Medical Device Daily

With melanoma rates on the rise — and the prognosis for the disease poor unless detected early — Electro Optical Sciences (EOS; Irvington, New York) has initiated a pivotal trial for MelaFind, its noninvasive, point-of-care device to assist in the early diagnosis of this skin disease, which accounts for 80% of all deaths due to skin cancer.

The first patients have been enrolled in the study and the company said it expects to submit the premarket approval application to the FDA for review in the second half of this year.

“The primary endpoint of the study requires us to have 93 eligible and evaluable biopsied melanomas,” Joseph Gulfo, MD, president/CEO of Electro Optical, told Medical Device Daily. “We estimate that will require a total of 1,200 melanomas and melanoma mimics, and that would take about 1,000 patients.”

EOS has a binding protocol agreement with the FDA, and in October, it was notified by the agency that it was granted expedited review.

Gulfo said the PMA application is being submitted in three modules, the first and second of which already have been submitted. The third, final module will contain the clinical data from the current study.

MelaFind is a handheld device that Gulfo describes as a “smart system” that the company is “training” to evaluate —noninvasively — the images that it generates. The device is held over a lesion at the point-of-care, such as a physician’s office, and it indicates, by means of special algorithms, whether the lesion is in need of biopsy.

It makes this determination by evaluating each pigmented lesion that gives a physician cause for suspicion against an EOS database of in vivo images and corresponding histopathological results of more than 5,000 biopsied lesions from more than 3,500 patients.

The “output” — that is, how the device gives the physician a “yes” or “no” result — currently is being designed, Gulfo said.

Though just handheld, the device includes a variety of components: an illuminator that shines specific wavelengths of light, including near-infrared banks; a lens system composed of nine elements that create images of the light reflected from the lesions; a photon light sensor; and an image processor employing algorithms to extract specific characteristics, or features, of the images.

The algorithms are then trained to “separate the population of melanoma from the population of non-melanomas,” according to EOS.

“If that algorithm is set, it’s really a formula,” Gulfo said, adding that “whatever the classifier comes up with as the important features against the population of melanoma in training sets is what we will use in the commercial setting.”

Ultimately, there will be a link-up between the device and the Internet, so that the central offices of EOS can monitor the devices for quality control. The company will also “do some diagnostics on actually images,” Gulfo said.

The pivotal study recently initiated most recently is set up a little differently than how EOS expects the device to work when commercialized, assuming all goes well. In the study, a physician makes a determination as to biopsy a lesion, and once that biopsy is taken, the patient is given the opportunity to participate in the study, Gulfo said.

“We need to know unequivocally when you’re training a computer what this lesion means,” he said. “And ‘unequivocal’ means pathology analysis.”

And the company sees other applications for its technology.

For example, in December, EOS signed an exclusive licensing agreement with KaVo Dental (Biberach, Germany), a dental equipment manufacturer. That agreement calls for KaVo to further develop and commercialize Difoti, an FDA-cleared non-invasive imaging device developed by EOS for the detection of dental decay.

KaVo will pay EOS an undisclosed up-front sum as well as up to a double digit annual royalty based on the number of systems sold each calendar year. EOS says that KaVo has made a multi-million-dollar commitment to refine the DIFOTI product for commercial launch.

In November of last year, the company reported a $13.2 million private placement.

Regarding future long-term strategy, Gulfo said the board will “consider these strategic decisions as opportunities arise,” but for now will focus on its platform as it is demonstrated in Melafind for the early diagnosis of melanoma.

“There are many other applications of that platform for other diseases, and we are interested in pursuing those,” he said.

Founded in 1989, EOS, which trades on Nasdaq, has about 30 employees.

The American Cancer Society (ACS; Atlanta) estimates that in 2007 there will be 59,940 new cases of melanoma in the U.S., and there are expected to be 8,110 deaths. ACS says that while it accounts for only 3% of all new skin cancer cases, it causes the greatest number of deaths.