@article{Shafiee_Wong_2017, title={Discovery Radiomics via Deep Multi-Column Radiomic Sequencers for Skin Cancer Detection}, volume={3}, url={https://openjournals.uwaterloo.ca/index.php/vsl/article/view/177}, DOI={10.15353/vsnl.v3i1.177}, abstractNote={<p>While skin cancer is the most diagnosed form of cancer in men<br />and women, with more cases diagnosed each year than all other<br />cancers combined, sufficiently early diagnosis results in very good<br />prognosis and as such makes early detection crucial. While radiomics<br />have shown considerable promise as a powerful diagnostic<br />tool for significantly improving oncological diagnostic accuracy and<br />efficiency, current radiomics-driven methods have largely rely on<br />pre-defined, hand-crafted quantitative features, which can greatly<br />limit the ability to fully characterize unique cancer phenotype that<br />distinguish it from healthy tissue. Recently, the notion of discovery<br />radiomics was introduced, where a large amount of custom, quantitative<br />radiomic features are directly discovered from the wealth of<br />readily available medical imaging data. In this study, we present<br />a novel discovery radiomics framework for skin cancer detection,<br />where we leverage novel deep multi-column radiomic sequencers<br />for high-throughput discovery and extraction of a large amount of<br />custom radiomic features tailored for characterizing unique skin<br />cancer tissue phenotype. The discovered radiomic sequencer was<br />tested against 9,152 biopsy-proven clinical images comprising of<br />different skin cancers such as melanoma and basal cell carcinoma,<br />and demonstrated sensitivity and specificity of 91% and 75%, respectively,<br />thus achieving dermatologist-level performance and<br />hence can be a powerful tool for assisting general practitioners<br />and dermatologists alike in improving the efficiency, consistency,<br />and accuracy of skin cancer diagnosis.</p>}, number={1}, journal={Journal of Computational Vision and Imaging Systems}, author={Shafiee, Mohammad Javad and Wong, Alexander}, year={2017}, month={Oct.} }