AI tool may support dermatologists in assessing skin lesions for cancer risk
Mayo Clinic researchers have found that an artificial intelligence (AI) tool may provide dermatologists with additional information to help assess whether skin lesions are cancerous and inform their clinical decisions about whether to biopsy or monitor them, according to a study published in Mayo Clinic Proceedings.
Researchers from Mayo Clinic have discovered that an artificial intelligence (AI) tool called Dermoscopy Explainable Intelligence (DEXI) may assist dermatologists in determining the potential cancer risk of skin lesions. The tool, which is part of an advanced imaging system for capturing and tracking skin lesions, assigns a score between 0.0 to 10.0 based on various image features.
These features include asymmetry, border, and color. When a DEXI score is 3.0 or higher, it indicates a potential risk for skin cancer. The study published in Mayo Clinic Proceedings evaluated DEXI's accuracy in identifying cancerous and noncancerous lesions. With a threshold of 3.0, DEXI correctly identified 93.4% of cancerous lesions and 75% of noncancerous lesions.
Dermatologists previously reviewed patient DEXI scores but were uncertain about the implications of the scores. Now, the scores can serve as an additional data point to inform clinical decisions about whether to perform a biopsy or monitor the lesion. The AI tool, trained on over 66,000 annotated dermoscopy images, provides explainable AI by identifying which image features contributed to its cancer risk score.
The retrospective study included 178 patients who underwent total-body photography, digital dermoscopy imaging, and biopsy of one or more skin lesions. Researchers analyzed 261 lesions and compared DEXI scores with tissue examination diagnoses. Noncancerous lesions had an average DEXI score of 2.5, while nonmelanoma skin cancers had an average score of 7.8, and melanomas had a score of 6.0.
A DEXI score of 3.0 was identified as a potential threshold to help clinicians decide between monitoring or biopsy. However, the study also highlighted the limitations of DEXI. The tool does not consider patient-specific factors such as age, medications, and history of melanoma. Additionally, DEXI was less effective in evaluating severely atypical nevi and early evolving melanomas.
Therefore, the study emphasizes that DEXI should not be used in isolation but rather in conjunction with clinical findings and patient-specific factors. The AI tool requires further validation and refinement to enhance its accuracy and usefulness in dermatology.
Written by urgent.news from Medical Xpress's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
