Detection of Micrometastases in Histologically Negative Lymph Nodes in Esophageal Cancer

Edmund S. Kassis, MD

Section of Thoracic Surgery
University of Pittsburgh Medical Center

I would like to thank the Society for giving us the opportunity to present our data today. I will be talking about the detection of micrometastases in histologically negative lymph nodes in esophageal cancer. First I would like to acknowledge our multidisciplinary team. I would like to thank Dr. Luketich for the opportunity to work with him over the past two years.

Esophageal cancer accounts for 1.5% of all malignancies in the United States. In 1997, 12,500 new cases were diagnosed, and the overall 5-year survival is a dismal 5-10%. The epidemiology of esophageal cancer is changing dramatically.

In this slide borrowed from the NIH cancer surveillance program, on the y-axis esophageal cancer per 100,000 person-years from the mid-1970s to the mid-1990s. The circles represent adenocarcinoma and the squares represent squamous cell carcinoma. The open symbols indicate white males while the closed symbols indicate African-American males. From this slide, adenocarcinoma is increasing at an alarming rate in both white males and African-American males.

This quote by Dr. Kelsen in Chest describes the current role of neoadjuvant therapy in the treatment of esophageal cancer. "At present, chemotherapy given alone or with concurrent radiation prior to planned operation is investigational... Surgery alone is the standard treatment for esophageal cancer." (Kelsen DP, Chest 1995) So one may ask, "Why has effective neoadjuvant treatment of esophageal cancer been so elusive?"

Some of the problem areas in evaluating esophageal cancer are the late presentation of the disease, so that as many as 50% of patients have systemic disease at presentation. Also, the results of some phase II studies have improved survival, but compared to historic controls and have not been verified by randomized studies. There is a long accrual time to complete randomized studies. Fourth, accurate pretreatment staging has not been available in any of the randomized studies to date.

So how important is accurate local regional staging? Current data reported by Dr. Lierut of Belgium does show a clear-cut stratification of survival based on radical esophagectomy pathology. The presence or absence of nodal metastases is an important predictor of survival, and while Dr. Lierut has shown that accurate staging by radical esophagectomy is possible, we do not currently have an accurate pretreatment staging modality that would allow us to better assess the effects of new neoadjuvant protocols. Current methods of detecting lymph node metastases are inaccurate.

Conventional scanning is inaccurate in over 40% of patients, as reported by Dr. Krasnack at the University of Maryland and by our group at the University of Pittsburgh. Endoscopic ultrasound is accurate in assessing the depth of tumor invasion, but its utility in detecting lymph node metastases has been questioned. Minimally invasive surgical staging has been shown by our group to be accurate, but requires general anesthesia and extensive lymph node dissection to avoid sampling errors, and requires a 24 to 48-hour hospital stay. Also, histologic examination may miss small foci of metastatic disease since there have been reports where up to 20% of histologically negative lymph nodes are positive on reexamination, which suggests a significant sampling error by pathologists. Our group has found PET scanning to be useful in detecting distant metastatic disease, but only 44% sensitive in the detection of lymph node metastases. Today I will discuss our investigation into reverse transcriptase-polymerase chain reaction to detect CEA mRNA in lymph nodes.

Reverse transcriptase-polymerase chain reaction, or RT-PCR, is a molecular biologic technique that is capable of amplifying minute quantities of mRNA, and through the use of specific gene primers, we can detect gene expression. It is a sensitive, simple, and rapid technique to perform. RT-PCR will only detect the mRNA in metabolically active tissue and will not detect circulating CEA protein. So why did we search for the mRNA for CEA? CEA is an accepted tumor marker that is expressed in a variety of malignancies. It is expressed in normal epithelial tissue, in adenocarcinoma and squamous cell carcinoma of the esophagus. Normal lymph nodes do not express CEA to any quantifiable degree, and we hypothesized that metastatic cells in lymph nodes produce CEA that may be detected by RT-PCR.

The objective of our study was to determine if RT-PCR for CEA mRNA would identify lymph nodes micrometastases undetected histopathologically. We obtained lymph node samples during minimally invasive surgical staging for patients with esophageal cancer. We also obtained lymph nodes from patients with benign disease such as achalasia or gastroesophageal reflux disease. During esophagectomy, malignant and benign esophageal tissues were also obtained. Each sample was bisected and one half was frozen in liquid nitrogen and stored at -70 degrees until RT-PCR. The other half was examined histopathologically and the results were compared.

This is an overview slide depicting the RT-PCR process. As I mentioned, samples are obtained from the operating room and stored at -70 degrees until RT-PCR. The first step of the RT-PCR process is RNA isolation, and here we use a guanidine-thiocyanate process which inhibits ubiquitous RNAases to maintain the quality of the RNA. The RNA is then reverse transcribed into complementary DNA, or cDNA. Then, through the use of gene primers which can anneal to a specific segment of the gene of interest, we can amplify this segment by the polymerase chain reaction, and the polymerase chain reaction can amplify this millions of times. This is evaulated by using ethidium bromide gel electrophoresis and we use a known molecular weight base pair ladder so we can identify the products of the PCR based on their molecular weight.

Using this technology, we have evaluated 123 lymph nodes from 30 patients with esophageal cancer. Fifty of the 123 samples, or 41%, were histologically positive. Eigthy-six of the 123 samples, or 70%, were positive by RT-PCR. Therefore, of the 73 histologically negative samples, 36 or 495 were positive by RT-PCR. Importantly, all histologically positive samples were also positive by RT-PCR and we have had no false-negative results.

This is an ethidium bromide gel electrophoresis showing the products of an RT-PCR reaction. On the left, we have a 100-base pair marker; on the bottom, we have the result of histopathology. PC indicates a positive control, which is an adenocarcinoma from the esophagus, and indicates a positive CEA band which correlates with 131-base pair. NC is a negative control, which is a lymph node from a patient with achalasia. The positive sign indicates positive histopathology and the minus sign indicates negative histopathology. Here we have a specimen which is positive both by histopathology and by RT-PCR. This is a specimen that is negative by histopathology and negative by RT-PCR. This specimen here is a lymph node that showed a negative histopathologic result, but a positive RT-PCR result. This indicates that a site of micrometastasis was present in that lymph node that was undetected by histopathologic evaluation.

Also we studied 26 lymph nodes from patients with benign esophageal disease, such as gastroesophageal reflux. All 26 lymph nodes were negative by histopathology and by RT-PCR, indicating that we have no false-positive results. Sixteen specimens of esophageal cancer and 11 specimens of normal esophagus were studied and these all showed a positive band for CEA.

Although the objective of the study was to determine of RT-PCR to detect lymph node metastases, we did take a preliminary look at survival. In 10 patients, all local regional lymph nodes were histologically negative. In 5 of these patients, RT-PCR was positive. In this group of 5, 3 patients have recurred and 2 patients have died of their disease within a few months of presentation. Median survival for this group of patients was 6.6 months. In the five patients that were negative by histopathology and by RT-PCR, 4 patients are alive without evidence of disease and 1 patient is alive with disease at a median survival of 10.8 months.

In conclusion, RT-PCR increases the detection of lymph node micrometastases in esophageal cancer. Fifty percent of histologically negative lymph nodes have molecular evidence of micrometastases. In our study, RT-PCR changed the stage from N0 histopathologically to N1 by RT-PCR in 5 of 30 patients. Additionally, an increased number of lymph node metastases detected by RT-PCR was found in two-thirds of patients. Positive RT-PCR in histologically negative lymph nodes may have similar clinical prognostic implications as histologic N1 disease. Negative RT-PCR with negative histopathology appears to translate into a favorable prognosis. We stress that the clinical results are only preliminary, and the objective of our study was to determine the feasability of this assay. Further clinical studies will be needed to confirm the clinical implications.

Thank you.