Background. The long-term prognosis after surgical therapy for esophageal c
arcinoma depends on tumor stage and completeness of resection. Similarly to
other epithelial tumors, the presence of micro deposits of neoplastic cell
s in the bone marrow may indicate residual disease and the potential for re
currence. This study assesses the prevalence of bone marrow-disseminated tu
mor cells in patients undergoing surgical resection for esophageal carcinom
a. In addition, we investigated the agreement between immunohistochemical a
nd molecular techniques for the detection of micrometastases in a subgroup
of patients.
Methods. Between January 1998 and November 1999, forty-eight patients with
andenocarinoma of the esophagogastric junction (n = 29) or squamous cell ca
rcinoma of the thoracic esophagus (n = 19) and no evidence of overt metasta
tic disease entered the study. An immunohistochemical assay (capable of det
ecting I carcinoma cell in 7x10(5) Done marrow cells) was used to test Done
marrow obtained by flushing a resected rib or by needle aspiration either
of iliac crest or Of a rib. A polymerase chain reaction (PCR) molecular tec
hnique was also used to identify bone marrow and peripheral blood epithelia
l cells.
Results. Cytokeratin-positive cells were found in 79.1% of the bone marrow
samples obtained from the rib, and in only 8% of the needle aspirates eithe
r from the iliac crest or from a contiguous rib: This difference is probabl
y explained the improved removal of metastatic cells with the flushing of t
he rib. Comparable results were obtained at a qualitative level by the PCR
technique on bone marrow. In addition, PCR-positive results were found in 3
of 18 peripheral blood samples. There was no association with tumor type,
neoadjuvant therapy or lymph node status. Patients with a pT3 or pT4 tumor
showed, at a borderline statistical level, a higher proportion of cytokerat
in-positive cells in the flushed rib.
Conclusions. Bone marrow-disseminated tumor cells are present in the resect
ed rib of a high proportion of patients undergoing esophagectomy for carcin
oma, and immunohistochemistry seems to Of the method of choice for their qu
antitative assessment. However, the prognostic and therapeutic implications
of this finding need further investigation.