https://www.qradiomics.com/tags/computer-aided-diagnosis/ Recent content in Computer-Aided-Diagnosis on Qualia Radiomics Hugo en-us Wed, 29 Jun 2022 21:01:32 -0400 https://www.qradiomics.com/posts/2022-06-29-clinically-interpretable-radiomics/ Wed, 29 Jun 2022 21:01:32 -0400 https://www.qradiomics.com/posts/2022-06-29-clinically-interpretable-radiomics/ <p><a href="https://arxiv.org/pdf/2206.14903.pdf">MICCAI'22 Paper</a> | <a href="https://arxiv.org/pdf/1808.08307.pdf">CMPB'21 Paper</a> | <a href="https://zenodo.org/record/6762573">CIRDataset</a></p> <p>This library serves as a one-stop solution for analyzing datasets using clinically-interpretable radiomics (CIR) in cancer imaging (<a href="https://github.com/choilab-jefferson/CIR">https://github.com/choilab-jefferson/CIR</a>). The primary motivation for this comes from our collaborators in radiology and radiation oncology inquiring about the importance of clinically-reported features in state-of-the-art deep learning malignancy/recurrence/treatment response prediction algorithms. Previous methods have performed such prediction tasks but without robust attribution to any clinically reported/actionable features (see extensive literature on the sensitivity of attribution methods to hyperparameters). This motivated us to curate datasets by annotating clinically-reported features at the voxel/vertex level on public datasets (using our published <a href="https://github.com/taznux/LungCancerScreeningRadiomics">advanced mathematical algorithms</a>) and relating these to prediction tasks (bypassing the “flaky” attribution schemes). With the release of these comprehensively-annotated datasets, we hope that previous malignancy prediction methods can also validate their explanations and provide clinically-actionable insights. We also provide strong end-to-end baselines for extracting these hard-to-compute clinically-reported features and using these in different prediction tasks.</p>