Three speaking engagements in the AI space earlier in the year have brought into focus various aspects of AI in Radiology and Imaging.
The first, the ISMRM workshop in Quantitative Imaging in New Delhi reinforced the fact that Quantitative Imaging forms the basis for all AI. What cannot be quantified or objectivized cannot form the basis for analysis by a computer algorithm. This also holds true for ground truth. From a day to day practical standpoint for clinical radiologists, we are moving towards more structured and templated reports (BIRADS and PIRADS are two examples) which allow for greater predictability and standardization of the interpretation process and more clear communication with clinicians. This will eventually be a pan-radiology phenomenon.
The second, during my visit to the spectacular Institute of Genomics and Integrative Biology also in New Delhi, highlighted the importance of cross-specialization and cross-fusion in disciplines in scientific advance. In particular the burgeoning field of Radiogenomics represents the utilization of genetic data to enhance the specificity of radiologic diagnosis (as in tumor analysis) and allows for more personalized imaging. With the increasing complexity of the data to be analysed, AI presents itself as the best option to improve the accuracy of radiologic analysis, especially in relatively unusual and complex disease pathologies, such as genetic diseases and oncology.
And the last at the Healthcare Business International Conference in London, showed me how the world of finance is eagerly tracking the developments in the AI space and waiting to see which unicorns emerge from the large numbers of startups that are currently working assiduously to make an impact on AI in healthcare.
Speaking opportunities, apart from offering the chance to share one’s experiences and thoughts, and to network with others with similar or related interests, are also a great learning experience. As Stephen Covey recommends, it is particularly important that one should first “seek to understand, then to be understood”
Patient Ca*r
Dr Smith at Middlesex Hospital
I freakin love Teleradiology