Digitization and computerized images will now help the field of pathology with a new technique developed by University researchers.
The innovative method, called Spatially-Invariant Vector Quantization, will allow pathologists to more accurately and efficiently pinpoint signs of disease. According to its developers, SIVQ carefully targets signs of a disease, such as cell or tissue abnormalities, by analyzing digital images.
The findings were published in the Journal of Pathology Informatics online on Feb. 26. Jason Hipp, co-lead author of the findings, and contributing author Ulysses Balis — both pathologists at the University Hospital — began working on the SIVQ software’s first algorithm in the late ‘90s.
After a series of successes and failures and collaboration with researchers from the Massachusetts General Hospital and Harvard Medical School, the current SIVQ tool emerged as a potential “game changer” for the field of pathology, Hipp said.
Different from other pattern recognition software currently being used by physicians, the SIVQ technique pinpoints cell or tissue abnormalities using circular vectors instead of the traditional square or rectangle-shaped search tools. The circular vectors can identify the features of an image no matter the image orientation.
“The idea of the tool is to help improve diagnoses, (and) help pathologists provide better patient care,” Hipp said. “It adds a quantitative component to a profession that is based on an art.”
Ideally, the tool will allow pathologists to use digital images of the body’s tissues in more prolific ways, according to Balis, who is the director of the Division of Pathology Informatics Division at the University Medical School.
“Now that pathologists have access to digital images, the challenge has been what to do with them,” Balis explained. “Pathologists generally don’t do anything with digital images other than look at them.”
Hipp said the tool can be used to identify carcinomas — an invasive malignant tumor — and certain cancer-prone areas like the prostate.
“The idea would be to have the computer program help the pathologist identify the cancer glands,” he said.
Though SIVQ has the potential to hold “tremendous value” for physicians, Hipp said, it isn’t intended to replace the work of pathologists.
“We want to stress that this is a tool to aid the pathologist,” he said.
Though the tool isn’t likely to enter the medical market anytime soon, Balis said it has been distributed throughout the academic community at institutions like Rutgers University. Cornell University has also expressed interest in the tool.
“We’re still in the very early phases of identifying how this technology can improve productivity … for researchers as well as at the clinical level,” Balis said. “The thought is that within a year or two’s time, this algorithm could be a critical part of workflow. The technology highlights the fact that there’s much to be optimistic about.”