Rxivist logo

Rxivist combines preprints from bioRxiv with data from Twitter to help you find the papers being discussed in your field. Currently indexing 70,138 bioRxiv papers from 306,359 authors.

A Deep Learning Approach for Rapid Mutational Screening in Melanoma

By Randie H. Kim, Sofia Nomikou, Zarmeena Dawood, George Jour, Douglas Donnelly, Una Moran, Jeffrey S Weber, Narges Razavian, Matija Snuderl, Richard Shapiro, Russell S Berman, Nicolas Coudray, Iman Osman, Aristotelis Tsirigos

Posted 16 Apr 2019
bioRxiv DOI: 10.1101/610311

DNA-based molecular assays for determining mutational status in melanomas are time-consuming and costly. As an alternative, we applied a deep convolutional neural network (CNN) to histopathology images of tumors from 257 melanoma patients and developed a fully automated model that first selects for tumor-rich areas (Area under the curve AUC=0.98), and second, predicts for the presence of mutated BRAF or NRAS. Network performance was enhanced on BRAF-mutated melanomas <=1.0 mm (AUC=0.83) and on non-ulcerated NRAS-mutated melanomas (AUC=0.92). Applying our models to histological images of primary melanomas from The Cancer Genome Atlas database also demonstrated improved performances on thinner BRAF-mutated melanomas and non-ulcerated NRAS-mutated melanomas. We propose that deep learning-based analysis of histological images has the potential to become integrated into clinical decision making for the rapid detection of mutations of interest in melanoma.

Download data

  • Downloaded 419 times
  • Download rankings, all-time:
    • Site-wide: 27,485 out of 70,053
    • In pathology: 85 out of 371
  • Year to date:
    • Site-wide: 25,084 out of 70,053
  • Since beginning of last month:
    • Site-wide: 10,333 out of 70,053

Altmetric data

Downloads over time

Distribution of downloads per paper, site-wide


Sign up for the Rxivist weekly newsletter! (Click here for more details.)