Challenges in circulating tumor cell detection by the CellSearch system

Abstract

Enumeration and characterization of circulating tumor cells (CTC) hold the promise of a real time liquid biopsy. They are however present in a large background of hematopoietic cells making their isolation technically challenging. In 2004, the CellSearch system was introduced as the first and only FDA cleared method designed for the enumeration of circulating tumor cells in 7.5 mL of blood. Presence of CTC detected by CellSearch is associated with poor prognosis in metastatic carcinomas. CTC remaining in patients after the first cycles of therapy indicates a futile therapy. Here we review challenges faced during the development of the CellSearch system and the difficulties in assigning objects as CTC. The large heterogeneity of CTC and the different approaches introduced in recent years to isolate, enumerate and characterize CTC results in a large variation of the number of CTC reported urging the need for uniform definitions and at least a clear definition of what the criteria are for assigning an object as a CTC.

Keywords: CellSearch; Circulating tumor cells; Epithelial cell adhesion molecule; Rare event detection.

Figure 1

Rare cell detection. Probability distribution of two cell populations one unstained (green) and one stained (blue) appearing at a ratio of 1:1, panel A; 1:1.000 panel B, 1:10.000 panel C, and 1:1.000.000 panel D. The blue and green lines depict the probability distribution.

Figure 2

Immunocytochemistry. Typical images of CTC immuno‐magnetically enriched using EpCAM labeled ferrofluids and stained by immunocytochemistry with hematoxylin nuclear stain (blue purple) and cytokeratin (red). The brown yellow color can be contributed to the ferrofluids that are approximately 175 nm in size and is visual due to the accumulation.

Figure 3

CTC versus survival in breast, colon and prostate cancer. Kaplan–Meier plots of metastatic breast (top), colon (middle) and prostate (bottom) cancer patients with 0, 1–4, 5–24 and >24 CTC before initiation of a new line of therapy. Plots were generated using the data from the studies presented by (Cohen et al., 2008; Cristofanilli et al., 2004; de Bono et al., 2008).

Figure 4

Different appearances of CTC. From top to bottom: EpCAM+, CK+ extracellular vesicles; small apoptotic CTC; CellSearch CTC; large apoptotic CTC; clusters of CTC and lymphocytes attached to CTC. The relative frequencies are indicated on the y‐axis.

Figure 5

Survival versus CTC definition. Kaplan–Meier plots showing the relation with survival of objects, from metastatic prostate cancer patients, subdivided into different groups. A) cells classified according to the CellSearch criteria; B) cells classified as intact CTC; C) cells classified as apoptotic and D) the EpCAM+, CK+ extracellular vehicles (EVs). Plots were generated using the data from the studies presented by (Coumans et al., 2010).

Figure 6

What is and what isn't a CTC. Left part of the figure shows seven areas with objects in which the CellTracks software identified cytokeratin (CK) as well as DNA staining. Shown are, thumbnails of the CK staining as presented by CellTracks software (Normalized CK); CK staining that is not scaled (Unscaled CK); DNA; CD45 and a CK/DNA overlay. To the right the decision tree is provided that the reviewer follows to sore the objects. All seven objects fit the size criteria.