Evidence for disseminated tumor cells in lymphatic vessels afferent to sentinel lymph nodes in patients diagnosed with cervical cancer

Abstract Background In patients diagnosed with cervical cancer, the purpose of lymphadenectomy is the removal of lymph nodes for diagnosis and potential treatment of metastasized tumor cells. It is unclear if afferent lymphatic vessels harbor tumor cells and, thus, may pose additional risk for recurrence or progression if not removed. Aim In this feasibility study, we analyzed the lymphatic vessels afferent to sentinel lymph node (SLN) using a highly sensitive and specific molecular marker for cervical cancer cells. Methods and Results Twenty patients diagnosed with cervical cancer of FIGO stage IA1 to IIB2 underwent laparoscopic SLN removal. Labeling was done using patent blue and the afferent lymphatic vessels were harvested from the parametric tissue and frozen at −80°C. HPV DNA type was evaluated in the primary tumor. Lymphatic vessels afferent to the sentinel lymph nodes were analyzed for the presence of viral oncogene transcripts of the respective HPV type. In one of 18 patients, all with tumor stage ≤IBI and pN0 by conventional histopathology, HPV mRNA could be detected in two of four lymphatic vessels, whereas at least one of the lymphatic vessel biopsies of both patients with tumors ˃4 cm and pN1 status was HPV mRNA positive. No clinical correlation with recurrence after a median follow‐up of 9 years was noticed. Conclusion HPV mRNA indicative of disseminated tumor cells could be detected in lymphatic vessels. The relevance of harvesting lymphatic vessels afferent to SLN in order to increase oncologic safety will have to be investigated in a future prospective study.


| INTRODUCTION
In patients with cervical cancer, metastatic spread to lymph nodes is the most important prognostic marker and is decisive for the selection of therapy. Histopathological examination is standard medical care for the past 100 years. In order to decrease morbidity associated with systematic lymphadenectomy, sentinel lymph node removal is accepted for women with low volume tumors. 1 To avoid falsenegative results associated with selective bilateral pelvic sentinel lymph node harvesting, ultra-staging was investigated in order to identify micrometastasis and isolated tumor cells. Immunohistochemical staging for cytokeratin markers is simple and convenient 2 but its clinical predictive value is controversially discussed. [3][4][5] In terms of molecular markers, it was shown that HPV type specific mRNA (viral oncogene transcripts) is not only more specific than cytokeratin19 mRNA for the detection of tumor cells but is also of prognostic significance in patients with pN0 status. 6,7 Several studies have recently been initiated to reduce radicality of treatment where only sentinel lymph nodes and no parametria are removed, [8][9][10] leaving afferent lymphatic vessels accompanied with minute lymph nodes in situ. Tracers such as patent blue or infrared fluorescence from indocyanine green 11,12 allow identification of small lymphatic vessels, whereas radioactive labeling is only suitable for lymph node identification alone.
Selective removal of SLN without the afferent lymphatic vessels and/or preservation of the parametria, which contain the afferent vessels, may lead to the risk that tumor cell containing lymphatic vessels remain in situ and may lead to loco-regional recurrence. In our study, we address the following questions: • Do lymphatic vessels afferent to the sentinel lymph node contain micrometastasis or isolated tumor cells?
• Does presence or absence of tumor cells in lymphatic vessels correlate with the histopathologic status of the respective SLN?

| Patients
In this monocentric prospective study, 20 patients (aged 26-47 years, median 33.5) with stage IA1 to IIB2 cervical cancer predominantly of squamous cell histologic type were included (Table 1) Sentinel lymph nodes were identified by laparoscopy using patent blue as a tracer. In our cohort, two patients with tumors >4 cm were included because of the high likelihood to detect HPV mRNA in lymphatic vessel biopsies. Of each patient, at least one tissue sample comprising afferent lymphatic vessels (ie, leading from the primary tumor to the SLN) was removed and shock frozen at −80 C for subsequent molecular analysis. For one of the two patients with large tumors, no SLN could be detected, instead one lymphatic vessel biopsy afferent to a lymph node and a lymph node biopsy were taken. After surgery, all lymph nodes were paraffin-embedded and examined by standard histologic examination. Ultra-staging of SLN was not performed.

| HPV-genotyping of primary tumors
HPV-genotyping was performed using genomic DNA extracted from FFPE primary tumors. Thirty-seven different genital HPV types, including all hrHPV types, can be amplified by the GP5+ and GP6+ bio assay. 13

| STR analysis
Sample identity of all primary cancers and their corresponding lymph vessel biopsies were confirmed by STR analyses. The standard ESS DNA profiles (amelogenin, vWA, SE33, THO1, D21S11, D8S1179,   D3S1358, FGA, D18S51, D16S539, D19S433, D1S1656, D2S441, D2S1338, D10S1248, D12S391, and D22S1045) were obtained by amplification with the multiplex PCR Kits Powerplex ESX17 Fast (Promega) and ESSPlexSE QS (Qiagen) following the manufacturer's instructions with reduced PCR volume of 12.5 μL employing 1 ng DNA. A positive control (corresponding to the respective multiplex kits) and a negative control (sterile water) were analyzed in each amplification. All amplifications were done in a T3000 Thermal Cycler (Biometra, Göttingen, Germany). Amplification products were separated and detected on the ABI3500 Genetic Analyzer (ThermoFisher).
Alleles were assigned in comparison to the corresponding allelic ladders. Electrophoresis results were analyzed with GeneMapper ID-X Software. Allele peaks were interpreted when greater than or equal to 100 RFUs. Instead of a lymph vessel biopsy, a lymph node biopsy was analyzed.

| Histological assessment of lymphatic vessels biopsies
The presence of lymphatic vessels could be confirmed in HE-stained sections in all but one biopsy ( Figure 1). In most biopsies, varying amounts of fat, connective tissue, blood vessels, and nerve cells were also evident. Moreover, in 10 biopsies, small lymph nodes (≤1 mm) were also detected ( Table 1 and Figure 1). One biopsy comprised a lymph node with metastatic disease.

| Correlation between HPV-positive lymphatic vessels and the histopathology of the corresponding lymph nodes
With exception of two cases (ID #9 and #12 in Table 1), all patients were pN0. One lymphatic vessel biopsy of each patient with pN1 status contained HPV16 oncogene transcripts. In another patient, two of four lymphatic vessel biopsies were HPV mRNA positive despite pN0 status (ID #19, Table 1). The details of the above three cases are as follows:

| DISCUSSION
The presence of circulating tumor cells in blood is an independent predictive factor in cancer, for example, breast, prostate, colorectal, 16  Somewhat unexpected is the observation that three lymphatic vessel biopsies afferent to sentinel nodes harbored HPV mRNA although the sentinel nodes were free of metastasis by histopathology (ID #12 and 19 in Table 1). The most likely explanation for this phenomenon is that micro-metastases were missed by conventional histopathology.
In two patients (ID #12 and #13), sentinel lymph nodes were detected only unilaterally, and in one patient (ID #9) detection failed completely. This finding is not uncommon for patients with tumors It should be emphasized that viral oncogene transcripts are highly specific molecular markers for the detection of viable tumor cells. The constitutive expression of viral E6 and E7 genes is required for the maintenance of the transformed phenotype. 20 In the context of lymph nodes or lymphatic vessels, the detection of HPV DNA would not allow to discriminate between viable tumor cells and fragmented viral DNA in lymphocytes or its presence in interstitial fluid. In two recent studies, using commercially available kits, 4 out of 125 (3.6%) and 9 out of 134 (8%) histologically free lymph nodes were positive for viral oncogene transcripts. 21,22 These observations are in line with previous studies 6,7 and underscore the high diagnostic potential of viral oncogene transcripts.
Due to the small number of patients and the surgical resection of afferent vessels in our study, we cannot hypothesize if removal of tumor cell containing lymphatic vessels is of therapeutic value. In the three patients with positive lymphatic vessels no recurrence of tumor was observed after more than 5 years follow-up. Only 1 out of 20 patients (ID# 6) recurred locally, underwent surgery and chemoradiation, and is free of disease with a follow-up of more than 6 years: her lymphatics were free of HPV mRNA.

| CONCLUSION
HPV mRNA indicative of disseminated tumor cells could be detected in lymphatic vessels. The relevance of harvesting lymphatic vessels afferent to SLN in order to increase oncologic safety will have to be investigated in a future prospective study. Moreover, in such a study, the tissue sections, which would normally be discarded during ultrastaging of SLN, could be used to determine HPV mRNA levels and would thus also permit a direct comparison of both approaches.

| Limitations
Although the primary aim of this study was to provide evidence for disseminated tumor cells in lymphatic vessels, ultra-staging of the respective SLN would have provided valuable additional information.
Clearly, routine histopathological examination of SLN is not sensitive enough to detect micro-metastases or small tumor cell clusters. This aspect will be kept in mind for conception of a subsequent study.

ACKNOWLEDGMENTS
We cordially thank all patients who participated in this study. This study was financed by intramural funds from the Jena University Hospital and the Charité-Universitätsmedizin Berlin. Open Access funding enabled and organized by Projekt DEAL.