MelaFind Improves Chances for Accurate Melanoma Diagnosis

Will insurers provide coverage for a new device that takes much of the guesswork out of screening for this deadly skin cancer?

Thomas Morrow, MD

Nearly 70,000 Americans will be diagnosed with an invasive melanoma, very difficult to treat, this year. In the past 30 years, the incidence of invasive melanoma has doubled to approximately 28 per 100,000 population per year.

In general, a younger population is affected by melanoma, compared to other cancers, accounting for the average of 22.1 years of lost life attributed to invasive melanoma. Severe blistering sunburn is thought to be the largest single risk factor but the cumulative effect of even natural exposure to sunlight can increase the risk. Other risk factors include a prior history, or a family history, of melanoma; light hair, skin, and eye color; and the use of tanning beds. People with many moles also have an increased risk. Obviously the use of sunscreen and avoiding sunburn are recommended for the prevention of melanoma.

Early detection

Early detection is the most effective approach to preventing death from melanoma. As is true of most cancers, the earlier a melanoma is found, the better the prognosis. Many public health initiatives have long been teaching that a skin lesion with asymmetric shape, irregular borders, varied or multiple shades of color, and changing size and shape are hallmarks of melanoma. But these characteristics are not easily recognized in small and thin lesions, explaining why most melanomas are larger than 6 mm in diameter (about the size of a pencil eraser) at the time of diagnosis.

Thickness also plays a huge role in prognosis — the thinner the lesion the better. Published reports document that a melanoma less than 0.76 mm thick has a 99.5 percent 10-year survival rate. If the lesion is thicker than 3 mm, fewer than half of patients are alive in 10 years. The only foolproof diagnostic approach is biopsy, a procedure that few really wish to have unless needed. The absolute best case for melanoma diagnosis would be for a skin lesion to be accurately diagnosed visually in an early stage. Biopsies would be limited to “dangerous lesions,” and no malignant lesions would remain undetected.

Unmet need

But the sensitivity and specificity of being able to detect a melanoma on physical examination varies widely from physician to physician, regardless of specialty. Some reports suggest that for dermatologists, as many as 50 biopsies need to be performed to detect one melanoma, and the rate can be as high as 80:1 for nondermatologists.

Because of the high biopsy-to-positive rate, an unmet need has existed in the diagnosis of melanomas. For years clinicians have been contemplating the use of computer-guided diagnostic assistance, and now a company called Electro-Optical Sciences has come up with a device that it has named MelaFind. It uses seven images from the visible and near-infrared spectral bands of light and compares the images with a data bank of biopsy-proven images, using a set of algorithms. The algorithms check the quality of the image and determine the properties of each lesion based on the various light frequencies. The device then classifies each lesion as malignant or nonmalignant.

Because the various spectral bands can penetrate the skin to a depth of 2.5 mm, the MelaFind device can actually “see” a lesion under the skin, which the human eye cannot do. The additional data can actually help it to make critical management and diagnostic recommendations.

The final output on this scan is a yes-or-no decision: Should the lesion be biopsied? This decision is based on the algorithms that examine and compare more than 9,000 biopsied skin lesions, including more than 600 melanomas.

Comparisons made

Diagnoses from the MelaFind device were compared to dermatologists’ diagnoses and found to be more accurate in detecting and suggesting biopsy. The results of a clinical trial were published in April 2008, leading to submission of the MelaFind to the FDA. (Friedman RJ, Gutkowicz-Krusin D, Farber MJ, et al. The diagnostic performance of expert dermoscopists vs a computer-vision system on small-diameter melanomas.

The manufacturer performed a study that basically compared the opinions of 10 blinded dermatologists who studied images of lesions to the output of the MelaFind. All of the skin lesions in the study had been biopsied and reviewed by up to three different blinded dermatopathologists to determine whether the lesion was a melanoma or other cell type.

The melanomas selected for the study were all small (≤6 mm), the most difficult type to diagnose. The two questions asked of all experts and of the MelaFind were:

“Is the lesion a melanoma?”

“Would you biopsy or excise the lesion?”

If the answer to the second question was “yes,” the expert was asked to state a reason for the biopsy. This is the most critical question to ask from the patient’s perspective, as a biopsy will “find” a melanoma even if the physician misdiagnosed it visually. Also, if the physician did not choose to biopsy and the lesion was actually a melanoma, the patient would obviously be placed in danger.

Of great interest is that the trial provided evidence that there is a high level of diagnostic and management disagreement among highly trained and experienced dermatologists. Of lesions that were known to be melanomas, the mean decision to biopsy was only 71 percent (range 37-88 percent). Stated another way, the average dermatologist would have missed 3 out of 10 melanomas. The MelaFind was correctly able to make a correct management decision 98 percent of the time. The MelaFind missed one melanoma in situ. Collectively, all 10 experts missed two melanomas in situ.

The study further evaluated lesions that were more advanced, such as invasive melanomas.

The experts recommended biopsy correctly 81 percent of the time (range 71-91 percent) and the MelaFind recommended biopsy for 100 percent of the lesions that were later to be determined to be melanomas.

If approved by the FDA, this tool will give physicians a much higher positive predictive value for the decision to biopsy and provide for a much lower risk of missing a melanoma.

From a managed care perspective, this may significantly decrease the number of invasive melanomas, thus saving money. But there is no reason to believe that this device will be used only by dermatologists. It might become a routine screening device for primary care physicians and nurses, who will want to be paid for this service.


How will health plans manage this? Will it become a standard of care? Will it be included in the normal office visit charge in the way that vital signs are now included? Will providers upcode the office visit to a more complex level? What about mass screenings?

Today many health fairs offer free cholesterol tests. Are free melanoma screenings using the MelaFind far off? These questions are not answered, but the MelaFind appears to be an answer for people who are wondering “Is this lesion dangerous?”

Thomas Morrow, MD, is the immediate past president of the National Association of Managed Care Physicians. He has 24 years of managed care experience at the payer or health plan level. Contact him at [email protected]
The author is a director in the value-based health department at Genentech Inc. He has had no other industry affiliations in the past three years. The views expressed in Tomorrow’s Medicine are the author’s alone.