Imagine 2016 - Workshop Abstract - Radiology and Pathology are Converging: The New Diagnostics

From the time of Virchow, histopathologic classification has been the backbone of classification of medical diseases and the ultimate tool in diagnosis. Radiology, a relative newcomer to the diagnostic scene, has in the past, also provided important information to facilitate a diagnosis and management.  In the era of digital information, these two disparate fields are rapidly coming together in a more integrated fashion.

Pathology and radiology both provide phenotypic evidence require for diagnosis and patient management. Reporting and anatomic pathology and radiology are both based on the analysis and interpretation of data derived from the review of images. Traditionally narrative reports have been used. Moreover, the workflow models are somewhat similar. While traditionally, the “scale” of the information provided and interpreted has been different, increasing awareness of the importance systems biology and disease heterogeneity has in many cases made these two disciplines fundamentally complimentary in diagnosis, management decisions, teaching and research.  As a result, efforts are underway to promote the integration of this information and in the broader sense, integrating pathologic, radiologic, genomic, proteomic and metabolomic information on both an individual patient level and on a population level.  

Currently, efforts are underway to converge the data from these two disciplines from an informatics and IT perspectives. These include homogenization of image archives and data services, the use of multimedia electronic health records and decision support systems.

Integrated reporting is currently available at the UCLA Medical Center.  Moreover, at the University of Rotterdam, the departments of radiology and pathology have been combined in division of integrated diagnostics.  The paradigm for future diagnostics includes the convergence of large diverse scale fused non-linear data streams vetted by domain knowledge expertise which are then integrated into disease and outcomes categories.  This approach, of course, will enable improved use of “big” “real world” data to better understand disease and outcomes.