|Through a collaboration with researchers at The University of Texas M.D. Anderson Cancer Center, UH Associate Professor of Computer Science George Zouridakis is working on a new method for early detection of skin cancer.
The method involves computer analysis of skin lesion images, which are obtained using techniques called transillumination and cross-polarization imaging.
Combined, these techniques create under-the-skin illumination, showing development of new blood vessels around freckles and moles, which is a likely indicator of cancer.
Current methods for skin cancer detection rely on a visual examination of the skin and skin-surface photographs to record any changes on the lesion’s pigmentation and geometry over time. Zouridakis and his collaborators are focusing on changes underneath the skin.
“Our method is a tool to assist physicians in making a diagnosis,” explains Zouridakis. “Preliminary studies show that with this method, the classification of lesions as malignant or benign jumps from 58 to 86 percent accuracy, which is a huge improvement.”
“If detected early, skin cancer has one of the highest cure rates of all cancers, and treatment is relatively simple,” says Zouridakis.
A hand-held device that uses this technology is currently under development and is expected to be available next year. For more information about this work, please visit www.cs.uh.edu/~zouridakis.