There are many examples across different medical specialties where image-derived radiomic features improve the accuracy of predictive, descriptive, or prescriptive clinical decision support algorithms. On the other hand, clinical context, such as patient history, complaints, or laboratory results are known to improve the performance of screening or diagnostic imaging algorithms.
In both cases, we are leveraging complementary contextual image or EHR information to overcome the limitations of image-only or EHR-only models, respectively. Despite the similarities, there are significant differences in clinical value, algorithm type, and even deployment platform associated with each algorithm type.
In this talk, we briefly review the state-of-the-art in each type of algorithm, compare and contrast their current and potential future clinical value, and explore if they can be hosted in the same run-time platform.
Watch the recording from our CTO Sinan Batman's RSNA22 Innovation Theater Presentation titled, "Is Imaging AI aided by EHR data more valuable than Clinical Decision Support AI aided by radiomics?"Opt-in is required in order to watch this recording.