Background
Curatively intended surgical resection is the standard therapy for operable patients with early-stage non-small cell lung cancer (NSCLC), and the prognosis of these patients is closely related to disease stage [
1]. The regional lymph node involvement is a major prognostic factor, and for complete surgical resection of NSCLC a systematic nodal dissection is recommended [
2]. This allows pathological staging of the disease according to standardized definitions, and thereby decision of further treatment strategies. The fact that approximately half of the patients undergoing surgery experience disease relapse, suggests that disseminated tumor cells (DTCs) may be present already at the time of surgery [
3]. In routine clinical practice, pathological evaluation of resected lymph nodes is done by standard histopathology, a method by which DTCs cannot be identified. The high recurrence rate after surgical resection of NSCLC indicates that current staging classifications are not able to accurately predict patient outcome and that the nodal staging might be suboptimal. Detection of DTCs to regional lymph nodes at the time of surgery could possibly facilitate identification of subcategories of patients with high risk of disease relapse, and thereby stratification of patient groups for adjuvant therapy.
Occult metastatic spread to the lymph nodes or distant sites has been the focus of research over many years, and has been reported under different terminology. The Union for International Cancer Control (UICC) has defined micrometastasis as clusters of tumor cells measuring between 0.2 and 2 mm in diameter, and isolated tumor cells as single tumor cells or small clusters of cells smaller than 0.2 mm [
4]. Tumor cells that have spread to lymph nodes or bone marrow are often referred to as DTCs, whereas circulating tumor cells (CTCs) are used for single cells in blood [
5]. A number of previous studies have addressed the prognostic value of detecting micrometastasis and DTCs in lymph nodes of NSCLC patients [
6‐
23], but due to considerable differences in terminology, methodology and results, no conclusion can be drawn based on the existing literature. The methods used for detection have traditionally been immunohistochemistry (IHC) with antibodies targeting epithelial-specific proteins like cytokeratins [
6‐
18], and molecular methods using RT-PCR for detection of tumor- or epithelial cell specific mRNA transcripts [
3,
19‐
25]. Our group has previously published a study where we investigated the presence of DTCs in bone marrow aspirates from patients undergoing lung cancer surgery by the use of immunomagnetic selection (IMS) [
26]. In the present study we have used IMS to investigate the presence of DTCs in lymph node tissue prospectively collected from patients with early stage NSCLC undergoing curatively intended surgery. In the IMS method small magnetic particles coated with the antibody MOC31 which recognizes the epithelial marker EpCAM, are used to isolate tumor cells from the lymph nodes, allowing fast screening of as much as 2 × 10
7 cells. The objective of our study was to determine the incidence of lymph node DTCs, and to compare results obtained with IMS to the pathological staging obtained by histopathology. Additionally, we wanted to investigate the associations between the presence of DTCs and clinical and histopathological variables, as well as to patient outcome.
Discussion
In the present study we have examined the presence and prognostic implication of DTCs in prospectively collected lymph node samples from early stage NSCLC patients undergoing surgical resection. By using the technique IMS with the EpCAM-targeting antibody MOC31, we found a strong association between pN status as assessed by standard histopathology and detection of EpCAM+ cells. The pN status, but not the IMS result, was significantly associated with patient outcome. Among the patients staged as pN0 by histopathology, 32 % were categorized as IMS-positive, but IMS was not able to predict prognosis in this patient subcategory. Taken together, we conclude that detection of lymph node DTCs by IMS does not improve nodal staging in NSCLC patients, and cannot be used for identification of patients with high risk of relapse after surgical resection.
A number of publications have investigated the incidence of micrometastatic disease or DTCs in lymph nodes from patients with NSCLC, the majority using IHC or RT-PCR, with conflicting results. Several groups have shown an association between the presence of micrometastasis or DTCs and poor prognosis [
7,
8,
10,
11,
13,
14,
19‐
23] while other have concluded like in our study, that their presence has no impact on patient outcome [
15‐
18]. Data from the diverse publications are not easily comparable due to differences in methodology as well as in size and composition of patient cohorts. Also, various definitions of the terms “micrometastases”, “occult tumor cells” and “isolated tumor cells” have been used, and many authors who demonstrated negative prognostic value did not specify whether they investigated isolated tumor cells or true micrometastasis [
6,
14,
30]. In 2010, Herpel et al. [
8] showed that nodal micrometastasis specifically defined as small tumor deposits had prognostic significance, however two other studies investigating isolated tumor cells in regional lymph nodes concluded that this did not affect survival in resected NSCLC patients [
16,
17]. One might hypothesize that NSCLC patients with detectable single DTCs have similar prognosis to patients with node negative disease, while patients with micrometastasis may have higher risk of relapse. Interestingly, similar results have been found in breast cancer [
31]. In a paper from 2008, Marchevsky et al. investigated both isolated tumor cells and micrometastases in lymph nodes from 266 stage I NSCLC patients, and concluded that the presence of neither of these were significantly associated with survival. The authors also performed a meta-analysis including their own data and 13 other studies, and concluded on no significant association between micrometastasis and survival [
18]. Thus, based on the current literature and our own results, detecting DTCs in regional lymph nodes does not seem to provide a clinically useful method for identifying subcategories of NSCLC patients with high risk of disease relapse. Previously we have also shown that detection of DTCs in bone marrow by IMS has no prognostic impact in early stage NSCLC patients [
26].
The present study is, to our knowledge, the first to use IMS as method for detection of DTCs in regional lymph nodes from NSCLC patients. This technique was developed in our lab, and has been used in several published papers investigating DTCs in bone marrow [
26,
28,
32‐
35] and lymph nodes [
36] in various cancer types. Evaluation of morphology, size and shape of the MOC31-bead-bound cells is challenging when using IMS, and is dependent on the observer’s skills and experience. In our study we found a high number of MOC31-bead bound cells in the samples, and the cut off value for an IMS-positive sample was set at 10 cells based on experience and our previous publication [
26]. The high number of discovered cells might indicate that some nonmalignant cells are also detected with this method, and in fact lymph node reticulum cells can express epithelial antigens [
37,
38], and weak binding of MOC31 to lymphocytes has been described [
39]. Interestingly, however, in a study using RT-PCR to identify tumors cells in lymph nodes in a NSCLC cohort comparable to ours, Nordgård et al. also reported a high number of positive findings in histopathologically node negative patients [
40]. Another important issue when using antibodies targeting EpCAM is that the expression of this protein may vary in tumor cells during disease progression [
41], and expression can be down-regulated as a consequence of epithelial-to-mesenchymal transition (EMT) [
42]. This might perhaps explain why no EpCAM
+ cells were discovered in 3 of 17 lymph node samples from patients who had metastases detected by routine histopathology. There are clearly several methodological issues associated with the use of IMS, leading to considerable challenges for the further development and application of this method.
Our study is limited by the fact that we only had access to one lymph node sample from one node station per patient. Analyzing a number of nodes from various stations from each patient would of course have been preferable. Moreover, a larger patient cohort and longer follow-up period would increase the reliability of our results. However, we were able to detect EpCAM+ cells in the vast majority of samples from lymph nodes in which metastases had been shown on histopathology, and there was a strong association between results from the IMS method and nodal staging by standard histopathology.
Abbreviations
NSCLC, non-small cell lung cancer; DTCs, disseminated tumor cells; IMS, immunomagnetic selection; UICC, Union for International Cancer Control; PBS, phosphate buffered saline; HSA, human serum albumin; CTCs, circulating tumor cells; IHC, immunohistochemistry
Acknowledgements
The authors would like to thank Ingjerd Solvoll and Ildri Haltbakk for excellent technical assistance.