作者: Martina Sollini , Luca Cozzi , Gaia Ninatti , Lidija Antunovic , Lara Cavinato
DOI: 10.1016/J.YMETH.2020.01.007
关键词: Computed tomography 、 PET-CT 、 Outcome prediction 、 Medical physics 、 Radiomics 、 Medline database 、 Breast cancer 、 Generalizability theory 、 Medicine 、 Positron emission tomography
摘要: The aim of the present review was to assess current status positron emission tomography/computed tomography (PET/CT) radiomics research in breast cancer, and particular analyze strengths weaknesses published papers order identify challenges suggest possible solutions future directions. Various combinations terms "breast", "radiomic", "PET", "radiomics", "texture", "textural" were used for literature search, extended until 8 July 2019, within PubMed/MEDLINE database. Twenty-six articles fulfilling inclusion/exclusion criteria retrieved full text analyzed. studies had technical clinical objectives, including diagnosis, biological characterization (correlation with histology, molecular subtypes IHC marker expression), prediction response neoadjuvant chemotherapy, staging, outcome prediction. We reviewed discussed selected investigations following workflow steps related clinical, technical, analysis, reporting issues. Most evidence on role PET/CT cancer is at feasibility level. Harmonized methods image acquisition, post-processing features calculation, predictive models classifiers trained validated sufficiently representative datasets, adherence consensus guidelines, transparent will give validity generalizability results.