Lola Butcher reported earlier this week in her Practice Matters blog on how smartphones were changing clinical practice. Now, a study published in this week’s Science Translational Medicine shows that smartphones are even making their way into cancer diagnostics
Researchers from Massachusetts General Hospital and Harvard Medical School report that a micro-nuclear magnetic resonance (µNMR) system linked to a smartphone accurately diagnosed patients with suspected intra-abdominal malignancies 96% of the time.
In fact, the system, which simultaneously measured four cancer-related proteins in fine needle aspirates, was more accurate than standard histology techniques (74% accuracy for fine needle aspirates and 84% for core biopsies).
In the initial clinical test, the investigators, led by Ralph Weissleder, MD, PhD, Director of the Center for Systems Biology and Attending Clinician at Massachusetts General Hospital, measured nine proteins in fine needle aspirates from 50 patients. (Patients then underwent conventional core biopsies for standard of care diagnosis and 44 patients were found to have malignant disease.)
The proteins measured included EpCAM, B7-H3, HER2, EGFR, MUC-1, CK18, Vimentin, p53, and Ki-67.
The best any single protein could do for diagnostic accuracy was 70%, with CK18. However, when the team combined multiple proteins into one assay, the accuracy jumped to 92% for a three-protein combination (MUC-1, HER2 and EGFR) and to 96% for a four-protein combination (MUC-1, HER2, EGFR and EpCAM).
The sensitivity and specificity for the four-protein combination was 100% and 67%, respectively.
The bedside test also was faster than standard histology-based diagnosis. Instead of several days, the µNMR test typically took less than one hour.
Remarkably, that speed might not just be a comfort to patients who otherwise have to wait days for the results, but it might also prove critical in proteomic analysis.
The researchers found that the protein markers degraded rapidly after removal from the body, with a 100% decline in one hour. The degradation continued over the next three hours until it leveled off with approximately a 400% decline in the markers.
Though this point on protein marker stability may seem like a technical side note, it is critically important. And the field should pay keen attention to it. As Carolyn Compton, MD, PhD, Director of the NCI's Office of Biorepositories and Biospecimen Research, told me last fall, “The simple way to think about this is you can have the perfect analytical tool and still get the wrong answer if you compromise the specimen you are analyzing.” (The complete interview was in OT’s 12/10/10 issue.)
Dr. Weissleder’s group validated their four-protein µNMR assay on an independent set of 20 patients. In that case, the assay accurately diagnosed 100% of patients, 14 of whom were found to have malignant disease by standard methods.
The work was funded by NIH grants and the authors declare no competing interests. The µNMR system, however, is described in the paper as a “prototype” and presumably has some commercial value; it is not clear from reading the paper who owns the intellectual property rights.
(Am I showing my renewed sensitivity to potential conflicts of interest? Yep, I’m afraid the Duke saga has me thinking twice, even in the face of very cool science.)