Circulating tumour cells (CTCs), which are found in the blood of cancer patients at low concentrations, are probably the cause of metastatic disease. A new microfluidic chip can isolate these cells, providing new possibilities for both diagnosis and understanding the biology of metastasis.

Mehmet Toner and colleagues have developed a microfluidic chip that can sort CTCs from millilitre whole-blood samples. Previous applications of such lab-on-a-chip devices for cell sorting have been successful on microlitre samples of buffer solutions, but scaling up to large amounts of unprocessed whole blood has been difficult. The authors developed a chip on which microposts are coated with antibodies against a protein that is overexpressed on the surfaces of cancer cells, epithelial cell adhesion molecule (EpCAM). These microposts sequester tumour cells as the sample flows across them at an optimal rate that was worked out by the authors.

To test their CTC chips, the authors carried out several experiments in which known quantities of cancer cell lines were spiked into fluid samples. Under optimal conditions they recovered over 65% of CTCs, and this was independent of the level of EpCAM expression, which varies by several orders of magnitude between cancer types. The authors then tested actual blood samples from patients. They detected CTCs in 115 of the 116 cancer samples and none of the 20 healthy samples. The concentration of CTCs varied between 5 and 1,281 per millilitre and, interestingly, localized prostate cancers produced similar levels of CTCs to metastatic prostate cancers.

As a further test, the authors assessed whether the CTC counts from their chips could predict tumour volumes in patients undergoing treatment. Although absolute tumour volume did not correlate well with CTC count (presumably because of other factors affecting inter-patient variability), changes in CTC count were reasonable predictors of changes in tumour volume.

This newly developed system has several important advantages, such as its use of unprocessed whole blood samples and its high sensitivity and specificity. One of the most important is that the sorted tumour cells are highly pure and viable, allowing them to be subjected to a full range of molecular techniques for both therapeutic and basic research.