ECR 2014 - RC1605 - Improving workflow efficiency and quality 2

reporting has four strategies:

1. text mode

2. fill in the blanks

3. serial assembly

4. mixed or nested assembly

Many different factors have been published that relate to errors in radiology reporting.  Therefore, a more structured approach to reporting is required since the images are not the product of radiology but the conclusion of the report is.

The ESR and RSNA has presented a joint initiative on structured reporting. The different affiliated, subspecialty and national societies of the ESR will be uploading structured reports through the website of the RSNA. The IHE MRRT Management of Radiology Report Templates defines the technical framework of building report templates.

The HTML reports downloadable from the RSNA repository can be used and locally adapted to have optimal reports.

In  the coming years ESR subcommittees should start to contribute to the RSNA and although practical clinical impelentations are limited, interest from radiologists and willingness to start testing these report is growing.

ECR 2014 - session SS105 - Trends in reporting, image management and mobile computing

10 presentations on different topics were covered in this scientific session:

CAD

Evaluation of a Siemens software tool called FastSpine to perform analysis of the spine with automatic labeling and curved reconstruction. They showed a high percentage of correct automatic labeling of 93.5%. FastSpine took about a minute to perform on average with manual correction where needed. The CAD provided time saving in more complex cases and had a comparable performance.

Two presentations focused on RECIST implementation. One looked at the implementation of dedicated tools for RECIST analysis and showed a clear saving in reading time while maintaing reproducibility. A Japanese group presented a cloud-based Lesion Management Solution that is used to track patients in lung cancer screening workup.

MOBILE DEVICES

This session showed again that mobile devices are of great interest and interesting applications are identified and implemented. Four of the presentations covered this topic.

A by dr von Falk showed that an iPad could be used for diagosis of CT data although screen size and security are still hampering the implementation. Another issue with mobile devices is the diiffering reading environments that cannot be controlled and thus could influence the reading quality. Mobile devices should probably be implemented as additional possibility and not for primary use. 

The dkfz demonstrated the use of an iPad for offline analysis of radiology information at a distant location to be used during autopsy. 

A group from Pisa also evaluated the iPad on CT images and showed the iPad to be faster for evaluation with even higher sensitivity and specificity for the iPad.

An important point to keep in mind with all the evidence for using iPads in radiology is that it is all focused on a limited, well defined, clinical question and mostly limits to CT.

Another group from Germany introduced the iPad in anatomical courses. The developed AnatomyMap a dedicated software tool for interactive exploration of volume rendered data running on a remote server using a web based interface.

QUALITY CONTROL

A group from switzerland showed that after transfer from free text to structured reporting they needed quaity control of those structured reports. They analysed about 250 reports on over 30 criteria. They showed that one third was excellent, one third good and one third non-optimal. The two worst performing radiologists were those not (properly) using the structured reporting. 

USABILITY

A Dutch group presented their work on interface customization. Most PACS systems have an adaptable interface, but the radiologist do not optimally use this capability. The showed that using adaptive support suggestions based on usage could help to improve radiology workflow and workstation usage. 

New advances in speech recognition and translation

Demonstration of new technology developments in speech recognition and transcription combined with translation that even allows voice translation. Very promising example and perhaps the future in tele-radiology allowing radiologists all over the world to dictate in their own language and having automatic translation of their report into the language of the requesting party.

NEW! IHE Radiology White Papers Published

The IHE Radiology Technical Committee has published the following white papers as of December 17, 2012:

• Cross-Enterprise Screening Mammography Workflow Definition
• Management of Radiology Report Templates 

ECR 2012 - SS505/B0376 - Semantic navigation in radiology

In this presentation (abstract here) a method was proposed that links semantic information from the radiology report directly to images. Additional value of the methodology would be the capability to auto compare what is written in the radiology report to the pathology reports that are constructed later in the diagnostic process. To achieve their objective the researchers developed tools to analyze and automatically index text to include them into a search engine. They suspect possible applications in clinical, research and teaching environments.

Clinical:

• Validate diagnosis
• Looking for best-practice reports
• Mobile app for clinicians
• Integration in referring physician portal

Research:

• Collecting patient cohorts

Teaching:

• Database for teaching cases
• Prepare lectures and talks

The concluded that their analysis shows that comprehensive syntactic and semantic text analysis is required to achieve high-quality information extraction from radiology reports. This could be the basis for a number of value-added applications in clinic, research and teaching.

ECR - MIR@ECR - Managing Radiology

Three presentations were given in this session.

The first one was on the implications for radiology of Imaging and benchmarking by dr. Schouman. Benchmarking is the comparison to best practices based on measurements. She promoted benchmarking to achieve improvement of the productivity and performance of a radiology department. The current situation with EHR, RIS and PACS starts to enable efficient benchmarking. Problems are
- Cumbersome data capture
- Poor data quality
- Lack of standardization
- Reductive analysis
- Meaningless comparisons
One of the key points is the assurance that manual data entry is done correctly. For example the information going into the RIS.

Second was the added value of in-house radiological IT by dr. Jakobsen. He showed, based on examples, that IT is crucial for the operation of a radiology department and that the absence of IT would take down that complete radiology department. The in-house IT support is critical to keep the IT up and running since they can either solve the problem or knows who to contact. Concluding in-house radiology IT can effectively prevent a stop in workflow.

Finally, dr. Senol talked about Risk management in radiology. As definition of risk he gave: The potential that an chosen action or inaction leads to a negative outcome. Risks in radiology are:
- related to reporting
- direct harm to the patient
- related to decision making
Risks related to reporting are:
- delay
- errors in reporting (up to 4% of reports)
- false positive or negative errors
- incorrect diagnosis
When an error occured patient and hospital should be informed and the root cause of the error should be assessed.

ECR 2012 - MIR@ECR - Communication with our Partners

Three presentations were given in this session.

First dr. Kahn talked about computerized physician order entry and decision support. He approached this from the side of the appropriateness criteria. He stated that those criteria are not meant for the radiologist but for the referring physicians. However, delivering it to them in an enormous pdf file is not very efficient. It should be integrated into the exam ordering process using computerized systems inside the electronic health record. By doing this the referring physician can enter the properties of the patients and get decision support on their order. This should reduce the number of wrong diagnostic tests or may even reduce the number of diagnostic tests. Kahn remarked that we should be talking about a radiology examination request and not an order. Furthermore he stated that decision support should be regarded as advise and not the law and that a physician should be able to reject the advice.

Next, dr. Centoze gave a presentation about communication of urgent and unexpected findings. He started to state that efficient and correct communication is crucial. ESR has release guideliness on communication of urgent and unexpected findings published in insights in imaging. Ideally an automatic alert and feedback system should be implemented. Urgent information should always be communicated directly with the referring physician. Unexpected findings could be communicated with other methods like e-mail .

Finally, a presentation on implications of reporting infrastructure - general reading room vs. Individual office was given by dr. Strickland. She presented her personal preference of a general reporting room and discussed pros and cons of that choice against private offices. The pros are:
- less costs because of maximum use of the workstations
- optimized ergonomics in the central reading rooms
- camaraderie: reporting is a social activity/teambuilding
- facilitates clinician access
- enhances inter-radiologist consultation
The perceived cons are
- noise
- personalized workstation settings lost
- pre-existing building architecture
Extra small reporting rooms could be beneficial to have. But they should not be personal.

ECR 2012 - SS505/B0373 - Automated semantic navigation and synchronized alignment

Semantic navigation is used to synchronize baseline and follow-up CT examinations. The algorithm detects words from the radiology report and puts links on it to automatically travel to the proper region in the images. To achieve this the software automatically detects landmarks in the anatomy and connects this to semantic information using rekationships like 'close to', 'above', 'to the right of', etcetera. this method allows matching of non-corresponding exams and provides clinically sufficient automatic alignment of corresponding findings in consecutive examinations.