A Rare Event Imaging System for Detection of Cancer Cells in Blood and Bone Marrow

Abstract & Commentary

Synopsis: An automated microscope system is described for the detection of rare cancer cells in samples of blood and bone marrow. The Rare Event Imaging System could detect rare cancer cells (1 in 1 million) in hematopoietic tissue samples. The Rare Event Imaging System facilitated analysis for minimal residual tumor (MRT), and additional correlations of rare event imaging and patient disease status are eagerly awaited.

Source: Kraeft SK, et al. Clin Cancer Res 2000;6:

Improved detection strategies are needed for the evaluation and monitoring of patients with cancer. Identification of microscopic subclinical disease in blood or bone marrow may identify patients in need of systemic therapy at a time prior to clinical presentation with disseminated disease. In addition, identification of malignant cells with a demonstrated capability to spread to blood and bone marrow has significant implications for the patient, as disseminated disease is a major cause of cancer-related mortality.

Several strategies have been used to detect minimal residual tumor cells in the blood of patients with cancer. These strategies include attempts to grow, culture, and characterize tumor cells,1 molecular approaches such as reverse transcription PCR,2,3 and immunological approaches to identify and characterize circulating malignant cells.4,5 While flow cytometry can evaluate a large number of cells in a matter of minutes, it remains important to confirm the malignant nature of positive events detected by flow cytometry. Thus, an automated system to identify and analyze rare malignant cells in blood and bone marrow could have significant potential clinical applications.

The manuscript by Kraeft and colleagues describes an automated microscopic system (Rare Event Imaging System) for the detection and analysis of cancer cells in samples of peripheral blood and bone marrow. Samples are obtained from blood and bone marrow, red cells are lysed, and peripheral blood mononuclear cells (PBMC) or nucleated bone marrow (BM) cells are counted and allowed to attach to adhesive slides. The adhered cells are then analyzed by immunological methods for possible tumor cells. The instrumentation required for this analysis include an automated fluorescence microscope (Nikon Microphot-FXA) with a cooled, charged, coupled device camera and a 60-MHz Pentium personal computer.

Kraeft et al initially described the efficiency of the cell deposition method following analysis of PBMC from normal donors and cancer patients, as well as BM samples from cancer patients and stem cell samples from cancer patients. The efficiency of cell deposition did vary, with 89% recovery from normal donor peripheral blood samples, 64% recovery from cancer patient peripheral blood samples, 58% recovery from cancer patient bone marrow samples, and 73% from cancer patient stem cell samples. The sensitivity of the rare event imaging system was then evaluated by spiking breast cancer cells into PBMC samples. The anticytokeratin antibody could readily detect the breast cancer cells in PBMC samples. A serial dilution analysis was performed, and breast cancer cells were detected at the lowest tested dilution corresponding to a detection of one breast cancer cell in 1 million PBMC. To allow additional phenotypic characterization of these cancer cells, and to improve the specificity of the rare event detection analysis, a staining procedure involving double labeling of tumor cells was developed. The double-labeling procedure consisted of labeling with an antibody specific for a surface molecule expressed on various malignancies and combining that analysis with staining for intracellular cytokeratin. The antibodies used included an antibody reactive with epithelial cell adhesion molecule (Ep-CAM) (breast, ovarian, colon, and lung carcinoma antigen) and an antibody reactive with disialo-ganglioside (GD2) antigen (small-cell lung carcinoma, neuroblastoma, and melanoma antigen). Visualization of the MCF-7 breast cancer cells were shown with the Ep-CAM/cytokeratin double-stained approach, and the SW2 small cell lung cancer cell line was shown to have double staining with the GD2/cytokeratin labeling. Thus, dual characterization with a surface molecule and intracellular cytokeratin was clearly demonstrated with these cancer cell lines.

The specificity of the single- and double-staining protocols was evaluated with peripheral blood samples from healthy donors. Sixteen to 18% of the normal donor peripheral blood samples had a low level of cytokeratin-positive cells that ranged from one to 26 labeled cells per 106 white blood cells. In contrast, use of the double-labeling protocol almost entirely eliminated the background positivity, with only a single double-positive cell detected in a total of 77 peripheral blood samples.

Three hundred fifty-five peripheral blood, bone marrow, and stem cell samples were then evaluated from patients with breast cancer before autologous bone marrow transplantation but after high-dose chemotherapy using the single cytokeratin labeling method. Cytokeratin-positive cells were detected in 52% of the bone marrow cells, 34% of peripheral blood samples, and 27% of stem cell samples. The frequency of these cytokeratin-positive cells in the positive samples ranged from one to 1020 cytokeratin-positive cells per million cells evaluated. To control for background reactivity in the normal donor samples, a threshold cut-off value was defined as the mean number of cytokeratin-positive cells plus two times the standard deviation as observed in the control sample. Kraeft et al then report cytokeratin positivity in 40% of the bone marrow samples, 24% of the peripheral blood samples, and 12% of the stem cell samples.

COMMENT by Mark R. albertini, MD

The Rare Event Imaging System was demonstrated to be a sensitive means of detection of rare cancer cells (1 in 1 million) in samples of peripheral blood and bone marrow. The double-labeling protocol resulted in a significant reduction in false-positive determinations. This analysis allowed phenotypic characterization of tumor cells with the double-labeling analysis. Thus, the double-staining protocols for cytokeratin/Ep-CAM and cytokeratin/GD2 can detect rare cancer cells in hematopoietic tissues from cancer patients.

The Rare Event Imaging System appears ready for clinical testing to determine correlations between positivity with this analysis and subsequent clinical outcome. The current article describes and validates the ability of this system to detect rare malignant cells. However, clinical implications of this finding require further investigation. The proposed monitoring system appears ready for this clinical testing, and studies analyzing clinical samples from cancer patients are eagerly awaited. Potential incorporation of events detected with the Rare Event Imaging System together with traditional cancer outcome measurements would provide valuable testing of this technology.

Kraeft et al also describe potential additional phenotypic analysis of samples with markers that are correlated with metastatic potential. The use of this Rare Event Imaging System with additional markers requires further investigation. However, the use of a double-staining procedure with additional surface marker analysis appears to be a logical extension of this technology.

In summary, Kraeft et al describe an exciting technology to identify rare malignant cells in samples from peripheral blood and bone marrow. Additional clinical testing will be required to determine potential use of this monitoring strategy for patients with cancer.


1. Braun S, et al. J Natl Cancer Inst 1998;90:1099-1101.

2. Datta YH, et al. J Clin Oncol 1994;12:475-482.

3. Krismann M, et al. J Clin Oncol 1995;13:2769-2775.

4. Racila E, et al. Proc Natl Acad Sci USA 1998;95:

5. Gross HJ, et al. Proc Natl Acad Sci USA 1995;92:

Which one of the following statements about the Rare Event Imaging System is false?

a. Cytokeratin-positive cells can be detected in peripheral blood samples from healthy donors.

b. The double-labeling protocol can improve the specificity of rare event detection by combining analysis for intracellular cytokeratin with an analysis for epithelial cell adhesion molecule

c. The detection of cytokeratin-positive cells following cytokeratin single-staining is a specific marker for cancer cells.

d. The double-labeling protocol can improve the specificity of rare event detection by combining analysis for intracellular cytokeratin with an analysis for disialo-ganglioside (GD2) antigen.