MIT Scientists Developed AI to Detect Melanoma

Scientists and researchers from the Massachusetts Institute of Technology (MIT) have developed an artificial intelligence tool that can detect melanoma, a type of skin cancer. Following this technology, physicians are able to detect suspicious pigmented lesions (SPLs) through wide-field photography.

Melanoma is a type of skin-related cancer that starts manifesting in cells called melanocytes, states the Tech Explorist. According to MIT News, this malignant tumor accounts for over 70% of deaths related to skin cancer around the world.

To identify signs and symptoms of melanoma, physicians have long relied on visual cues and inspection to detect suspicious pigmented lesions, most can be indicative of skin cancer. MIT News states that the early detection of  SPLs can improve melanoma prognosis and can aid in alleviating treatment costs.

Scientists Developed AI to Detect Melanoma

With most physicians relying on visual cues and indicators in SPLs, the challenge lies not only in detecting these early but also in addressing the high volume of suspicious pigmented lesions found in the body for biopsy.

Through the artificial intelligence technology pipeline developed by the scientists and researchers from MIT and elsewhere, physicians and other medical sectors can now easily detect the presence of suspicious pigmented lesions via wide-field photography, notes Tech Explorist.

This technology is made possible by the use of deep convolutional neural networks (DCNNs) and the use of wide-field photography. The latter is reportedly found in mobile devices and other personal cameras notes MIT News.

Touted as neural networks, DCNNs work by determining and clustering images together. DCNNs leverage machine learning algorithms to work. MIT News states that by combining wide-field photography with DCNNs, artificial intelligence can easily identify SPLs and skin lesions.

The system has since been tested with 133 patients at the Hospital Gregorio Maranon in Madrid, Spain. The system resulted in 90.3% sensitivity in detecting SPLs and differentiating this from nonsuspicious skin lesions, reports MIT News.

Post-doctoral and medical device expert, as well as MIT’s Venture Builder, Luis R. Soenksen said that the use of these cameras can help in the early detection and mitigation of skin cancer.

In a statement made by Soenksen, he said, “Our research suggests that systems leveraging computer vision and deep neural networks, quantifying such common signs, can achieve comparable accuracy to expert dermatologists,” Soenksen explains. “We hope our research revitalizes the desire to deliver more efficient dermatological screenings in primary care settings to drive adequate referrals.”

Following the development, the team over at MIT News aims to try its system with different types of cameras. Moreover, the scientists over at MIT will also look into working with various light settings and images taken by different people and photographers, states AZO Robotics.