Type:
Educational Exhibit
Keywords:
Quality assurance, Cancer, Structured reporting, Screening, Outcomes analysis, RIS, MR, CT, Oncology, Computer applications
Authors:
D. J. Vining1, A. Pitici2, C. Popovici2, A. Prisacariu2, M. Kontak2; 1Houston, TX/US, 2Houston/US
DOI:
10.26044/ecr2019/C-3592
Findings and procedure details
We developed a structured reporting solution that works in the following manner:
- Records key images and a radiologist’s verbal descriptions of image findings (Figure 3).
- Tags the findings in a database with metadata describing the anatomical locations,
radiological diagnoses,
and common data elements using natural language processing (NLP) (Figure 4).
- Assembles a multimedia report (Figure 5).
When a particular finding is identified (e.g.,
lung nodule),
the system prompts the radiologist to dictate salient disease features using displays of common data elements (CDEs) (10). As the features are described by the radiologist,
the system populates the common data element template from which the appropriate RADS assessment score is automatically calculated and inserted into the report. If the radiologist fails to mention required features,
the system performs a compliance check and alerts the radiologist to complete the data entry either by dictating an additional description or by selecting the CDE elements manually from the displayed menu.
Upon complete data entry,
the system automatically calculates the appropriate RADS assessment based on the anatomy and common data elements for incorporation into a multimedia structured report along with any associated recommendations for further patient care.
The automatic calculation of the RADS score is accomplished via a versatile concept that we developed known as a “diagnostic template” (Figure 6).
A diagnostic template is a copy of the CDE for a particular type of finding with selected features that are matched to certain actions,
such as labeling an image finding with a Recommendation and/or RADS assessment based on the combination of CDE elements (Figure 7).
This process of creating diagnostic templates to determine RADS assessments has been implemented for each of the RADS reporting schema listed above.
Other benefits of creating discrete data using CDE’s linked to RADS assessments include (1) the generation of anatomy-specific roadmaps incorporating RADS information,
and (2) using the RADS data for determining medical outcomes (Figure 8).