Supriya Gupta MBBS, MD, Presenter: Nothing to Disclose

Priyanka Prakash MBBS, Abstract Co-Author: Nothing to Disclose

Thomas J. Schultz BS, Abstract Co-Author: Nothing to Disclose

Daniel Ira Rosenthal MD, Abstract Co-Author: Nothing to Disclose

Keith J. Dreyer DO, PhD, Abstract Co-Author: Medical Advisor, General Electric Company Medical Advisor, Siemens AG Medical Advisor, Nuance Communications, Inc Medical Advisor, Carestream Health, Inc Medical Advisor, Vital Images, Inc Medical Advisor, Amirsys, Inc Medical Advisor, Life Image Inc Medical Advisor, McKesson Corporation

James H. Thrall MD, Abstract Co-Author: Board Member, Mobile Aspects, Inc Medical Advisor, Genelux Corporation Stockholder, General Electric Company Stockholder, Apple Inc Stockholder, Achillon Pharmaceuticals, Inc Stockholder, Ferrumax Pharmaceuticals Inc

Emphasis on quality and appropriateness is gaining momentum for reporting radiology reports. To evaluate the effectiveness of RADLEX-based Intelligent Interactive data Mining System (RIIMS) to track the nature of pathology findings in Important Finding Alerts (IFA) sent by radiologists to alert the referring physicians for an imaging finding.Insight into trends of sending alerts to physicians for certain findings can raise awareness about patterns of pathology, which need urgent physician’s attention.

Radiology reports were imported into RIIMS from Radiology Information systems (RIS) and various demographic data, exam codes, date of exam, ordering physicians names were recorded. We identified some common pathology terms for Brain (Br, Lung (Lu), Liver (Li), kidney (Ki) and Heart (Hr) using RADLEX ontology trees and reports classified as IFA present (IFA+) and IFA not present (IFA-). The most common pathology (MCP) for sending IFA amongst each organ was found.

Of the 1721531 cases imported into RIIMS from Radiology Information Systems (RIS), IFA+ were 82884 cases (4.81%). We evaluated 224,706 exams for 5 major organs for which the average IFA+ was 11.54% (BR-4.15%,LU-14.07%,LI-14.08%,KI-12.23%,HR-9.28%). MCP associated with Br IFA+ exams were- mass (31.99%), hemorrhage (28.94%), ischemia (23.91%) , aneurysm (17.3%); MCP for Lu IFA+ were-nodule (79.4%),mass (24.4%),effusion (15.04%);MCP for Li IFA+ were-metastasis (37.5%), mass (32.65),hemangioma (14.06%), MCP for Ki IFA+ were-cysts (69.72%),mass (33.96%),carcinoma (18.05%);MCP for Hr IFA+ were-ischemia(30.15%),mass (25.4%),infection (24.76%).

The most common IFA overall are pulmonary nodules followed by mass. For brain , lung, liver, lidney and heart, IFAs are most often sent for mass, pulmonary nodules,metastasis, renal cysts and ischemia, respectively. Clinical significance of these findings can be gauged by assessing the distribution patterns of pathology for IFA. Guidelines for tracking such alerts can be standardized for a more uniform practice based on scientific data.

The spectrum of trends could landscape disease burden in a given setting besides providing useful feedback for improving imaging apractice and set the platform for creating guidelines for reporting.

Gupta, S, Prakash, P, Schultz, T, Rosenthal, D, Dreyer, K, Thrall, J, To Alert or Not: Tracking Pathology Trends in Important Findings Alerts in Radiology for Improving Medical Practice Using Ontology-based Language Processor.  Radiological Society of North America 2010 Scientific Assembly and Annual Meeting, November 28 - December 3, 2010 ,Chicago IL. http://archive.rsna.org/2010/9009748.html