Skin cancer is one of the most common types of cancer in humans. Traditional skin cancer diagnostic methods are costly, require a professional physician, and take time. Therefore, AI tools such as shallow and deep machine learning are used to help diagnose skin cancer. Skin cancer is detected using computer algorithms and deep neural networks.
According to MIT News, melanoma is a type of malignant tumor responsible for more than 70 percent of skin cancer deaths worldwide. For about four years, doctors have relied on visual examination to identify suspicious pigmented lesions (SPLs) indicative of skin cancer. Such early identification of SPLs in primary care can improve melanoma prognosis and significantly reduce treatment costs.
Due to the large amount of pigmented lesions that often need to be examined for possible biopsies, the challenge is that it is difficult to find SPLs quickly. Researchers from MIT and elsewhere have developed a new AI pipeline that takes deep convolutional neural networks and applies them to analyze SPLs through wide-field photography, common in most smartphones and personal cameras.
DCNNs are neural networks that can be used to classify (or “name”) images in order to then group them (e.g. when performing a photo search). These ML algorithms belong to the subset of deep learning.
Cameras are used to capture large area images of large areas of a patient’s body. The program uses DCNNs to quickly and effectively identify and screen early-stage melanoma, according to Luis R. Soenksen, a postdoc and medical device expert who currently serves as MIT’s first venture builder in the AI and healthcare space.
Soeksen conducted the research with MIT researchers, including MIT Institute for Medical Engineering and Science (IMES) faculty members Martha J. Gray, W. Kieckefer Professor of Health Sciences and Technology, Professor of Electrical Engineering and Computer Science; and James J. Collins, Termeer Professor of Medical Engineering and Science and Bioengineering.
Soeksen explained in a recent article that early detection of SPLs can save lives; However, the current capacity of medical systems to provide comprehensive skin screening for anxiety is still insufficient.
The paper describes the development of an SPL analysis system that uses DCNNs to more quickly and efficiently identify skin lesions that require more investigation and screening than routine primary care visits or by the patient themselves. The system used DCNNs to facilitate identification and Optimize classification of SPLs in widefield images.
Result of using AI tools
Using AI, the researchers trained the system using 20,388 wide-field images of 133 patients. The photos were taken with a variety of common cameras readily available to consumers. Dermatologists who worked with the researchers visually classified the lesions in the images for comparison. They found that the system achieved greater than 90.3 percent sensitivity in distinguishing SPLs from non-suspect lesions, skin and complex backgrounds by avoiding the need for time-consuming imaging of individual lesions.
Research suggests the system uses computer vision and deep neural networks. Quantifying such common signs can achieve comparable accuracy to experienced dermatologists. The study is expected to provide faster and more accurate assessment of SPLS and could lead to earlier treatment of melanoma.