Bedside Cancer Detector
Portable NMR and smartphone combo could diagnose in minutes
Photo: C. Min/H. Lee/R. Weissleder/Harvard Medical School
2 March 2011—A handheld device that detects proteins produced by tumor cells could give doctors a fast, accurate way to diagnose and monitor cancer. Tumor-marker testing usually requires pea-size tissue samples and can take days in a laboratory to yield results. The new detector, by contrast, requires a tiny speck of tissue, takes less than an hour to process samples, and could be used in a doctor’s office instead of a hospital.
The device, which was developed by researchers at Harvard Medical School, attaches to a smartphone, providing a user-friendly interface for doctors to view results. In tests on patient tissue samples, the researchers accurately detected cancer 96 percent of the time. They reported the results last week in the journal Science Translational Medicine.
Doctors typically use tumor-marker testing together with a biopsy to diagnose cancer. Marker levels can help them choose the appropriate therapy and later check to see if it is working. To get the tissue samples, they typically use 2-millimeter hollow-core needles, which carry a risk of bleeding and infection. Pathologists then examine the sample under a microscope.
The new tool can detect standard tumor markers and their concentrations in samples taken with a fine needle—a 0.5-mm core. It gives results in as little as a half hour. ”We think the device could support big medical decisions like whether to do surgery on a patient to look at the malignancy or refer them to more intensive CT or MRI scans,” says Hakho Lee, a biomedical engineering professor at Harvard who developed the technology. ”If a patient is already getting chemotherapy, the doctor could quickly tell whether a treatment is working.”
The detector is a miniature nuclear magnetic resonance (NMR) machine. NMR is used to determine the chemical structure of organic molecules. It involves aligning atoms in a strong magnetic field and making them vibrate under the influence of a radio-frequency signal. Conventional machines use huge magnets in order to make the atoms vibrate strongly enough to measure their natural oscillation frequency.
The new detector’s magnet can be smaller because the sample volume is also smaller and because the system measures only how quickly the atoms’ vibrations decay over time rather than measuring the frequency, as a conventional NMR does. The test sample goes on a thumb-size probe, which also contains the chips that process the data and communicate with the smartphone.
The researchers used the device to analyze abnormal stomach tissue samples from 50 patients who were suspected to have malignant tumors. They detected NMR signals for nine different cancer protein markers and identified a group of four markers as the best tumor indicators. Then, looking only for NMR signatures of these four markers, they correctly identified cancer in 96 percent of the cases. They managed to improve accuracy to 100 percent in a later set of 20 patients. Standard pathology testing on a larger biopsy specimen from the same tumors was only 84 percent accurate.
”It’s an exciting technology and could be used for cancer prescreening at a patient’s bedside,” says Shan Wang, an electrical engineering professor at Stanford University who is developing magnetic nanoparticle-based sensors for tumor-marker detection that rely on detecting a small electrical resistance change. He points out, however, that while the device is very sensitive, it also gave false positives, identifying some healthy people as having cancer.
Andrew Seidman, a clinical oncologist at Memorial Sloan-Kettering Cancer Center, says that the detector’s appeal lies in being less invasive than core-needle biopsies. He believes it would be most helpful for guiding cancer therapy. ”Tumors evolve during treatment,” Seidman says. ”I can see this as a safer way to do serial biopsy—to reassess the biology of a patient’s tumor and adjust treatment as the cancer progresses.”
In the future, the technology could replace core-needle biopsies for cancer diagnosis, Seidman says, but right now, ”it’s not ready for prime time. It’s a promising technology but needs to be further refined and validated in larger cohorts.”
About the Author
Prachi Patel is a contributing editor to IEEE Spectrum and a freelance journalist in Pittsburgh. In the February 2011 issue, she wrote about the benefits and choices that engineers have in business school.