Microfluidic device that captures circulating tumor cells could give doctors a noninvasive way to diagnose and track cancers

Doctors typically diagnose cancer via a biopsy, which can be invasive and expensive. A better way to diagnose the disease would be to detect telltale tumor cells floating in the bloodstream, but such a test has proved difficult to develop because stray cancer cells are rare, and it’s difficult to separate them from the mélange of cells in circulation.

Now researchers from Massachusetts General Hospital and Harvard Medical School say they’ve built a microfluidic device that can quickly grab nearly any type of tumor cell, an advance that may one day lead to simple blood tests for detecting or tracking cancer.

In a given blood sample, circulating tumor cells are rare—there may be only one tumor cell for every billion cells. The device combines existing microfluidic techniques of cell sorting into a single device, he says. The result is that the tumor cells can be pulled out of a blood sample quicker, and without prior knowledge of their molecular characteristics.

Their latest chip can isolate circulating-tumor cells in the blood, and could apply to all types of cancer in one or two hours.

The device developed by Toner’s group combines magnetic labeling of cells and microfluidic sorting to process a sample of blood in about an hour or two. To capture tumor cells regardless of their cancer type, the system first tags white blood cells with magnetic beads that are covered with antibodies that recognize proteins on the surface of the immune cells. The sample is then passed into microfluidic chambers that clear out red blood cells, plasma, and unused free magnetic beads based on their size. Then the device discards the tagged white blood cells using a magnetic field. “In the past, we were focused on tumor cells that we know very little about,” says Toner. “Here, we throw away the cells we know everything about, the blood cells,” he says.

If you liked this article, please give it a quick review on ycombinator or StumbleUpon. Thanks