Computer-aided Multispectral Digital Analysis (MelaFind) for Assessing Atypical Skin Lesions

Evidence Review

Many insurers have not yet updated their coverage policies since the FDA cleared this scanning tool in November 2011

A summary of ECRI Institute’s Emerging Technology Evidence Report

Managed care leaders are striving to make evidence-based decisions about new and emerging health technology. Managed Care and ECRI Institute have joined in a collaboration to bring bi-monthly summaries of the Institute’s Emerging Technology Evidence Reports. ECRI Institute is an independent not-for-profit organization that researches the best approaches to improving patient care by analyzing the research literature and data on clinical procedures, medical devices, and drug therapies.

Melanoma is a potentially deadly skin cancer that arises in cells called melanocytes, which make the brown pigment melanin.

Early diagnosis and prompt treatment improves survival. Melanoma that is confined to the epidermis (i.e., melanoma in situ) is curable in virtually all cases by simple excision. However, when melanoma advances or metastasizes, it is difficult to treat. When treated with conventional chemotherapy, most patients die within 6 to 10 months. However, with the advent of newer treatments, including targeted inhibitors and immunotherapies, median survival time, in some patients, has increased to more than one year.

Histopathologic examination of biopsy samples is the standard for definitive diagnosis of melanoma. Clinicians identify and biopsy suspicious lesions, but early malignant lesions are difficult to distinguish from benign irregular and atypical pigmented skin lesions and moles. Current screening methods involve visual examination with the naked eye or a dermascope. Visual assessment alone has been associated with a risk of false-negative diagnoses (missed melanomas) as well as false-positive diagnoses, which lead to unnecessary biopsies. Dermoscopy provides 10X or higher magnification and close surface lighting to improve visual clinical examination that enables inspection of structures that the naked eye cannot see. Lesion features that a clinician may inspect during dermoscopy include pigmentation patterns, lesion shape, and changes suggestive of cancer. Use of dermoscopy is intended to reduce false-negative melanoma diagnoses and may often address the false positives (such as the identification of characteristics of pigmented seborrheic keratoses, a benign epithelial lesion).

The MelaFind System

The MelaFind System is a “noninvasive, multispectral computer vision system that uses light from visible to near-infrared wavelengths to evaluate skin lesions up to 2.5 mm beneath the skin.” Multispectral analysis is intended to enable dermatologists to assess skin lesion properties that are not visible to the human eye to improve accuracy of lesion categorization. According to the manufacturer, MelaFind can help dermatologists improve the accuracy of lesion categorization, potentially increasing melanoma detection rates and reducing the number of unnecessary biopsies used to rule out melanoma in patients with benign lesions.

The findings to our key questions are as follows:

1. Does use of the MelaFind system added to standard practice (i.e., visual examination, dermoscopy) increase melanoma detection rates?

The results of one study suggest that the addition of MelaFind to standard practice increases melanoma detection rates; however, the study’s small size and case-control design preclude definitive conclusions about its impact or value in a clinical setting.

2. Does use of the MelaFind system added to standard practice (i.e., visual examination, dermoscopy) decrease the number of unnecessary biopsies?

The results of two low-quality studies suggest that adding MelaFind to standard practice is unlikely to decrease the number of unnecessary biopsies to a clinically relevant degree.

3. What adverse events (AEs) are reported in studies of the MelaFind system?

A small number of false-negative results (0.17 percent) were reported in two studies of the MelaFind system, which could delay diagnosis and treatment. In the study conducted for the premarket approval application submission, one melanoma in situ, one invasive melanoma, and one high-grade lesion were not detected. In the smaller Friedman et al. study, one melanoma in situ was missed. If clinicians rely heavily on the “negative” findings of the MelaFind device to recommend against biopsies, some melanomas may be missed.

Diffusion

MELA Sciences planned to launch MelaFind to about 200 selected “high-volume,” integrated dermatology and skin cancer specialists’ practices. As of June 2012, MELA Sciences has distributed four systems in the United States and two systems in Germany. Physician practices pay a one-time fee of about $7,500 to lease MelaFind and undergo company training. There is a fee of $50 per procedure to unlock the system for each use.

Diffusion is likely to be slow given the competition from numerous available dermoscopic devices and lack of reimbursement at the time of the product launch. Dermatologists may be reluctant to adopt this technology given the added expense and the fact that most third-party payers currently reimburse all or part of the expense of performing biopsies for clinically atypical lesions.

Coverage

In the United States, the Centers for Medicare & Medicaid Services (CMS) does not have a national coverage policy for multispectral analysis of skin. Thus, coverage for use of the MelaFind device to assess suspected skin lesions is at the discretion of local Medicare contractors.

ECRI Institute routinely searches 11 major private, third-party payers that publish their policies online (i.e., Aetna, Anthem, Blue Cross/Blue Shield Alabama, Blue Cross/Blue Shield Massachusetts, CIGNA, HealthPartners, Humana, Medica, Regence, Wellmark, UnitedHealthcare). Many private payers have not updated their coverage policies since FDA approval of the MelaFind device in November 2011. However, coverage determination for multispectral analysis using this device will likely be included in policies that describe dermoscopy in general. Payers that decide to reimburse for this test are likely to restrict coverage to exams performed only by dermatologists.

Our searches found two major payers with policies that describe coverage for dermoscopy in general (i.e., Aetna, CIGNA), five payers with policies that deny coverage for dermoscopy (i.e., Anthem, BCBS Alabama, BCBS Massachusetts, Humana, Wellmark) and four payers without a specific policy (i.e., HealthPartners, Medica, Regence, UnitedHealthcare).

State of Evidence Base

Quantity of evidence: Low

The evidence base for this technology consists of two studies: a prospective, blinded trial that enrolled 1,383 patients with 1,831 pigmented lesions that had one or more clinical or historical characteristics of melanoma. The findings of MelaFind and dermatologists’ clinical judgment were compared to histopathology results and a blinded, comparison trial that evaluated the performance of 10 specialized clinicians in diagnosing 99 small pigmented skin lesions (diameter ≤6 mm) compared with using the MelaFind system.

Quality of evidence: Low

Weaknesses of the evidence base include the following: results may not be applicable to standard practice; the case-control study may overestimate the test’s accuracy; a considerable number of lesions in the study prepared for the U.S. Food and Drug Administration premarket approval submission were ineligible for analysis or not evaluable due to operator error. Also, as with many new technologies, potential conflict of interest exists as the MelaFind manufacturer funded all studies available for analysis.

Consistency of evidence: Low

Too few studies are available to evaluate the consistency of patient-oriented outcomes of interest among different studies.

Excerpted with permission from ECRI Institute’s database of Emerging Technology Evidence Reports. To download the full report, visit www.ecri.org/managedcare [1].

For inquiries about this report or membership in ECRI Institute’s Health Technology Assessment Information Service, send an e-mail to htais@ecri.org [2].