Health

Skin cancer app fails to identify rare, aggressive cancers

(HealthDay)-The consumer machine learning model for detecting skin cancer is effectively September 29-October 2.

Lloyd Steele, MB Ch.B, University of London. We evaluated the performance of machine learning models of Merkel cell carcinoma (MCC) and amelanotic melanoma.Performance of direct sales models available in Europe as certified Medical equipmentWas evaluated using a series of images containing 28 MCCs, 35 amelanotic melanomas, 28 seborrheic keratosis, and 25 hemangiomas.

Researchers found that consumer apps misclassified 5 of 28 MCCs (17.9%) and 7 of 35 amelanotic melanomas as low risk. Almost two-thirds (62.2 percent) of benign lesions were classified as high risk. The sensitivity of the model for detecting malignant tumors was 79.4% and the specificity was 37.7%.

“Machine learning to improve model “The assessment needs to take into account the extent of the actual disease,” Steele said in a statement. Skin cancer.. ”


Pale melanoma masked by the albino gene


For more information:
For more information

Copyright © 2021 HealthDay.. all rights reserved.

Quote: Skin Cancer App is a rare and aggressive cancer obtained from https://medicalxpress.com/news/2021-10-skin-cancer-app-rare-aggressive.html on October 14, 2021. Unable to identify (October 14, 2021)

This document is subject to copyright. No part may be reproduced without written permission, except for fair transactions for personal investigation or research purposes. The content is provided for informational purposes only.



Skin cancer app fails to identify rare, aggressive cancers Source link Skin cancer app fails to identify rare, aggressive cancers

Related Articles

Back to top button