Melanoma thickness is equally difficult to judge for algorithms and dermatologists

Assessing melanoma thickness is difficult, whether performed by an experienced dermatologist or a well-trained machine learning algorithm. A study from the University of Gothenburg shows that the algorithm and dermatologists had an equal success rate in interpreting dermoscopic images.

When diagnosing melanoma, dermatologists assess whether it is an aggressive form (“invasive melanoma”), where cancer cells are growing in the dermis and there is a risk of spreading to other parts of the skin. body, or of a more benign form (“melanoma on the spot,” MIS) that grows only in the outer layer of the skin, the epidermis. Invasive melanomas that grow deeper than a millimeter into the skin are considered thick and, as such, more aggressive.

Importance of thickness

Melanomas are assessed by investigation with a dermatoscope – a type of magnifying glass fitted with a bright light. Diagnosing melanoma is often relatively straightforward, but estimating its thickness is a much greater challenge.

“In addition to providing valuable prognostic information, thickness can affect the choice of surgical margins for the first operation and how quickly it should be performed,” says Sam Polesie, associate professor (docent) of dermatology and venereology at the Sahlgrenska Academy of the University. from Gothenburg, Polesie is also a dermatologist at Sahlgrenska University Hospital and first author of the study.

Link between man and machine

Using a web-based platform, 438 international dermatologists evaluated nearly 1,500 melanoma images captured with a dermatoscope. The dermatologists’ results were then compared to those of a machine learning algorithm trained to classify the depth of melanoma.

Among dermatologists, the overall accuracy was 63% for correctly classifying MIS and 71% for invasive melanoma.

“Interestingly, work history and dermoscopy experience had no bearing on diagnostic accuracy in predicting melanoma thickness.

The area under the curve, which is a 0 to 1 measure of performance, was 0.83 for the pre-trained machine learning algorithm and 0.85 for the combined AUC of the individual drives. Collectively, the dermatologists’ assessment was on par with an algorithm trained to distinguish between MIS and invasive melanoma,” says Polesie.

Difficult to assess

Artificial intelligence (AI) is making great strides in healthcare. This technology should in particular be able to be developed as a support for medical imaging, ie for doctors who evaluate and interpret images, such as x-rays and images of the retina and skin changes. The technology is also applicable to areas other than image recognition.

“Our study highlights the difficulties of correctly assessing melanoma thickness based on dermoscopic images,” adds Polesie.

“In future studies, we aim to explore the usefulness of predefined dermoscopic structures for distinguishing. We also want to test whether clinical decision-making in this situation can be improved by means of machine learning algorithms.”

The results are published in the Journal of the European Academy of Dermatology and Venereology, JEADV. The study was conducted in collaboration with researchers from the Medical University of Vienna, Austria.

Source of the story:

Materials provided by University of Gothenburg. Note: Content may be edited for style and length.

Sharon D. Cole