stage esophagealSCC, because radiocolloid is injected into the submu-cosa of the esophageal wall, making it difficult to coverthe lymphatic basin in the deepest region of advanced(T3 or T4) tumors.

A comparison of the number of sentinel lymphnodes and clinicopathologic factors demonstrated asignificant association between the number of senti-nel lymph nodes and each of the following factors:tumor location, lymph node status, pathologic stage,and the number of metastatic lymph nodes. The iden-tification rate of sentinel lymph nodes was slightlylower for lower esophageal tumors than for upperesophageal and mid-esophageal tumors, because ra-dioactivity was detected occasionally in the liver. Moreadvanced esophageal SCCs were associated with ahigher number of sentinel lymph nodes. Several rea-sons account for this finding. First, because the sites ofradiocolloid injection are separated in advanced tu-mors, the regional lymphatic drainage from theesophagus shows a widespread distribution. Second,primary tumor tissue might destroy the lymphatic

drainage from the submucosal layer at the injectionsites around an advanced tumor. Similarly, lymphaticdrainage from metastatic lymph nodes might also bedestroyed.33

Many investigators have reported results obtainedwith lymphatic mapping and sentinel lymph nodebiopsy in patients with breast carcinoma. Currently,staining with blue dye, radiocolloid, or both is used forsentinel lymph node biopsy in these patients, yieldingdetection rates ranging from 66% to 98%,7–14 from 82%to 98%,15–22 and from 90% to 96%, respectively.23–29

Generally, radiocolloid localization appears to detectsentinel lymph nodes more frequently than vital bluedye. Therefore, a combination of the two may yieldbetter results.24 –26 In a study of gastrointestinal tu-mors, lymphatic mapping with isosulfan blue dyeidentified sentinel lymph nodes in 62 of 65 patients. Ofthese 62 patients, 4 with rectal tumors had lymphnode metastases identified only in the nonsentinellymph nodes.34 In colon carcinoma, the accuracy ofsentinel lymph nodes using the blue dye methodranges from 94% to 100%.33,35 Owing to the wide-spread distribution of lymph node metastasis inesophageal SCC, it is difficult to detect multiple sen-

FIGURE 1. Preoperative lymphoscintigraphy shows a sentinel lymph node

(arrow) in the upper mediastinum with intensive radioactivity and a primary

tumor (arrowhead) in the lower esophagus.

TABLE 2Sentinel Lymph Node and Clinicopathologic Characteristics in 25Patients with Esophageal Carcinomas

Characteristic No. of patientsNo. of sentinel lymph nodes(mean � SD) P value

GenderMale 20 2.15 � 1.49 0.3840Female 5 2.80 � 1.30

LocationUpper 2 2.00 � 0.00 0.0317Middle 16 2.81 � 1.47Lower 7 1.14 � 0.90

Tumor statuspT1 11 1.73 � 1.35 0.1620pT2 3 2.00 � 0.00pT3 9 2.67 � 1.58pT4 2 4.00 � 1.41

Lymph node statuspN0 10 1.20 � 0.92 0.0010pN1 15 3.00 � 1.31

Pathologic stageI 9 1.33 � 0.87 0.0432II 5 2.20 � 1.79III 6 3.33 � 1.37IV 5 2.80 � 1.30

Histologic gradeG1 9 1.78 � 1.56 0.4236G2 9 2.44 � 1.24G3 7 2.71 � 1.60

a SD: standard deviation; G1: well-differentiated carcinoma; G2: moderately differentiated carcinoma;

G3: poorly differentiated carcinoma.


tinel lymph nodes when vital blue dye is used as thesole mapping agent. Sentinel lymph node biopsy isdifficult because vital blue dye rapidly diffuses intosentinel lymph nodes in the cervical and abdominalregions. Kitagawa et al.36 reported that sentinel lymphnodes were detected with technetium-99m tin colloidsolution in 14 (88%) of 16 patients with esophagealSCC, which was similar to the detection rate in thecurrent study. Yasuda et al.37 reported that lymphnode metastases were observed in 12 of 23 patientswith esophageal carcinoma and that sentinel lymphnodes were negative for metastasis in 3 of the 12patients.These were patients with esophageal carci-noma whose tumor invasions extended beyond themuscle layer (T3), a finding similar to the results in thecurrent study.

When lymphatic mapping is performed with aradiolabeled colloid, the choice of radiocolloid is vital,because distribution depends greatly on particlesize.21 Radiocolloids for lymphatic mapping shouldhave an appropriate particle size to permeate the lym-phatic vessels and remain in the sentinel lymphnodes. The ideal radiocolloid particle size is reportedlybetween 10 and 100 nm or between 10 and 200 nm.2

We assessed lymphatic mapping using technetium-99m colloidal rhenium sulfide with a mean particlesize of 100 nm (range, 50 –200 nm). The particles oftechnetium-99m human serum albumin are too small(� 5 nm) to be retained in the sentinel lymph nodes,and the particles of technetium-99m tin colloid aretoo large (� 500 nm) to migrate through the lymphaticvessels.26 However, it has been reported that small-particle colloids (� 50 nm) are often trapped in several

FIGURE 2. Relation between the number of sentinel lymph nodes and

metastatic lymph nodes. A significant correlation was observed between these

two variables (P � 0.05).

FIGURE 3. Representative photograph of a positive sentinel lymph node on

cytokeratin immunohistochemical staining. A malignant cell cluster in a peripheral

sinus was stained. (A) Hematoxylin-eosin staining. (B) Cytokeratin staining.

FIGURE 4. Results of sentinel lymph node analysis by lymph node status and

occult metastasis. pN: pathologic lymph node metastasis; OM: occult metas-

tasis detected by cytokeratin immunostaining.


lymph nodes, whereas larger-particle colloids (200 –1000 nm) are often taken up by 1 lymph node only.38

In the current study, an average of 2.3 sentinel lymphnodes were detected in each patient. This is becauseof the widespread distribution of the regional lym-phatics from the esophagus, extending to the cervical,mediastinal, and abdominal regions. Watanabe et al.21

reported that the lymphoscintigraphy positivity rateobtained with colloidal rhenium sulfide was 100%,suggesting that this agent is particularly suitable forsentinel lymph node mapping in breast carcinoma.They reported that each patient had a mean of 2sentinel lymph nodes (range, 1–5 sentinel lymphnodes).

To clarify the optimum interval between injectionof the radiocolloid and detection of sentinel lymphnodes, we performed lymphoscintigraphy 2, 4, 6, and17 hours after injection in some patients. Hot spotswere seen on the 2-hour scan and remained identifi-able up to the 17-hour scan (data not shown). No newhot spots were detected in the 17-hour scan and theradiocolloid did not show a marked tendency for ex-cessive diffusion. These time data were similar tothose of Watanabe et al.21 Therefore, technetium-99mcolloidal rhenium sulfide was very suitable for thepurposes of the current study.

Preoperative lymphoscintigraphy has been shownto be inaccurate in determining the difference be-tween the strong radioactivity of the injection site andthe sentinel lymph node adjacent to the injectionsite.39 However, preoperative lymphoscintigraphy isimportant because it defines sentinel lymph nodes atsites distant from the primary tumor. It also can detectlymph node skip metastases, which occur in greaterthan 30% of patients with lymph node metastasis,particularly patients with esophageal carcinoma.40 Inthe current study, we investigated the sentinel lymphnodes identified by preoperative lymphoscintigraphybecause new sentinel lymph nodes are detected rarelyintraoperatively by a hand-held probe, and the accu-racy of this latter approach depends on several tech-nical issues.9 Although preoperative lymphoscintigra-phy provides information that is useful duringsubsequent surgery, intraoperative evaluation with ahand-held probe is essential for detecting the extent ofsentinel lymph node distribution.

In the current study, one patient in whom nosentinel lymph node was detected had occult metas-tases in a paraesophageal lymph node. This lymphnode adjacent to the primary tumor was not detectedbecause the lymph nodes in this region were adjacentto the radiocolloid injection site, the so-called ‘shine-through’ phenomenon. If the paraesophageal lymphnodes are considered to be sentinel lymph nodes, the

accuracy of sentinel lymph nodes might be increasedif they are subjected to a focused evaluation. In 9 ofthe 23 patients in whom sentinel lymph nodes weredetected, positive lymph nodes adjacent to the radio-colloid injection site were detected by routine evalu-ation and CK immunostaining. Among the remaining14 patients, negative sentinel lymph nodes were de-tected in 5 by both HE and CK staining, positive sen-tinel lymph nodes were detected in 4 by both HE andCK staining, and positive sentinel and nonsentinel andnegative lymph nodes adjacent to the primary tumorwere detected in 5 by both HE and CK staining. Tenpatients with T1 lesions who had sentinel lymphnodes detected were included in this group of 14 pa-tients, suggesting that this technique is better suited topatients with early-stage tumors. In addition, a fo-cused analysis of both sentinel and paraesophageallymph nodes using CK staining will increase the accu-racy of the sentinel lymph node concept. It may beimportant to investigate paraesophageal lymph nodesadjacent to the radiocolloid injection site, the so-called lymph node inside shine-through, in the sameway as with sentinel lymph nodes.

Wood et al.33 reported a focused evaluation of sen-tinel lymph nodes in patients with colorectal carcinoma.They detected occult lymph node micrometastases thatupstaged approximately 20% of their cases. In the cur-rent study, an additional 3 cases (12%) were upstagedafter a focused evaluation of sentinel lymph nodes.Bilchik et al.41 reported that CK staining and multima-rker reverse transcriptase–polymerase chain reaction as-says of sentinel lymph nodes identified patients at highrisk of recurrence of colorectal carcinoma. However, fo-cused evaluations based on multisectioning and immu-nostaining are too costly and time-consuming for rou-tine evaluation of nonsentinel lymph nodes.

In conclusion, the findings of the current studyare consistent with the hypothesis that sentinel lymphnode mapping can be used to identify the lymphaticbasin and is a feasible technique in patients withesophageal SCC. An analysis of sentinel lymph nodesbased on CK immunohistochemistry increases the ac-curacy of sentinel lymph node staging. The results ofthis technique for sentinel lymph nodes warrant fur-ther investigation. However, in the future, the tech-nique of sentinel lymph node mapping will be usefulfor planning the treatment of esophageal SCC.

REFERENCES1. Albertini JJ, Lyman GH, Cox C, et al. Lymphatic mapping

and sentinel node biopsy in the patient with breast cancer.JAMA. 1996;276:1818 –1822.

2. Stewart KC, Lyster DM. Interstitial lymphoscintigraphy forlymphatic mapping in surgical practice and research. J In-vest Surg. 1997;10:249 –262.


3. Morton DL, Wen DR, Wong JH, et al. Technical details ofintraoperative lymphatic mapping for early stage mela-noma. Arch Surg. 1992;127:392–399.

4. Krag DN, Weaver DL, Alex JC, et al. Surgical resection andradiolocalization of the sentinel lymph node in breast can-cer using a gamma probe. Surg Oncol. 1993;2:335–339.

5. Wells KE, Rapaport DP, Cruse CW, et al. Sentinel lymphnode biopsy in melanoma of the head and neck. Plast Re-constr Surg. 1997;100:591–594.

6. Kelemen PR, Van Herle AJ, Giuliano AE. Sentinel lymphad-enectomy in thyroid malignant neoplasms. Arch Surg. 1998;133:288 –292.

7. Giuliano AE, Kirgan DM, Guenther JM, et al. Lymphaticmapping and sentinel lymphadenectomy for breast cancer.Ann Surg. 1994;220:391– 401.

8. Giuliano AE, Dale PS, Turner RR, et al. Improved axillarystaging of breast cancer with sentinel lymphadenectomy.Ann Surg. 1995;222:394 – 401.

9. Giuliano AE, Jones RC, Brennan M, et al. Sentinel lymphad-enectomy in breast cancer. J Clin Oncol. 1997;15:2345–2350.

10. Noguchi M, Tsugawa K, Kawahara F, et al. Dye-guided sen-tinel lymphadenectomy in clinically node-negative andnode-positive breast cancer patients. Breast Cancer. 1998;5:381–387.

11. Flett MM, Going JJ, Stanton PD, et al. Sentinel node local-ization in patients with breast cancer. Br J Surg. 1998;85:991–993.

12. Guenther JM, Krishnamoorthy M, Tan LR. Sentinel lymph-adenectomy for breast cancer in a community managedcare setting. Cancer J Sci Am. 1997;3:336 –340.

13. Dale PS, Williams JT IV. Axillary staging utilizing selectivesentinel lymphadenectomy for patients with invasive breastcarcinoma. Am Surg. 1998;64:28 –32.

14. Koller M, Barsuk D, Zippel D, et al. Sentinel lymph nodeinvolvement—a predictor for axillary node status withbreast cancer— has the time come? Eur J Surg Oncol. 1998;24:166 –168.

15. Veronesi U, Paganelli G, Galimberti V, et al. Sentinel-nodebiopsy to avoid axillary dissection in breast cancer with clini-cally negative lymph-nodes. Lancet. 1997;349:1864–1867.

16. Pijpers R, Meijer S, Hoekstra OS, et al. Impact of lympho-scintigraphy on sentinel node identification with techne-tium-99m-colloidal albumin in breast cancer. J Nucl Med.1997;38:366 –368.

17. Borgstein PJ, Pijpers R, Comans EF, et al. Sentinel lymphnode biopsy in breast cancer: guidelines and pitfalls of lym-phoscintigraphy and gamma probe detection. J Am CollSurg. 1998;186:275–283.

18. Miner TJ, Shriver CD, Jaques DP, et al. Ultrasonographicallyguided injection improves localization of the radiolabeledsentinel lymph node in breast cancer. Ann Surg Oncol. 1998;5:315–321.

19. Crossin JA, Johnson AC, Stewart PB, et al. Gamma-probe-guided resection of the sentinel lymph node in breast can-cer. Am Surg. 1998;64:666 – 669.

20. Krag D, Weaver D, Ashikaga T, et al. The sentinel node inbreast cancer—a multicenter validation study. N Engl J Med.1998;339:941–946.

21. Watanabe T, Kimijima I, Ohtake T, et al. Sentinel nodebiopsy with technetium-99m colloidal rhenium sulphide inpatients with breast cancer. Br J Surg. 2001;88:704 –707.

22. Offodile R, Hoh C, Barsky SH, et al. Minimally invasivebreast carcinoma staging using lymphatic mapping withradiolabeled dextran. Cancer. 1998;82:1704 –1708.

23. Barnwell JM, Arredondo MA, Kollmorgen D, et al. Sentinelnode biopsy in breast cancer. Ann Surg Oncol. 1998;5:126–130.

24. O’Hea BJ, Hill AD, El-Shirbiny AM, et al. Sentinel lymphnode biopsy in breast cancer: initial experience at MemorialSloan-Kettering Cancer Center. J Am Coll Surg. 1998;186:423– 427.

25. Smillie T, Hayashi A, Rusnak C, et al. Evaluation of feasibilityand accuracy of sentinel node biopsy in early breast cancer.Am J Surg. 2001;181:427– 430.

26. Imoto S, Fukukita H, Murakami K, et al. Pilot study onsentinel node biopsy in breast cancer. J Surg Oncol. 2000;73:130 –133.

27. Cox CE, Pendas S, Cox JM, et al. Guidelines for sentinel nodebiopsy and lymphatic mapping of patients with breast can-cer. Ann Surg. 1998;227:645– 653.

28. Pendas S, Dauway E, Cox CE, et al. Sentinel node biopsy andcytokeratin staining for the accurate staging of 478 breastcancer patients. Am Surg. 1999;65:500 –506.

29. Branagan G, Hughes D, Jeffrey M, et al. Detection of micro-metastases in lymph nodes from patients with breast can-cer. Br J Surg. 2002;89:86 – 89.

30. Kitagawa Y, Ohgami M, Fujii H, et al. Laparoscopic detectionof sentinel lymph nodes in gastrointestinal cancer: a noveland minimally invasive approach. Ann Surg Oncol. 2001;8:86S– 89S.

31. Japanese Society for Esophageal Diseases. Guidelines for theclinical and pathological studies on carcinoma of theesophagus, 9th edition. Tokyo: Kanehara, 1999.

32. Beck JR. Likelihood ratios. Another enhancement of sensi-tivity and specificity. Arch Pathol Lab Med. 1986;110:685–686.

33. Wood TF, Tsioulias GJ, Morton DL, et al. Focused examina-tion of sentinel lymph nodes upstages early colorectal car-cinoma. Am Surg. 2000;66:998 –1003.

34. Tsioulias GJ, Wood TF, Morton DL, et al. Lymphatic map-ping and focused analysis of sentinel lymph nodes upstagegastrointestinal neoplasms. Arch Surg. 2000;135:926 –932.

35. Waters GS, Geisinger KR, Garske DD, et al. Sentinel lymphnode mapping for carcinoma of the colon: a pilot study. AmSurg. 2000;66:943–945.

36. Kitagawa Y, Fujii H, Mukai M, et al. The role of the sentinellymph node in gastrointestinal cancer. Surg Clin North Am.2000;80:1799 –1809.

37. Yasuda S, Shimada H, Chino O, et al. Sentinel lymph nodedetection with Tc-99m tin colloids in patients with esopha-gogastric cancer. Jpn J Clin Oncol. 2003;33:68 –72.

38. De Cicco C, Cremonesi M, Luini A, et al. Lymphoscintigra-phy and radioguided biopsy of the sentinel axillary node inbreast cancer. J Nucl Med. 1998;39:2080 –2084.

39. Molpus KL, Kelley MC, Johnson JE, et al. Sentinel lymphnode detection and microstaging in vulvar carcinoma. JReprod Med. 2001;46:863– 869.

40. Hosch SB, Stoecklein NH, Pichlmeier U, et al. Esophagealcancer: the mode of lymphatic tumor cell spread and itsprognostic significance. J Clin Oncol. 2001;19:1970 –1975.

41. Bilchik AJ, Saha S, Wiese D, et al. Molecular staging of earlycolon cancer on the basis of sentinel node analysis: a mul-ticenter Phase II trial. J Clin Oncol. 2001;19:1128 –1136.


Documents

Sentinel lymph nodes with technetium-99m colloidal rhenium sulfide in patients with esophageal carcinoma