In the present study, we investigated the correlation between the presence of occult neoplastic cells (ONCs) in lymph node sinuses and recurrence/metastasis of stage II/III gastric cancer in 164 patients who underwent radical curative resection. We calculated the five-year relapse-free survival rate (5Y-RFS) and five-year overall survival rate (5Y-OS) of the ONC(+) and ONC(-) groups. The 5Y-RFS was 71.4% in the ONC(-) group and 47.5% in the ONC(+) group (P=0.003). The 5Y-OS was 68.8 and 48.4%, respectively (P=0.008). ONCs were found in 34.8% of stage II patients and were also detected in 66.7% of stage III patients. For distinguishing between the recurrence and non-recurrence groups, the sensitivity of ONC(+) was 64.5% (40/62; P=0.003), the positive predictive value (PPV) was 49.4% (40/81), the specificity was 59.8% (61/102) and the negative predictive value (NPV) was 73.5% (61/83). This high sensitivity indicates that ONC positivity may be a significant indicator for high-risk patients in the early postoperative period, and a lack of ONCs may be a useful indicator for identifying low-risk patients, as patients without ONCs had a high NPV.

Here we evaluated the prevalence and prognostic impact of epithelial cell adhesion molecule (EpCAM)-positive disseminated tumor cells (DTCs) in stage I rectal cancer. Further we tested the association of these single tumor cells or small tumor cell groups with the extent of peritumoral lymphangiogenesis. A total of 845 regional lymph nodes (LN) of 44 patients classified as negative on conventional histopathology were retrospectively reanalyzed with immunohistochemistry (IHC) using the monoclonal antibody Ber-Ep4 directed against EpCAM for the detection of DTCs. The degree of lymphangiogenesis in the primary tumors was assessed by IHC of the primary tumor tissue using the monoclonal antibody D2-40, which reacts with the lymphatic endothelium. The IHC results were correlated with clinico-pathologic parameters and clinical follow-up data. EpCAM-positive DTCs in LNs were detected in 8 (18.2%) of the 44 patients. During a median follow-up of 59 months, 3 (37.5%) of the 8 patients with EpCAM-positive DTCs relapsed, whereas none of the DTC-negative patients developed tumor recurrence (P=0.004). Survival analysis revealed a significant effect of the prevalence of DTCs on overall survival (P=0.0009) and on recurrence-free survival (P=0.0001). Finally, the prevalence of EpCAM-positive DTCs in perirectal LNs was significantly correlated with a high density of peritumoral lymphatic vessels (P=0.015). Our results show that DTCs may occur in stage I of rectal cancer and are associated with poor prognosis. Their occurrence seems to be linked to a high density of newly formed lymphatic vessel at the primary tumor site. According to our data, patients with DTCs in their LN might benefit from adjuvant therapy.

Lymph node (LN) involvement is the most important prognostic factor in colorectal cancer (CRC), and pN-positive status identifies patients who require adjuvant chemotherapy. Approximately 15% to 20% of patients without nodal metastases (pN0) develop recurrent disease. In this study, we tested the prognostic significance of isolated tumor cells (ITCs) in LNs of patients with pN0 CRC (stages I and II).

ITCs in LNs regional to CRC were assessed in 312 consecutive patients with pN0 CRC who were followed up clinically and/or endoscopically for at least 6 months after surgery (mean, 67 months; median, 64 months; range, 8 to 102 months). LNs were dissected from gross surgical specimens according to a standardized protocol (with a mean of 17 LNs per patient; range, five to 107 LNs). In all, 5,313 pN0 LNs were collected and assessed by using cytokeratin immunostaining in two serial histology sections from each LN, which amounting to a total of 10,626 specimens. The correlation between ITC status and cancer recurrence was tested by using univariate and multivariate statistics.

ITCs were documented in 185 of 312 patients (59%). CRC relapsed in 31 of 312 patients (10%), and 25 of 31 recurrences (81%) were documented among ITC-positive patients. CRC recurrence rates among ITC-positive and ITC-negative patients were 14% (25 of 185 patients) and 4.7% (six of 127 patients), respectively. In both univariate and multivariate analyses, ITC status was the only variable significantly associated with cancer relapse (Cox model; hazard ratio, 3.00; 95% CI, 1.23 to 7.32; P = .013).

In patients with pN0 CRC, cancer relapse was significantly associated with ITCs in regional LNs. ITCs should be considered among the clinicobiologic variables that identify high-risk patients who can benefit from adjuvant chemotherapy.

The sentinel lymph node (SLN) technique has practical applications in multiple solid tumors including colorectal carcinoma. Identifying the SLN(s) provides better staging of the regional lymphatics beyond standard H&E analysis. This additional information assists in predicting biology and may be useful in guiding adjuvant therapy. We postulate the era of sentinel node has ushered in a new generation of node-negative patients; patients that have an exceptionally favorable outcome when compared to historic controls.

The prognostic relevance of lymphatic micrometastases in colorectal carcinoma is unclear. To determine the prognostic significance of micrometastases in colorectal cancer, a meta-analysis was performed on all studies, which reported 3-year disease-free survival (DFS) and overall survival (OS).

Published studies selected for meta-analysis contained sufficient data from which to extrapolate estimates of 3-year DFS and/or OS. From 1991-2003, 25 studies re-examined N0 lymph nodes by serial sectioning and immunohistochemical (IHC) staining or reverse transcriptase-polymerase chain reaction (RT-PCR) assay. Eight studies (566 patients) with IHC detected micrometastases and three (173 patients) with RT-PCR micrometastases were used to determine DFS and OS. Weighted estimates of 3-year survival were combined across studies within each group, and the combined survival estimates were compared across groups using a binomial test.

Micrometastases were identified in all IHC studies; upstaging, including N1, N1mi and N0(i+), was achieved in 32% (179/566 patients). All RT-PCR studies identified micrometastases; upstaging to N0(mol+) was achieved in 37% (64/173 patients). There was a statistically significant difference in 3-year OS between RT-PCR positive N0(mol+) patients (77.8%) and those for whom micrometastases were not detected (96.6%) (P < .001).

The prognostic value of micrometastases detected retrospectively by RT-PCR is significant in AJCC stage II colorectal patients. Studies utilizing RT-PCR performed a more complete nodal analysis when compared to studies using IHC techniques. RT-PCR may also be more specific for the detection of clinically relevant micrometastases compared to IHC detected cytokeratins. Prospective studies are needed to evaluate the potential benefit of systemic chemotherapy in patients with molecular metastases.

Approximately 30% of patients with lymph node (LN)-negative colorectal carcinoma (CRC) die of tumor recurrence, which can be related to the presence of tumor cells in LNs not detected by conventional histopathologic analysis. However, the prognostic significance of occult cancer cells still remains uncertain. We evaluated the incidence and the prognostic significance of occult cancer cells in LNs from 395 consecutive patients with curatively resected stage IIA CRC using immunohistochemistry for cytokeratin 20. Immunostained tumor cells were categorized as micrometastases (MCMs) or isolated tumor cells (ITCs) according to the American Joint Committee on Cancer criteria. The detection rates were compared with the clinicopathologic characteristics of the patients and with cancer-specific survival. The median follow-up time was 128 months. Micrometastases were detected in 39 patients (9.9%), whereas ITCs were found in 112 (28.4%), for an overall frequency of 38.2%. None of the clinicopathologic parameters examined was correlated with the presence of occult cancer cells. Patients with ITCs and those with negative LNs showed a similar survival rate (77.7% and 78.3%, respectively), whereas patients with MCMs had a lower survival rate (64.1%). At the univariate analysis, MCMs, tumor growth pattern, extent of tumor spread, and Crohn's-like lymphoid reaction influenced the survival rate significantly. Nevertheless, at the multivariate analysis, only the pattern of tumor growth and the extent of tumor spread were independent prognostic factors. The detection of immunostained tumor cells in the LNs of patients with stage IIA CRC occurs relatively frequently but has no significant effect on prognosis.

The prevalence of isolated tumour cells (ITCs) in regional lymph nodes from colorectal cancer (CRC) is controversial and has never been prospectively assessed in large groups of consecutive patients. pN0 early-relapsing CRC can be explained by lymph node-ITC.

To assess the prevalence of ITCs in regional lymph nodes from 309 consecutive patients with pN0M0 (pathological (p)-tumour-node-metastasis (TNM) staging system) CRCs.

ITCs were assessed by immunohistochemistry (MNF116 monoclonal antibody (1:100); Dako, Glostrup, Denmark) in two serial histological sections obtained from 5016 mesenteric lymph nodes from 309 patients with pN0 CRCs (mean number of lymph nodes per patient = 16.2; p-TNM stage 0, n = 25; p-TNM stage I, n = 123; and p-TNM stage II (A+B), n = 161). Tumour histology, vascular cancer invasion and pathological stage were also recorded.

ITCs were detected in the regional lymph nodes of 156 of 309 (50.5%) patients with CRC, mostly in nodes located within 3 cm from the neoplasia. ITC status correlated with (a) tumour p-TNM stage (Pearson's chi(2): p<0; ordered logistic regression: odds ratio (OR) = 4.6; 95% confidence interval (CI) = 2.88 to 7.33; p<0) and (b) pT value (Pearson's chi(2): p = 0; ordered logistic regression: OR = 4.9; 95% CI = 3.1 to 7.7; p<0). By multivariate analysis, including p-TNM stage, vascular invasion and ITC status, both stage (OR = 5.1; 95% CI = 2.9 to 8.9; p<0) and vascular invasion (OR = 4.2; 95% CI = 1.94 to 8.98; p<0) were found to be independent variables associated with ITC+ lymph nodes.

More than 50% of pN0-CRC patients have ITCs in the mesenteric lymph nodes. ITC status is significantly correlated with cancer stage and vascular cancer invasion. The clinicopathological effect of ITC remains to be prospectively evaluated.

The objective was to determine the presence and frequency of micrometastasis in lymph nodes of patients with rectal cancer treated by preoperative chemoradiation followed by curative resection.

All 56 patients included were treated with 5-FU and leucovorin plus 5,040 cGy, followed by radical surgery and were diagnosed with stage II distal rectal adenocarcinoma after complete pathological examination (ypT3-4N0M0). Immunohistochemistry was assessed with cytokeratin monoclonal antibody AE1/AE3. Three 4-microm paraffin sections were obtained from each lymph node, cut at 50 microm apart from each other. The results were reviewed by two independent pathologists.

Mean number of lymph nodes was 9.6 per patient. Four patients (7%) and seven lymph nodes (1.35%) were positive for micrometastasis. Three patients had pT3 and one a pT4 tumor. One of the patients had positive micrometastasis and the presence of mucinous deposits. One other patient had mucinous deposits without any micrometastasis. All four patients are alive with no evidence of recurrent disease. Fourteen patients negative for micrometastasis had recurrent disease (25%), eight systemic (14.7%) and six locoregional (10.3%). There were two cancer-related deaths. The mean follow-up period was 39 months.

Patients with rectal cancer treated by preoperative chemoradiation showed a surprisingly low rate of micrometastasis detection (7%), even in high-risk patients (T3 and T4 tumors). Lymph node micrometastasis was not associated with decreased overall or disease-free survival. The identification of mucinous deposits on lymph nodes with no viable tumor cells may be direct evidence of lymph node downstaging. The downstaging effect of preoperative chemoradiation therapy may be significant in reducing even micrometastasis detection in low rectal cancer managed by this treatment strategy.

Accurate staging in colorectal cancer is important to predict prognosis and identify patients who could benefit from adjuvant therapy. Patients with lymphatic metastasis, Stage III/Dukes C, are generally treated with adjuvant chemotherapy. Still, patients without lymphatic metastasis do have relapse as high as 27 percent in five years in Dukes B2. It is hypothesized that these patients have occult (micro)metastasis in their lymph nodes. If these (micro)metastasis can be identified, these patients might benefit from adjuvant therapy. We reviewed the literature on procedures to improve lymph node staging.

An extensive literature search was performed in PubMed (www.pubmed.com). Using the reference lists, more articles were found.

We found 30 articles about sentinel node in colorectal cancer describing original series. Some groups reported several studies including the same patients. We reported their largest studies. For all other techniques, we only included key articles.

Many techniques to improve staging have been described. The finding of occult (micro)metastasis is of prognostic significance in most studies. The sentinel node technique has been recently described for use in colorectal cancer. Although it seems clear that this technique has prognostic potential, it is not yet been shown in a follow-up study. Furthermore, the finding of occult (micro)metastasis in any technique used has not been shown to be clinically significant. Whether to treat patients with adjuvant therapy if occult (micro)metastasis are found needs to be proven in future studies.

Standard practice is to take one section from every lymph node found in colorectal carcinoma resection specimens, to look for metastatic carcinoma. This study evaluates whether assessing three sections separated by 100 microm detects significantly more metastases in nodes than the conventional single section.

A retrospective study of 100 colorectal carcinoma resection specimens. All blocks containing lymph nodes had two extra histological sections cut (separated by 100 microm) and stained with haematoxylin and eosin. The original slide was called level 1, and the extra two sections levels 2 and 3.

Twenty Dukes's A (equivalent to WHO-UICC stage grouping I, pTNM stage pT1/2N0), 43 Dukes's B (equivalent to WHO-UICC stage grouping II, pTNM stage pT3/4N0), and 37 Dukes's C (equivalent to WHO-UICC stage grouping III, pTNM stage at least pN1) cases were examined (total 1453 nodes). Twelve extra metastases (in 11 patients) were discovered in nodes at levels 2 and 3, which were negative in level 1. Ten cases were Dukes's C and, in one patient, this led to upstaging from N1 to N2 (pTNM classification system). One case was Dukes's B and the discovery of a single metastasis on level 2 upstaged it to Dukes's C.

Triple levelling detected more tumour deposits than the conventional single section. In two patients, the staging classification of the lesion was changed, with potentially important implications for prognosis and management.

The value of immunohistochemical detection of disseminated tumor cells in histopathologically tumor-free lymph nodes (pN0) of patients with colorectal cancer is still of uncertain prognostic value. We therefore evaluated the immunohistochemical detection rates and their prognostic value comparing three different monoclonal antibodies.

A total of 170 lymph nodes of 85 patients with curatively resected colorectal carcinoma at UICC stage I or II were evaluated for disseminated tumor cells. Frozen sections of each lymph node were immunohistochemically stained using three antibodies directed against CEA, CK20, and Ber-EP4. The detection rates were compared with histopathological tumor parameters and with the patient's survival. The median follow-up time was 86 months.

CEA-, CK20-, and Ber-EP4-positive disseminated tumor cells were identified microscopically in lymph nodes of 23 patients (27%), 24 patients (28%), and 23 patients (27%), respectively. In 18 patients (21%) disseminated tumor cells were found in consecutive sections and stained positive for all three monoclonal antibodies. The lymph nodes of 10 of 18 patients (56%), which developed tumor recurrence, contained CEA- and CK20-positive disseminated tumor cells. Ber-EP4-positive cells were present in lymph nodes of 9 of 18 patients (50%) with tumor recurrence. The 5-year overall survival of the 23 patients with CEA-positive disseminated tumor cells was 72% compared to 91% of the patients without immunohistochemical evidence of tumor cells (p<0.01). While the identification of CK20-positive tumor cells was also correlated significantly with a worse overall patient survival (p<0.01), the application of Ber-EP4 failed to reach significance (p=0.057). Multivariate analysis identified the tumor site (colon versus rectal cancer) (p<0.006) and the presence of CEA-positive disseminated tumor cells (p<0.03) as independent prognostic factors.

In colorectal carcinoma, the immunohistochemical detection of disseminated tumor cells in histopathologically pN0 peritumoral lymph nodes allows the identification of a subgroup with a significantly worse prognosis. Nevertheless, the prognostic value of immunohistochemically detected disseminated tumor cells remains controversial due to the nonuniform data in the literature.

Most patients with stage I and stage II colon adenocarcinomas do not have disseminated disease, and the group is not offered adjuvant therapy. However, more than 30% of stage II colon adenocarcinoma patients get metastases to remote organs. Thus, it is important to identify patients in this group at risk of disease relapse.

We have examined the prognostic value of isolated tumour cells (ITC) in mesenteric lymph nodes in a consecutive series of 156 colon carcinoma patients with stage II disease. Immunohistochemistry, using antibodies to cytokeratins, and morphology were used to identify presence of ITC.

ITC were detected in 59 (37.8%) patients. Presence of ITC in mesenteric lymph nodes was independently associated with reduced relative survival both in univariate (p=0.0199) and in a multivariate analysis (p=0.041).

The results strongly suggest that presence of ITC in mesenteric lymph nodes is associated with reduced relative survival in colon carcinoma patients stage II, and that detection of ITC may be important in treatment of these patients.

Inconsistent conclusions have been drawn about the clinical significance of micrometastases in lymph nodes (LNs) of node-negative colorectal cancer (CRC) patients. We performed a comparative study of detection of micrometastases using immunohistochemistry (IHC) by anti-cytokeratin antibody and carcinoembryonic antigen (CEA)-specific reverse-transcriptase polymerase chain reaction (RT-PCR) in the same patients, in an attempt to move closer to their clinical application.

Sixty-four CRC patients, with RNA of good quality available from paraffin-embedded LN specimens, were selected from 84 stage II patients who underwent curative surgery between 1988 and 1996. We investigated associations between the presence of micrometastases by each method and prognosis.

Micrometastases were detected in 19 (29.6%) of 64 patients by RT-PCR and in 35 (54.7%) of 64 patients by IHC. By RT-PCR analysis, patients exhibiting a positive band for CEA mRNA had a significantly worse prognosis than those who were RT-PCR-negative, with respect to both disease-free and overall survival (P =.027 and.015, respectively). By IHC analysis, the presence of micrometastasis did not predict patient outcome in terms of either disease-free or overall survival. Infiltrating pattern of tumor growth characteristic was significantly associated with shorter disease-free survival among various clinical or pathologic factors. By multivariate Cox regression analysis, micrometastasis detected by RT-PCR and the Crohn's-like lymphoid reaction were both independent prognostic factors.

Micrometastases detected by RT-PCR, but not IHC, may be of clinical value in identifying patients who may be at high risk for recurrence of CRC and who are therefore likely to benefit from systemic adjuvant therapy.

Sentinel lymph node (SLN) mapping techniques have been validated in breast cancer and melanoma. This study summarizes our experience with SLN mapping for colon cancer.

Fifty-five patients with colon cancer underwent intraoperative SLN mapping. One mL of 1% isosulfan blue was injected subserosally around the tumor. The first nodes highlighted with blue were identified as the SLNs. SLNs underwent multiple sectioning and immunohistochemical staining for cytokeratin. The overall learning curve was calculated.

Lymphatic mapping adequately identified at least 1 SLN in 45 patients (82%). SLNs adequately predicted regional status in 44 of 45 (98%) cases. In 9 of 45 cases (20%), the SLNs were the only sites of metastases. Among the 14 cases that were SLN positive, 6 of 55 patients (11%) were positive only by immunohistochemistry. Of the 31 cases with negative SLNs, 1 case had a 3.5-mm pericolonic tumor-replaced non-SLN (3% false-negative rate). The overall learning curve stabilized after five cases.

Intraoperative SLN mapping is a feasible technique, with a quick learning curve, and had a reasonable SLN identification rate. Negative SLNs accurately predict the status of non-SLNs 97% of the time. Eleven percent of patients were upstaged by demonstration of micrometastases and may benefit from adjuvant chemotherapy.

A variety of techniques have been employed for the detection of occult tumour cells in the blood, bone marrow and lymph nodes of patients with colorectal cancer. This review examines the methods used, results obtained and the clinical significance of studies in this field.

A Medline literature search was performed using the terms colorectal cancer, minimal residual disease, micrometastasis, polymerase chain reaction, reverse transcriptase polymerase chain reaction and immunocytochemistry; further references were obtained from key articles.

Immunocytochemical examination of bone marrow is the benchmark for detecting clinically significant occult disease. Larger standardized studies are required to confirm the prognostic significance of molecular assays for the detection of tumour cells in blood and bone marrow. The prognostic significance of lymph node tumour cells detected by either immunohistochemical or molecular methods awaits further affirmation.

Standardization of terminology and techniques used, combined with large prospective clinical studies, is required if detection of occult residual disease is to become a prognostic marker for recurrence in colorectal cancer.

The detection of lymph node metastases is the single most important prognostic factor for patients with colorectal cancer. This review outlines the difficulties and methods of detecting positive lymph node metastases in this disease. An outline of traditional diagnostic methods including preoperative ultrasound and cross sectional imaging techniques are evaluated alongside newer modalities including immunoscintography and PET scanning and intraoperative radioguided imaging. Pathological methods of detecting positive nodal disease using standard histopathological staging, enhanced lymph node harvesting and determination of micrometastases are also discussed.

Lymph node status is a major determinant of disease recurrence after patients undergo curative resection for gastric carcinoma. A proportion of patients without lymph node metastasis develop systemic recurrences. Recent studies in a range of solid tumors have found a high incidence of micrometastases in the regional lymph nodes of patients with apparently negative lymph nodes. In patients with breast and colorectal carcinoma, the presence of micrometastases has been associated with a poorer prognosis. In patients with gastric carcinoma, the significance of micrometastases in lymph nodes remains controversial. Most published reports on this subject suffer from the problems of small sample size and selection bias.

One hundred seven patients with pathologic T2N0M0 (tumor invades muscularis propria or subserosa [T2], no regional lymph node metastasis [N0], and no distant metastasis [M0]; pT2N0M0) gastric carcinoma who underwent gastric resection between 1984 and 1990 at the National Cancer Center Hospital were studied. Two consecutive sections were newly prepared from each lymph node for hematoxylin and eosin staining and immunohistochemical staining (IHC) with antibody against cytokeratin. Associations between clinicopathologic factors and the presence of micrometastases as well as micrometastases and survival were sought.

Micrometastases were identified in 38 of 107 patients (35.5%) and in 87 of 4484 lymph nodes (1.94%) by IHC. The incidence of micrometastases was significantly higher in patients with infiltrative tumors than in patients with expansive, growing tumors (P = 0.02). Other clinicopathologic findings had no statistically significant correlation with the incidence of micrometastases. The 5-year survival rates of patients with and without micrometastases were 94% and 89%, respectively. Similarly, the 10-year survival rates were 79% and 74%, respectively. The survival curves of patients with or without micrometastasis were nearly superimposed (P = 0.86).

The presence of immunohistochemically detected micrometastases in the regional lymph nodes did not affect the survival of Japanese patients with pT2N0M0 gastric carcinoma who had undergone gastrectomy with D2 lymph node dissection.

The sentinel lymph node (SLN) mapping technique has been used in breast cancer and melanoma, and was recently described for colon cancer.

Thirty-five patients with colon cancer underwent intraoperative SLN mapping. One milliliter of 1% isosulfan blue was injected subserosally around the tumor. The first nodal area that was highlighted with blue was identified as the SLN. All lymph nodes underwent examination with hematoxylin and eosin (H&E) stain. SLNs underwent additional sectioning and were stained with CAM 5.2.

Lymphatic mapping adequately identified the SLN in 25 patients (71%). In the 15 cases where the SLNs were negative for metastases, all other non-SLNs were also negative (0% false negative rate). The SLN was the only site of metastases in 6 (17%) of 35 patients. CAM 5.2 staining provided the only evidence of micrometastases in 4 (11%) of 35 patients.

Intraoperative SLN mapping is a feasible technique with a reasonable SLN identification rate (71%). The absence of metastases in the SLNs accurately predicts the status of the non-SLNs. Tumors in 11% of patients were upstaged by the demonstration of micrometastatic involvement, and these patients may benefit from further adjuvant chemotherapy.

The number of examined lymph nodes and metastases in lymph nodes smaller than 5 mm (small lymph nodes) are a determining factor in the stage of rectal cancer although the clinical significance of occult micrometastases is controversial. We are reporting our preliminary results on the identification and prognostic utility of metastases in small lymph nodes and occult micrometastases.

We searched small metastatic lymph nodes in 101 cases of adenocarcinoma of the lower third of the rectum. We used the manual technique to dissect mesorectal fat and occult micrometastases in the lymph nodes of 52 Dukes' A and B patients, using a pool of anticytokeratin antibodies.

Forty-five percent of the metastatic lymph nodes were smaller than 5 mm in diameter and determined the Dukes' stage in 15 (30.6%) of 49 Dukes' C patients. Occult micrometastases were found in 21 (40.4%) patients: five recurred but vascular invasion, positive distal margin of the rectum, and positive circumferential margin of the mesorectum were present.

Small metastatic lymph nodes, vascular invasion, positive distal margin of the rectum, and positive circumferential margin of the mesorectum were found to be more important than occult micrometastases in predicting early recurrence of rectal cancer.

Colon cancer has been assumed to spread sequentially through the regional lymphatic bed, with skip metastases occurring in only 1% to 3% of cases. Molecular techniques allow the detection of occult metastases, but to date have not been applied to assess the pattern of regional lymphatic spread of colon cancer.

Fifty-five tumors from 54 patients with colonic adenocarcinoma were studied. Lymph node mapping was performed on fresh colonic specimens recording the position of each node on an anatomical diagram. Half of each lymph node was submitted for routine histology examination and half assayed for keratin 20 gene expression by reverse transcription-polymerase chain reaction. Logistic regression was used to analyze the distribution of histologic and occult metastases.

A total of 1084 lymph nodes were dissected (median, 19 nodes; range, 4-52). Sixty-four lymph nodes from 20 tumors had histologically evident metastases and 76 lymph nodes from 13 tumors had occult metastases. There was no difference in the distribution of either histologic or occult metastases among paracolic, intermediate, and apical node groups. Ten patients had evidence of anatomical skip lesions after lymph node mapping and molecular analysis, only 1 of which was histologically detectable.

This study demonstrates a higher incidence of skip metastases in colon cancer assessed by molecular techniques than has previously been reported, challenging the concept of sequential development of early lymph node metastases.

In the absence of other metastatic disease, the presence of lymph node metastasis remains the most important determinant of survival in colorectal cancer (CRC). Cluster designation 44 variant 6 (CD44v6) over-expression is associated with worse outcome in all stages of CRC. The CD44v6 is believed to confer metastatic potential through its facilitation of migration, extravasation and proliferation, although the specific means by which it conveys an adverse prognosis in CRC is unknown. The aim of the present study was to determine if CD44v6 over-expression in Stage II CRC subjects was associated with the presence of lymph node micrometastases.

We assessed tumour CD44v6 expression in 43 randomly sampled subjects who had resections for Stage II CRC between 1984 and 1991 by using immunohistochemistry. Micrometastases were sought in corresponding lymph node (LN) sections using keratin immunohistochemistry.

There was a statistical trend between tumour CD44v6 over-expression and mortality (P = 0.09) and a significant relationship between LN cytokeratins and mortality (P = 0.01). There was no association between the detection of LN cytokeratins and tumour CD44v6 over-expression.

We conclude that the adverse survival effect of CD44v6 over-expression is not mediated though lymphatic spread and postulate that it may therefore facilitate haematogenous metastasis.

In a previous study the authors demonstrated, using immunohistochemical methods for epithelial antigens, that the regional lymph nodes of gastric adenocarcinoma contained individual tumor cells or small clusters of these cells (tumor cell microinvolvement [TCM]) in over 90% of cases. In the current study the authors used the same method to investigate a series of gastric adenocarcinoma cases treated with neoadjuvant chemotherapy prior to tumor resection; their aim was to determine the effect of chemotherapy on TCM in regional lymph nodes.

Resection specimens from 17 patients with adenocarcinoma of the stomach, resected after neoadjuvant treatment and classified by routine histology as ypN0, were included in this study. One section from each of the 622 lymph nodes dissected from these specimens was stained by immunohistochemical methods for cytokeratins and Ber-Ep4.

Six patients (35%) and 25 of the 622 lymph nodes (4.0%) had TCM, compared with 93% of patients and 21.8% of lymph nodes in the previous study of patients treated with surgery alone. The lymph node response to chemotherapy correlated with the pathologic response of the primary tumor. Specifically, none of 5 patients with a complete or major pathologic response versus 6 of 12 (50%) patients with minor, partial, or no response had lymph node microinvolvement.

In comparison to our previous study, this study indicates that chemotherapy has a marked effect on tumor cells in regional lymph nodes and that the extent of this effect can be correlated with the degree of pathologic response of the primary tumor to chemotherapy.

Recent reports suggest that genetic examination of K-ras or p53 mutation is more sensitive for the detection of occult lymph node metastasis in colorectal carcinomas than conventional examination by haematoxylin and eosin (H & E) staining or immunohistochemistry for gene products. The aim of this study was, first, to define the microscopic characteristics of metastatic cancer cells in lymph nodes stained by the anti-cytokeratin antibody CAM5.2 for cytokeratins 8 and 18, and, second, to compare the detection rate of occult lymph node metastasis for immunohistochemical vs genetic methods.

K-ras mutations were first examined in primary tumours of seven cases which showed distant metastasis or local recurrence within 5 years of the initial surgery in spite of the original reporting of no lymph node metastasis by routine H & E staining. K-ras mutations were positive in three cases in primary tumours and lymph nodes, and the remaining four primary tumours were negative for p53 mutation as well as K-ras mutation. Therefore, genetic analysis of occult lymph node metastasis was uninformative, but occult metastasis was detected by cytokeratin staining in two of these four cases. Comparative study of cytokeratin-positive cells was performed on each of the 43 lymph nodes from three cases with K-ras mutations. Cancer cells were detected in 28 of the 43 lymph nodes (65.1%) by cytokeratin staining and in 10 of the 43 corresponding lymph nodes (23.3%) by genetic analysis. Artefactual contamination by cancer cells was present in eight of the 28 cytokeratin positive lymph nodes, and three of the eight nodes were genetically positive.

This study suggests that cytokeratin immunohistochemistry is more sensitive and specific for the detection of occult lymph node metastasis than genetic diagnosis by K-ras mutation in cases with genetic alterations as well as in cases without them.

There is controversy about the prognostic significance of occult lymph node metastases detected by immunohistochemistry with the anti-cytokeratin antibody CAM 5.2. The aim of this study was to characterize occult lymph node metastases in colorectal carcinomas that might be associated with a higher risk of recurrence. Three hundred fifty-eight lymph nodes from 10 recurrent and 9 nonrecurrent cases of colorectal carcinoma were examined. All these patients had been reported originally as having no lymph node metastases by routine hematoxylin and eosin staining. Three 10-micron sections or ten 3-micron sections (30-micron total thickness) from each lymph node were stained with CAM 5.2 and examined for the presence of occult lymph node metastases. Occult metastases were detected in 67 of 175 lymph nodes from the recurrent cases, and in 23 of 183 lymph nodes from the nonrecurrent cases. The frequency of positive nodes was significantly higher in the recurrent cases. The recurrent cases had metastases in nodes more distant from the main tumor than did the nonrecurrent cases. Detection of occult lymph node metastases with cytokeratin immunohistochemistry may make it possible to identify patients with a higher risk of recurrence after the removal of a primary colorectal tumor.