Gigapixel camera has eye for melanoma

A recently developed gigapixel whole-body photographic camera is like 63 cameras in one, with the potential to image an entire body -down to each mole, freckle and melanoma. 

Daniel Marks, Ph.D., who helped to develop optics for U.S. military surveillance, created a mega-camera that can take a picture of an entire football stadium, revealing the detail of each fan’s face. Mounted on a drone, the camera cand vividly document individual city streets, each car and each person. 

About three years ago, Dr. Marks, an associate research professor of electrical and computer engineering at Duke University, says a colleague suggested he study the technology for use in dermatology. And he did.

“It’s very similar to what’s in the surveillance [technology], but is more portable and photographs a much smaller object than an entire city,” Dr. Marks says.

Promising, but research is needed

Dr. Marks presented the technology and a whole body image of himself at The Optical Society's annual meeting October 2014 in Tucson, Ariz. The purpose was to showcase the camera’s resolution of features with which dermatologists would be familiar.

He and his research team have not yet started, but are planning a pilot study on patients.

The goal is to use the camera to find potentially cancerous skin lesions earlier than would be possible with most skin examination techniques. The technology fills a void in skin cancer detection, he says. A dermatologist might examine a small region with a high-resolution, handheld dermatoscope or a large region at low resolution or visually.

The gigapixel image doesn’t require a compromise between the two, he says.

Using this camera, dermatologists would photograph whole-body images in detail much greater than what is possible with digital photography mole mapping or the human eye. The images, however, would not be as vivid as a dermatoscope’s 10-fold magnification of an individual lesion.

“To kind of give you an idea what the resolution is, it’s about 75 microns, which is about the width of a human hair. You could potentially see a human hair as one pixel wide on the picture,” he says.  

While all this seems promising, its efficacy is currently unproven. Future studies will determine whether the technology helps dermatologists detect skin cancers earlier than possible with what’s available today, as well as how accurately the technology pinpoints worrisome lesions without increasing the rate of false positives. Studies may show, for example, that low-risk patients might not be good candidates for the technology because the potential for a false positive or false negative outweighs the possibility of catching a skin cancer early.  

Potential applications

Dr. Marks sees applications for high-risk patients, where dermatologists would use the camera for whole-body images; then, let a computer do the work of analyzing lesions that need attention.

“It’s much more likely that a computer would have to look through that quantity of data because to examine an entire megapixel image at that resolution is a very time-consuming process. For example, if you printed it out, it would be like a 20-foot wide poster on the wall,” Dr. Marks says.

An important advantage of the technology is that it would provide a complete record of a patient’s skin at a given time, and could be used at intervals to detect the early changes.

“Another potential benefit is telemedicine. A lot of places may have high-speed internet connections but are underserved by specialists,” Dr. Marks says. “It makes the process a lot more accessible to more people.”

How it’s used

The camera is about the size of a small television set or a large microwave oven. It’s not as easy to use as a handheld device. This camera would be set on a clinical cart with mobile lighting and should be used in rooms that have a backdrop for taking pictures and controlled lighting conditions.

Next step

Dr. Marks says he and his team at Duke are building a database of skin cancer biopsy pathology and photographic records.

“If we can correlate specific conditions with retrospective studies, where if we had seen this in the image we could have known to investigate that earlier, then we can design a study that basically tests for those particular signatures and tells us whether or not we’re actually finding the disease earlier than if done visually,” he says. “We’re trying to push this technology forward as fast as we can so dermatologists can get their hands on it.”