Functional imaging: translation of radiomics and artificial intelligence into clinical practice

Background: Radiomics and artificial intelligence (AI) are among the most important topics in radiologic research of recent years. However, translation from research to practice has so far only occurred sporadically.Objective: The current and future clinical benefits of radiomics and AI in oncology are critically examined in the present article.Materials and methodsA literature search was performed.Results: AI systems are already available for radiologic oncology practice and are increasingly used in diagnostics. However, systems that provide referrers with information for prognosis, treatment response, or progression are not yet approved and will not become available in the foreseeable future.Conclusion: Radiomics and AI already show practical value for improving radiologic diagnostic quality. However, to develop reliable systems that can provide quantitative statements useful for patient treatment, there is still a lack of larger validation studies and financial incentives for manufacturers.