RSNA 2014 

Abstract Archives of the RSNA, 2014


Story of Stickr - Design and Usage of an Automated Biopsy Follow Up Tool

Scientific Papers

Presented on December 3, 2014
Presented as part of SSK12: ISP: Informatics (Quality and Safety)


Marc D. Kohli MD, Abstract Co-Author: Research Grant, Koninklijke Philips NV Research Grant, Siemens AG
Aaron P. Kamer MD, Presenter: Nothing to Disclose


Mammographers are legally required to evaluate pathology from each biopsy in order to determine concordance. Many other sub-specialist radiologists find large-scale followup challenging due to task complexity. We set out to design and implement a web-based biopsy follow up worklist application. Important quality metrics such as adequacy rates and diagnostic rates would also be be calculated from data collected. 


Prior to implementation of the worklist, radiology faculty who regularly perform biopsies were surveyed about their biopsy practices. Our application was built to receive biopsy reports and pathology reports in real-time from HL7 feeds. Each radiology report is processed to assign a radiologist and a resident (if applicable).  Upon logging in, the faculty or resident is presented with a list of biopsies performed. The biopsies that have associated pathology reports are highlighted. With just two clicks, a biopsy can be marked as adequate/concordant. If biopsies are flagged as inadequate/discordant, an option to visit the hospital paging webpage is presented. 


Of the 21 faculty survey respondents (with 8 mammographers), only 43% follow up the pathology results every time. 3 faculty (14%) follow up on their biopsies up to 20% of the time. Over 1300 image-guided biopsy reports have entered the successfully deployed application, with 82% of these reports having been linked with respective pathology, a rate much higher than before discarding a body part matching requirement between reports. The participating physicians have noted concordance/discordance in 23% of biopsies that have pathology. 


Radiologists, particularly mammographers, have a high rate of biopsy follow up. Many other faculty do not as reliably follow up on their pathology results, instead depending on the referring clinician to determine repeat biopsy necessity. Use of NLP for body part matching in biopsy/pathology reports results in a low number of report matching, but reports matched using only time and patient ID number criteria results in a high number of reports delivered. A biopsy-pathology follow up worklist can be well-integrated into current radiology practice systems. 


By automatically populating a web-based worklist with radiology and pathology reports, an otherwise time consuming and tedious task can be educational and add value to patient care.

Cite This Abstract

Kohli, M, Kamer, A, Story of Stickr - Design and Usage of an Automated Biopsy Follow Up Tool.  Radiological Society of North America 2014 Scientific Assembly and Annual Meeting, - ,Chicago IL.