JVIR
’
Scientific Session
Wednesday
was significantly higher in post-implementation period than that in pre-implementation period (Po0.001). Conclusions: Implementation of Mallampati score and ASA classification in the pre-procedural evaluation allows for identification of patients with potential cardiorespiratory complications who could benefit from the utilization of general anesthesia. An increase in utilization of anesthesia was identified as was a trend toward reduction in cardiorespiratory complications during IR procedures following implementation of this simple pre-procedural assessment.
3:09 PM
’
S153
existing classification systems were evaluated and compared statistically using chi-squared test, Mann-Whitney test, or Student’s T-test as appropriate. Overall comparison between the two surveys was performed using multivariate regression analysis employing F statistics. Results: The proposed AE classification system performed better (po0.05) than the existing system for accuracy and consistency of reporting and ability of integration into local QI programs. Additionally, the responders gave a higher overall rating to the educational and research value of the new system compared with the existing system (po0.05). Conclusions: The proposed AE classification system is superior to the existing SIR system in the domains evaluated.
Abstract No. 353
Proposal of a new adverse event classification system by the SIR Standards of Practice Committee
Purpose: To develop and evaluate a new adverse event (AE) classification for interventional radiology (IR), with comparison to the current Society of Interventional Radiology (SIR) AE classification. Materials: A survey was developed via 15 conference calls by a group of 17 members from the SIR Standards of Practice Committee and Service Lines. Twelve clinical AE case scenarios were generated that encompass a broad spectrum of procedural anatomy, technique and complexity in IR as well as all levels of potential adverse event severity and preventability. Survey questions were designed to evaluate practicality of AE system use, educational value, suitability to improve quality, accuracy, consistency and completeness of AE reporting, ability of integration into local quality improvement program and utility for scientific research. Survey participants were instructed to answer survey questions based on each AE scenario for the proposed and existing SIR classification systems. SIR members were invited between 10/15/2016 to 3/16/2016 via an online survey link, and 140 members participated. Answers on new and
Abstract No. 354
Utilization of deep learning techniques to assist clinicians in diagnostic and interventional radiology: development of a virtual radiology assistant K. Seals, B. Dubin, L. Leonards, E. Lee, J. McWilliams, S. Kee, R. Suh; David Geffen School of Medicine at UCLA, Los Angeles, CA Purpose: Improved artificial intelligence through deep learning has the potential to fundamentally transform our society, from automated image analysis to the creation of self-driving cars. We apply these techniques to create a virtual radiology assistant that offers clinicians many non-interpretive radiology skills. A wide range of functionality is offered, from helping select the optimal imaging study or procedure for a given clinical scenario to describing the follow-up of incidental findings, guiding contrast administration in renal failure/contrast allergy, and describing optimal peri-procedural patient management. Materials: The application was built in Xcode using the Swift programming language. The user interface consists of text boxes arranged in a manner simulating communication via traditional SMS text messaging services. Natural Language Processing (NLP) was implemented using the Watson Natural Language Classifier application program interface (API). Using this classifier, user inputs are understood and paired with relevant information categories of interest to clinicians. For example, if a clinician asks whether IVC filter placement is appropriate for a particular patient, they will be paired with an IVC filter category and relevant information will be provided. This information can come in multiple forms, including relevant websites, infographics, and subprograms within the application. Results: Model performance improves significantly as more training data is provided. Successful input categorization is performed in most cases, particularly for focused queries providing sufficient detail. Using a confidence probability threshold of 0.7, the natural language classifier provides optimal information categorization with minimal inclusion of extraneous categories. Conclusions: Deep learning techniques can be used to create powerful artificial intelligence tools to assist clinicians. These tools both allow clinicians to rapidly access useful information and reduce the need for radiologists to perform redundant communication tasks that potentially distract from patient care.
WEDNESDAY: Scientific Sessions
O. Khalilzadeh1, M. Baerlocher2, D. KATSARELIS3, P. Shyn4, A. Devane5, C. Morris6, A. Cohen7, B. Connolly8, M. Midia9, R. Thornton10, K. Gross11, D. Caplin12, G. Aeron13, S. Misra14, N. Patel15, T. Walker16, G. Martinez-Salazar17, J. Silberzweig1, B. Nikolic18; 1Icahn School of Medicine At Mount Sinai, New York, NY; 2 Royal Victoria Hospital, Barrie, ON; 3Society of interventional radiology, Fairfax, VA; 4Brigham and Women’s Hospital, Boston, MA; 5Greenville Health System University Medical Group, Greenville, SC; 6 University of Vermont Medical Center, Burlington, VT; 7 McGovern School of Medicine, University of Texas, Houston, TX; 8Hospital for Sick Children, Toronto, ON; 9 McMaster University, Hamilton, ON; 10Memorial Sloan Kettering Cancer Center, New York, NY; 11Greater Baltimore Medical, Owings Mills, MD; 12Hofstra Northwell School of Medicine, Manhasset, NY; 13 VA Medical Center, Providence, RI; 14Mayo Clinic and Foundation, Rochester, MN; 15University of Vermont Medical Center, Burr Ridge, IL; 16Massachusetts General Hospital, Boston, MA; 17Massachussetts General Hospital, Chestnut Hill, MA; 18Stratton Medical Center, Moorestown, NJ
3:18 PM