BACKGROUND

Patients with pancreatic cancer often have tumor recurrence despite curative resection. Cancer cells detected in blood or bone marrow at the time of diagnosis may relate to tumor stage and to prognosis. Recent research emphasis has centered on tumor cells in bone marrow aspirates, but whether these represent early micrometastases or blood-borne cells in transit is unknown.

PATIENTS AND METHODS

We developed a specific immunocytochemical assay that evaluated more than 5.3 x 10(6) extracted mononuclear cells per sample of blood and bone marrow and that could identify a single tumor cell in that population. The assay was applied to samples of blood and bone marrow from 105 patients with pancreatic cancer and 66 controls. The prevalence of isolated tumor cells was compared with Union Internationale Contre le Cancer (UICC) stage. A multivariate Cox regression analysis for survival was performed.

RESULTS

Pancreatic cancer cells were detected in 26% of blood samples and in 24% of bone marrow specimens. Specificity for cancer was 96%. The prevalence of isolated tumor cells in patients with proven resectable cancer was 9% in blood and 13% in bone marrow. The prevalence increased with UICC tumor stage in blood (P =.04) but not in bone marrow (P =.52) and correlated in blood with resectability (P =.02), progression of disease (P=.08), and peritoneal dissemination (P =.003). While survival correlated significantly with tumor stage (P <.001) and isolated tumor cells in blood correlated with tumor stage, the finding of cancer cells in blood or bone marrow, or both, was not independently associated with survival in patients with pancreatic cancer.

CONCLUSIONS

Isolated tumor cells in blood but not bone marrow reflect the stage of growth and spread of pancreatic cancer, particularly in the peritoneal cavity. The findings are consistent with cells in bone marrow aspirates being in transit, not implanted. These disseminated cancer cells may be the consequence, rather than the cause, of progression.