AbstractPurposeThe global COVID-19 pandemic has accelerated the development of numerous digital technologies in medicine from telemedicine to remote monitoring. Concurrently, the pandemic has resulted in huge pressures on healthcare systems. Medical imaging from chest radiographs to computed tomography and ultrasound of the thorax have played an important role in the diagnosis and management of the coronavirus infection.MethodsWe undertook a systematic review of the literature focused on MI in COVID-19 and the utility of AI. Keyword searches were performed on PubMed and preprint servers including arXiv, bioRxiv and medRxiv; 338 papers were included in a meta-analysis and manually reviewed to assess solutions in AI according to their clinical relevance. The maturity of the papers was evaluated based on four criteria: peer-review, patient dataset size, algorithmic complexity and usage of the AI in clinical practice.ResultsIn the first three quarters of 2020, we identified 3444 papers on MI in COVID-19, of which 556 had at least some focus on AI. 2039 of 3444 were specific to imaging modalities and predominantly (80.7%) focused on CT (9.9% on LUS and 9.5% on CXR). The AI literature was predominantly focused on CXR (51.2%), 36.1% on CT and 1.8% on LUS. Only a small portion of the papers were judged as mature (3.8%) and most AI papers focused on disease detection (72.8%).ConclusionsThis review evidences a disparity between clinicians and the AI community, both in the focus on imaging modalities and performed tasks. Better collaboration is needed to allocate resources optimally for the development of clinically relevant solutions that are validated on large-scale patient data.Clinical implicationsAI may aid clinicians and radiologists by providing better tools for localization and quantification of disease features and changes thereof, and, with integration of clinical data, may provide better diagnostic performance and prognostic value.