MIAMI -- A computer-automated device detected melanoma and high-grade dysplastic nevi lesions with 98% sensitivity in a large, prospective, multicenter study.
The digital dermoscopy device "sees" at different skin depths using 10 spectral bands ranging from 430 nm to 950 nm. The technology, which was more than a decade in development, objectively assesses up to 75 factors to differentiate melanoma from low-grade and high-grade dysplastic nevi, Dr. Gary D. Monheit said.
Dr. Monheit was an investigator at one of seven sites that used the MelaFind (Electro-Optical Sciences) to assess a total of 1,632 evaluable lesions. Lesions included 70 invasive melanomas and 57 melanoma in situ. "This is the largest prospective, blinded study ever conducted in melanoma detection," he noted.
The noninvasive device has a handheld wand for image capture at the point of care. Because of these features and a proprietary algorithm that analyzes multiple spectrums, it is "totally objective with a yes or no algorithm for excision," said Dr. Monheit, a private practice dermatologist in Birmingham, Ala., and an associate clinical professor in the departments of dermatology and ophthalmology at the University of Alabama at Birmingham. Detection is automatic with immediate feedback, Dr. Monheit said. "If we don't get a clear image, the machine tells us the image is not possible." The technology was developed after assessment of more than 10,000 pigmented lesions from more than 7,000 patients.
The aim of this "pivotal study" was to establish safety of the device and sensitivity for melanoma detection. No adverse events were reported, Dr. Monheit said. Lesions had to be pigmented with melanin, keratin, and/or blood. Clinical management was biopsy, he added.
For melanoma and high-grade dysplastic nevi, the device had a 98% sensitivity. "These are the [lesions] you will want to take out," Dr. Monheit said.
"At same time we should look at specificity--we do not want to biopsy every lesion that comes into our office," Dr. Monheit said. The specificity of the device was 9.4%, statistically superior to the dermatologist evaluations at 3.7%.
The results with the device were compared with prebiopsy investigator diagnoses and with an objective pathologic review of lesions by a panel of three dermatopathologists. If two of the three dermatopathologists concurred on the diagnosis, their consensus was final. The device had a 98% sensitivity for biopsy detection.
Dr. Monheit and the other study investigators also collected patient data, including age, gender, ethnicity, patient in-house or referred, and any risk factors for melanoma. Anatomic locations of the lesions were also noted.
This study provides "evidence for safety and efficacy for aid in evaluating pigmented lesions," Dr. Monheit said.
Disclosures: Dr. Monheit is a consultant and researcher for Electro-Optical Sciences.