Annals of Surgical Oncology

, Volume 12, Issue 6, pp 504–509

Sentinel Nodes Are Identifiable in Formalin-Fixed Specimens After Surgeon-Performed Ex Vivo Sentinel Lymph Node Mapping in Colorectal Cancer


  • Fraser McLean Smith
    • Departments of Academic Surgery and PathologyCork University Hospital
  • John Calvin Coffey
    • Departments of Academic Surgery and PathologyCork University Hospital
  • Nurul Mod Khasri
    • Departments of Academic Surgery and PathologyCork University Hospital
  • Miriam Fiona Walsh
    • Departments of Academic Surgery, Pathology, and Clinical and Molecular OncologySt James’s Hospital
  • Nollaig Parfrey
    • Departments of Academic Surgery and PathologyCork University Hospital
  • Eoin Gaffney
    • Departments of Academic Surgery, Pathology, and Clinical and Molecular OncologySt James’s Hospital
  • Richard Stephens
    • Departments of Academic Surgery, Pathology, and Clinical and Molecular OncologySt James’s Hospital
  • M. John Kennedy
    • Departments of Academic Surgery, Pathology, and Clinical and Molecular OncologySt James’s Hospital
  • William Kirwan
    • Departments of Academic Surgery and PathologyCork University Hospital
    • Departments of Academic Surgery and PathologyCork University Hospital

DOI: 10.1245/ASO.2005.08.019

Cite this article as:
Smith, F.M., Coffey, J.C., Khasri, N.M. et al. Ann Surg Oncol (2005) 12: 504. doi:10.1245/ASO.2005.08.019



In recent years, the technique of sentinel lymph node (SLN) mapping has been applied to colorectal cancer. One aim was to ultrastage patients who were deemed node negative by routine pathologic processing but who went on to develop systemic disease. Such a group may benefit from adjuvant chemotherapy.


With fully informed consent and ethical approval, 37 patients with primary colorectal cancer and 3 patients with large adenomas were prospectively mapped. Isosulfan blue dye (1 to 2 mL) was injected around tumors within 5 to 10 minutes of resection. After gentle massage to recreate in vivo lymph flow, specimens were placed directly into formalin. During routine pathologic analysis, all nodes were bivalved, and blue-staining nodes were noted. These later underwent multilevel step sectioning with hematoxylin and eosin and cytokeratin staining.


SLNs were found in 39 of 40 patients (98% sensitivity), with an average of 4.1 SLNs per patient (range, 1–8). In 14 of 16 (88% specificity) patients with nodal metastases on routine reporting, SLN status was in accordance. Focused examination of SLNs identified occult tumor deposits in 6 (29%) of 21 node-negative patients. No metastatic cells were found in SLNs draining the three adenomas.


The ability to identify SLNs after formalin fixation increases the ease and applicability of SLN mapping in colorectal cancer. Furthermore, the sensitivity and specificity of this simple ex vivo method for establishing regional lymph node status were directly comparable to those in previously published reports.


ColorectalSentinel nodeEx vivoMicrometastasisUltrastaging

After curative surgery for colorectal cancer, the single most important factor in determining a patient’s prognosis is the presence or absence of lymphatic metastases.1 Patients with evidence of nodal tumor spread are at increased risk of developing systemic disease. As a result, these patients go on to receive adjuvant chemotherapy, given its proven benefits.2 Patients who are lymph node negative are followed up by observation alone. However, 20% to 30% of patients initially deemed lymphnode negative develop systemic recurrence.3,4 This stark fact has led to the hypothesis that these patients may, in fact, have had lymph node metastases that were missed by routine pathologic methods. If this is the case, then this subpopulation of lymph node–negative patients may benefit from adjuvant chemotherapy.

Sentinel lymph node mapping (SLNM), now routinely used in the management of both breast cancer and melanoma,57 has recently been the focus of much attention in the setting of colorectal cancer. SLNM works by tracing the lymphatic drainage of a tumor by using either blue dye or radioisotope that ultimately reaches the first-draining lymph node or nodes (i.e., the sentinel lymph node [SLN]). Theoretically, the SLN is most likely to harbor metastatic disease, if this is present.8,9 Focused examination of the SLN by using sensitive methods can identify occult metastatic spread otherwise missed by routine, nonfocused examination. This fact has generated considerable interest in the use of SLNM in colorectal cancer, as reflected in increasing numbers of reports that aim to characterize this technique. Most of these have mapped tumors in vivo by using blue dye alone, radioisotope alone, or a combination of the two.4,1020

To date there have been three reports of ex vivo mapping in colorectal cancer.2123 In these reports, blue dye was injected peritumorally, similar to mappings performed in vivo. Gentle massage of specimens recreated lymphatic flow and enabled identification of the SLN. The sensitivity, specificity, and overall percentage of patients upstaged after ex vivo SLNM are similar to those after in vivo SLNM. One of the advantages of performing mappings ex vivo is that anaphylaxis is no longer a concern. With the in vivo technique, anaphylactic reactions can occur in up to 2% of patients.24 Further benefits of ex vivo mapping are that unnecessary tumor handling is avoided and operating time is unaffected.

The aim of this study was to determine the feasibility of identifying, harvesting, and performing a focused examination of SLNs from specimens that were mapped ex vivo before being placed in formalin. We found that, with this technique, SLNs were identifiable and readily differentiated from non-SLNs. The focused examination of identified SLNs identified metastatic disease and closely correlated with regional lymph node status. Moreover, a focused examination of SLNs harvested in this manner upstaged a significant proportion of patients deemed node negative by conventional evaluation.


Consent and Inclusion Criteria

This study was designed to be compliant with ICH-GCP guidelines. Ethical approval was obtained from the regional ethics committee, and before mappings were conducted, patients gave fully informed consent. Thirty-seven patients with primary colorectal carcinoma and three with large adenomas in the rectum or lower sigmoid colon were prospectively included. Patients who had received neoadjuvant radiochemotherapy for locally advanced rectal cancer were excluded.

Mapping Technique

Once the tumor was resected, a member of the surgical team performed the mapping within 10 minutes. Specimens were divided along their antimesenteric border with a scalpel, thereby exposing the tumor with approximately 4 cm of normal mucosa proximally and distally. Lymphazurin (isosulfan blue dye; Tyco, OH) 1 to 1.5 mL was injected with a tuberculin syringe quadrantically around the tumor. This was performed subserosally if the tumor was above the peritoneal reflection or submucosally if below it, because the bowel lacks serosa below the peritoneal reflection.23 Injection areas were gently massaged for 2 to 3 minutes to generate flow of dye along the lymphatics. The specimen was then placed directly into formalin.

A designated pathologist performed routine pathologic evaluation 24 to 72 hours later. During this time, all lymph nodes were identified and bivalved. Those found to be stained blue were designated SLNs. These were archived in separate coded cassettes for later processing.

Selecting Nodes for Focused Examination

Once a formal report for each specimen was available, nodes found to contain metastases by routine methods were identified. The archival codes of these nodes were compared with those of the SLNs. If SLNs were negative but non-SLN metastases were found or if the patient had node-negative disease on routine reporting, SLNs were subjected to focused examination.

Focused Examination

SLNs were step-sectioned in accordance with the protocol described by Weise et al.25 Briefly, nodes were cut at six levels 20 to 30 μm apart. Initially, levels 2 to 5 were hematoxylin and eosin (H&E) stained and then examined for evidence of tumor spread. If negative for metastatic disease, levels 1 and 6 underwent cytokeratin immunohistochemical staining (CK-IHC) with the monoclonal anti-CK antibody Cam 5.2 (Becton Dickinson, NJ). All immunostaining was performed with a Ventana Nexes I-view automatic staining machine (Ventana Medical Systems Inc., Tucson, AZ) with a 1:2 dilution of primary antibody. Sections of colorectal cancer were used as a positive control.

Sections were examined for occult tumor spread in a blinded fashion by two pathologists. Findings were then reviewed by a senior pathologist.

Statistical Analysis

A power calculation was performed in conjunction with the Department of Statistics, University College Cork, to determine the number of routinely node-negative patients that would be required to determine whether patients who were upstaged were upstaged as a result of selective flow of dye to true SLNs rather than to random nodes in the bed. The number of node-negative patients needed was calculated on the basis that 30% would be upstaged with a minimum of 80% power. This calculation indicated that 45 patients with node-negative disease would need to be recruited to prove this association with statistical significance. The number of node-negative patients needed increased considerably when the percentage of anticipated upstaging decreased.26


Mappings were performed on 23 men and 17 women. A total of 592 lymph nodes were found in 40 patients, with a total SLN yield of 121 nodes. In total, 29 colonic tumors, 8 rectal tumors, and 3 adenomas were mapped. The average number of lymph nodes found per specimen was 16.9 (range, 7–37). SLNs were found in 39 of 40 cases (98% sensitivity), with an average of 4.4 per patient (range, 1–8; Table 1).

Breakdown of operative approaches with associated total and sentinel lymph node yields

Operation performed

No. performed

Mean (range) nodal collection

Mean (range) sentinel node collection

Right hemicolectomy


17.9 (11–37)

5 (2–8)

Transverse colectomy




Total colectomy




Left/sigmoid colectomy


16.6 (12–37)

3.9 (2–8)

APR/anterior resection


14.3 (7–28)

3.5 (1–8)

APR, abdominoperineal resection.

A total of 16 (44%) of 37 patients with colorectal cancer who were mapped were found to have gross metastases in their lymphatic beds on routine pathologic reporting. Thirteen of these were in colonic tumors arising above the peritoneal reflection, and three were from tumors arising below it.

Of these 16 tumors, 10 had grossly positive SLNs. In two cases, the SLNs were the only nodes that contained tumor. SLNs from the remaining six patients were step-sectioned, and the middle four sections were H&E stained. Another two patients were found to have gross metastases in these sections that had been missed by routine sectioning. CK staining was then performed on SLNs of the four remaining patients, and one CK-positive micrometastasis and a cluster of CK-positive cells were found in two patients. Thus, in patients reported as node-positive by routine methods, SLNM identified 14 of 16 (Fig. 1). This gave our technique a sensitivity of 88%.
Fig. 1

Bar chart comparing sentinel node posi-tivity with routinely node-positive patients throughall stages of the step-sectioning sequence. H+E,hematoxylin and eosin; CK, cytokeratin; +ve, positive.

Having established the sensitivity and specificity of SLNM in detecting metastatic disease in node-positive patients, we performed step sectioning of the SLNs from node-negative patients (n = 24). Twenty-one of these patients had histologically confirmed colorectal cancer, and three had large adenomas. In a similar fashion to SLNs from node-positive patients, only H&E staining of the four middle step sections was initially performed. This technique identified a .1-mm micrometastasis in 1 (5%) of the 21 patients with colorectal cancer (Fig. 2A). CK-IHC was then performed on levels 1 and 6 from all remaining SLNs. This revealed occult tumor cell spread in a further 5 (24%) of the 21 patients with colorectal cancer. One of these patients had a T2 tumor. The remaining five had T3 tumors. Focused examination by CK-IHC thus upstaged another five patients with colorectal cancer previously deemed node negative. In one of these patients with a T3 tumor, only a single CK staining cell was found. In the other four, however, CK-staining cell clusters or micrometastases were found (Fig. 2B).
Fig. 2

(A) Hematoxylin and eosin photomicrograph (×400) of occult micrometastasis in the sentinel node of a routinely node-negative patient. (B) Cytokeratin immunohistochemistry photomicrograph (×400) of a representative cluster of occult tumor cells.

Of note, three patients with node-positive rectal cancer who had not undergone neoadjuvant radiochemotherapy were successfully mapped. In each of these patients at least one SLN was found to contain metastatic disease at the time of routine pathologic analysis. Although five patients with node-negative rectal tumors were also mapped, none of these was upstaged by focused examination of associated SLNs. Similarly, no evidence of occult tumor spread was found in SLNs draining adenomas.


SLNM is now widely practiced in selected cases of melanoma and breast cancer.5 In both breast cancer and melanoma, the tumor burden in the SLN accurately reflects the regional lymphatic status. The advantage of the SLNM technique is that it obviates unnecessary and potentially harmful lymph node dissection.27 Staging for colorectal adenocarcinoma, however, requires that all regional lymph nodes be harvested. Hence, the ability to avoid unnecessary lymph node dissection in colorectal cancer is not an issue. Nevertheless, the potential for the SLNM technique in colorectal cancer lies in its ability to upstage patients and, in doing so, to identify the subgroup that may ultimately develop metastatic disease. This process is referred to as ultrastaging.11 Such patients could benefit from adjuvant chemotherapy.

SLNM is an evolving field in colorectal cancer, and most reports published to date have involved in vivo mappings. Only three series have reported on ex vivo mapping.2123 Numerous drawbacks have hampered the implementation of both in vivo and ex vivo SLNM in colorectal cancer. When performed in vivo, increases in operating time by 10 to 15 minutes28 and a risk of anaphylaxis are encountered.24 Additionally, there is the theoretical risk of tumor cell shedding due to increased tissue manipulation.21 In the ex vivo mappings described, specimens were immediately transported for expedient pathologic processing;21,22 however, this is impractical if the procedure is not part of a dedicated research project.

Both the in vivo and ex vivo techniques described to date have implemented immediate SLN exploration and collection.10,19,21 This is time consuming. Moreover, it disrupts the specimen before routine evaluation by the pathologist.

In this study, ex vivo mappings took approximately 5 minutes to perform and did not add to the overall operating time. Once mapped, the specimen was placed in formalin until a pathologist was able to identify SLNs during routine diagnostic analysis. These factors led to enhanced applicability and acceptability of SLNM as part of routine staging in our institution.

After the decision to place the unaltered specimen directly into formalin, the question remained as to whether SLNs identified in this manner could accurately reflect the regional lymphatic status when evaluated for tumor deposits. We found an 88% correlation between SLN status and regional lymph node status.

Two false-negative SLN mappings were identified. Here, evidence of metastasis to SLNs after focused examination was absent in the presence of positive non-SLNs. In both cases, blue-stained SLNs were found within 2 cm of the tumor, whereas the first involved nodes were located 5 cm from the tumor. These are therefore likely to represent skip metastases.29 The mapping process failed—i.e., no SLN was identified—in one patient with a tumor of the ascending colon. No obvious reason existed for this, because the mapping technique used was identical to that of all other tumors.

In this study, six patients were upstaged from node negative, according to routine histological processing, to node positive after focused examination of their SLNs. In all except one of the cases upstaged, micrometastases were detected by CK-IHC. Hence, not only did the technique of ex vivo SLNM described herein closely correlate with regional lymph node status, but occult disease was also identified. Moreover, similar proportions of patients have been upstaged in previous in vivo and ex vivo reports.10,12,21 These findings led us to conclude that our technique is at least comparable to that of other institutions.

In conclusion, we have developed a novel adaptation of ex vivo SLNM in colorectal cancer. This involved placing unaltered specimens immediately into formalin after SLNM. This simple measure greatly facilitated the incorporation of colorectal SLNM and subsequent focused examination at our institution. We found that ex vivo SLNM in this manner was feasible, accurately correlated with regional lymph node status, and identified occult micrometastatic disease at a frequency comparable to that with other techniques. We envisage that this adaptation will greatly enhance the applicability and introduction of SLNM and focused histological examination of lymph nodes into routine ultrastaging protocols.


The authors thank Edel Galvin for her technical assistance in performing CK-IHC and Eleanor Carton for her assistance in providing surgical specimens.

Copyright information

© The Society of Surgical Oncology, Inc. 2005