Laser-based imaging pinpoints proteins in cancer
JUNE 15--Researchers at Vanderbilt University (Nashville, TN; www.vanderbilt.edu) have combined molecular imaging with mass spectrometry to improve their ability to identify and quantify the production of proteins in potentially cancerous tissue.
JUNE 15--Researchers at Vanderbilt University (Nashville, TN; www.vanderbilt.edu) have combined molecular imaging with mass spectrometry to improve their ability to identify and quantify the production of proteins in potentially cancerous tissue. By pinpointing the precise location of cells that are producing high levels of a protein thought to allow tumors to grow, the 337-nm-laser-based technique is expected to lead to improvements in the diagnosis and treatment of cancer.
While traditional mass spectrometry can generate a spectrum of proteins found in extracts based on their molecular weight or mass, imaging mass spectrometry takes this one step further by determining the location of specific proteins in the tissue and creating molecular photographs of these proteins. The result is a chemically based means of imaging that shows the distribution of individual proteins within tissues and can distinguish between normal and disease states. Researchers expect this technique to be of benefit some day in intraoperative assessment of the surgical margins of tumors, replacing conventional microscopy methods.
"What we did was devise hardware and software that would allow you to take a piece of human or animal tissue and put it inside this instrument. Maintaining the integrity of the tissue is no easy trick," says Richard Caprioli, director of the Mass Spectrometry Research Center at Vanderbilt University School of Medicine.
"One image might have 30,000 pixels, and each pixel contains 100,000 data points. How does a human being sift through all these data? " he says."The idea is to use artificial intelligence and teach the computer to use the protein patterns to classify the tumors and develop molecular information for pathologists. And down the road, after many hundreds of samples, the computers would learn, and we would be able to look at protein patterns and identify the kind of cancer and where it is located."