Wilhelm Konrad Roentgen, Professor of Physics in Worzburg, Bavaria discovered X-Rays in 1895 by observing and deducing an accidental exposure of energy from his early design cathode ray tube onto a photographic plate. The first X-Ray was of his wife’s hand, shown below. X-Rays are one of the earliest great discoveries of the post-Renaissance age, even before E=mc2. Radiological Society of North America (RSNA) has been the definitive gathering place for the future of Radiology and Healthcare technology for as long as I can remember. X-Rays and its cousin spectra drive most of the new innovations in instrumentation, process and informatics.
The Western Roentgen Society, a predecessor of the RSNA, was founded in 1915 in St. Louis, Missouri. RSNA celebrated its centennial last year in Chicago (the anchor city for the conference for a long time). An interactive timeline of RSNA and Radiology events can be seen here.
I broke my almost 18-year attendance hiatus after my RSNA Associate membership acceptance this year; I started my career building X-Ray machines many, many moons ago and have worked in most Radiology modalities.
The scale of the conference was impressive, as has always been. The Technical and Exhibition Hall was massive at the McCormick Place Conference Center in Chicago. With about 670 exhibitors (105 new exhibitors) and the “who’s who” anchors like Bayer, Canon, CareStream, FUJIFILM, GE, Hitachi, Hologic, McKesson, Philips, Samsung, Shimadzu, Siemens, Terarecon and Toshiba this year’s technology innovation highlights were:
- GE 1.5 Tesla and 3 Tesla MRI (Magnetic Resonance Imaging) instrument with Total Digital Imaging (TDI) as well as CardioVascular Ultrasound systems with HDlive.
- Siemens 3D Advanced Visualization software, Cloud-based imaging network and xSPECT (Single Proton Emission Computed Tomography) for bone scans along with the combination of MRI-PET and PET-CT modalities.
- Virtual Reality (True3D), 3D printing, Human Connectome, Machine Learning and Deep Learning.
- RSNA Image Share, a Provider and Patient service.
- The maturing of Vendor Neutral Archives (VNA).
On Monday November 30th, the “New Horizons Lecture: Redefining Innovation” was delivered by Jeffrey R. Immelt, Chairman and CEO of GE. Mr. Immelt made the point that GE was both in the instrumentation innovator (US$20B) and payor (US$2.5B) revenue streams in healthcare. He emphasized that improving the ecosystem (consumerism + access, chronic disease outcomes, lower cost and behavior changes) as well as sustaining innovation (neural MRI, decision support, image guided interventions, automated image analysis and productivity) were its guiding principles. Precision Medicine, integration of Radiology with Pathology, cell therapy using Bioprocessing, mobile technologies at global scale and analytics were the central innovation themes for GE.
On Tuesday, December 1st, Dr. James H. Thrall, Chairman Emeritus, Department of Radiology at Massachusetts General Hospital, delivered the “Annual Oration in Diagnostic Radiology: Trends and Developments Shaping the Future of Radiology”. He outlined three themes: imaging technologies, infrastructure and information/communications systems, and the application of the imaging correlates of precision medicine. Dr. Thrall presented a Venn diagram of all imaging modalities. The various inter-modal intersection sets were highlighted with specific mention of PET-CT-MRI and the work of Dr. Ge Wang and Omni-Tomography was highlighted as shown in the figure below:
Of particular note to me was the official entry of “Precision Medicine” into the RSNA lexicon. This is the first year I have heard of the term “RadioGenomics” and “RadiOmics” in a major conference (first mentioned by Andreassen et al in 2002). Dr. Thrall made it a point to mention shorter acquisition times and lower radiation dosage to the patient.
Dr. Ronald Arenson, RSNA President, introduced both plenary speakers.
My focus for the 2015 Academic sessions was Informatics. Here are the condensed highlights:
Dr. Charles Kahn (U Penn) and Dr. Bradley Erickson (Mayo) led the inaugural “Year in Review” for Imaging Informatics. This session was jointly sponsored by RSNA, AMIA (American Medical Informatics Association) and SIIM (Society for Imaging Informatics in Medicine). The format was a dense, rapid-fire summary of key topics for 2015 as well as the seminal journal articles on various topics, which were the foci of other informatics sessions:
- Health Policy (Medicare Imaging Demonstration Project)
- Ontologies (RadLex Playbook)
- Standards and Interoperability (Management of Radiology Report Templates)
- Reporting (RSNA RadReport, Structured Reporting)
- Natural Language Processing (Information extraction from reports: e.g.: VTE)
- Workflow (MIRMAID, RadPath)
- RadioGenomics (Breast Cancer, GBA)
- Deep Learning (Brain Tumor Segmentation, Image Segmentation)
“The Text Information Extraction for Radiology Reporting” session presented techniques using NLP, Machine Learning and Deep Learning. A majority of radiologists would like to see structured reporting. The tools mentioned were OpenNLP, Mallet, cTAKES, eHOST, VINCI ChartReview and NCBO Annotator.
One of the more useful “hands-on” sessions that I attended was “Radio-Genomic Research: Accessing Clinical Imaging-Genomics-Pathology Data from Public Archives-The Cancer Imaging Archive” led by Dr. C. Carl Jaffe and Dr. Fred W. Prior. The Cancer Genome Atlas (TCGA) data portal is well known in the Genomics. The NIH has now created The Cancer Image Archive (TCIA) which has specific deidentified images for integrating Radiology with Genomics.
There was an entire morning session devoted to “Digital Information Security and Medical Imaging Equipment” which covered the instrumentation layers, protocols and regulations in some detail. It is interesting to note that Radiology Imaging client applications are still using OSx as the primary platform, with some “hands-on” sessions for DICOM not using other OSes at all. It is time for a web service based imaging application to come to the fore.
I strongly believe that integrating (and interoperating with) Radiology and Pathology phenotypic moieties into Genomics knowledge will be the real catalyst for the adoption of Genomics as an early clinical test (which is getting more complex by the month). Oh, and let’s not forget Proteomics. If multi-modal Radiology becomes reality soon (especially PET-CT-MRI), using biomarker-guided imaging, the data generated and the analytics required both grow exponentially. We are getting to that unified “Healthcare Data Lake” as shown below:
RSNA is now one of the top two academic conferences in the United States (and maybe worldwide) with a 2014 attendance of about 57,000. The 2015 attendance dropped, with registered attendants numbering about 48,000 (as of this writing). Here is hoping for more radiologists, technologists, innovators and patient advocates for this year’s RSNA!
Stay warm and healthy!
Author’s notes: The opinions expressed herein are my own and not necessarily those of EMC. Hyperlinks are embedded within specific words or phrases. Please contact me if you need details on any of the above topics.