From our deep roots in partnership to our intense focus on clinical value and technical excellence, Blackford is solely geared towards providing multiple imaging applications and AI that are integrated into existing workflow. But what does that entail?
At its simplest, workflow integration means bringing the benefits of all these apps into the radiologist’s existing tools and workflow. They don’t have to go to another machine, open another piece of software, or, ideally, even have to click on another button.
To understand what this means in practice, it can be useful to think about how applications of different types fit into existing workflow. Let’s consider some examples, covering four common segments of imaging applications and AI:
- Image enhancement
- Triage tools
- Value-add reports
- CAD tools
SubtlePET is an AI-powered software that denoises PET images acquired in a fraction of the original scan time. But how does the Blackford platform ensure that these PET images are sent for denoising to the SubtlePET application without additional work from the technologist?
After image acquisition, the relevant series are auto-forwarded to the Blackford Platform from the modality. The Blackford platform then utilizes relevant study tags in the DICOM data to identify the PET study as relevant for processing by SubtlePET and routes the study to the Subtle application for processing.
On completion of processing, results are returned to the imaging modality to allow for any manual QA steps and any additional image post processing. The images are then forwarded to PACS as per the site’s standard process. This workflow ensures that the SubtlePET processing fits seamlessly into the normal imaging workflow without any additional steps required from the technologist.
Once the study is received in PACS, the worklist will be updated to indicate that the study has been successfully processed and is ready for reading. On opening the study from their worklist, the radiologist will find the denoised PET images ready and waiting for them to report.
Accipio Ix is an AI triage tool that assesses non-contrast CT head images for intracranial hemorrhage to allow for notification and prioritization of suspected ICH.
CT Head imaging is performed to rule out intracranial hemorrhage in suspected stroke cases, but how does Blackford ensure that a stroke study is prioritized in the processing queue above routine outpatient CT Head studies?
The CT Head study is auto-forwarded to Blackford Platform for processing in parallel to sending to PACS. The Blackford platform identifies that study is labelled as “STAT” and therefore will prioritize this study above other non-STAT CT Head studies due to be processed by Accipio Ix. Sending studies directly from the modality workstation as opposed to waiting for them to be received by PACS before sending to the platform ensures a rapid turnaround of urgent stroke cases.
But how does the radiologist know if a study has been flagged as a potential ICH by Accipio Ix?
The best way to do this within the radiologist’s workflow is via their reading worklist. The Blackford platform communicates with the radiologist’s worklist via HL7 messages, which allow the platform to notify the worklist of positive findings from a triage application. This means the study can be flagged and/or prioritized on the radiologist’s worklist.
icobrain provides a solution for easy and reliable access to MR and CT quantification for conditions such as multiple sclerosis, dementia, brain trauma and epilepsy.
Consider an MS patient requiring an MRI imaging follow-up investigation at a site utilizing the icobrain application. How does Blackford ensure that this study is processed by icobrain without any additional manual steps being added to the tech or radiologist workflow?
Once acquired, the study is automatically forwarded to Blackford, which utilizes the HL7 order information to determine the relevant application for processing. The information in this HL7 order message is compared to customizable site-based relevancy rules for when to run the relevant icobrain application.
The “processing” status in the worklist shows the radiologist that the study is being processed by an AI application and should only be opened once complete. Once the study is identified as complete, the radiologist knows that the icobrain result is ready and waiting for their review.
In addition, using HL7 messaging, the Blackford platform can also transfer the relevant information from AI outputs to the reporting software in advance of the radiologist’s read, saving dictation time and allowing for AI to fit seamlessly within the reporting workflow.
ClearRead CT assists radiologists by suppressing normal structures in chest CTs, detecting and measuring solid, part-solid, and ground-glass nodules as well as assisting with longitudinal nodule comparisons.
The patient enters the imaging workflow through referral for lung cancer screening CT, and information from the HL7 order message is saved by the Blackford platform to allow for HL7-based relevancy determination. But how does Blackford ensure that the relevant prior study is sent along with the current for comparison by ClearRead CT?
The Blackford platform is aware that the ClearRead CT application is able to perform comparison to one prior CT chest study. The platform will therefore poll the PACS for most relevant prior to send for processing along with the current study. Once the study has been acquired, the images are automatically forwarded to the Blackford Platform in parallel to the standard workflow (along with the pre-identified prior study).
Utilizing the HL7 order information, the Blackford platform matches the information in the HL7 order message to determine whether the study is relevant for processing by ClearRead CT. For example, if the study is identified as “CT Chest without contrast”, and the reason for examination is “lung cancer screening study”, the platform will send the study to ClearRead CT for processing.
On reviewing their worklist, the radiologist will see that the chest study has been identified as relevant for processing by an AI application and the AI status will show as “processing”. Once the study is identified as complete, the radiologist knows that the AI result is ready and waiting for their review.
As before, results from ClearRead CT can also be sent to the radiologist’s reporting software using HL7 messaging to auto-populate the report in preparation for the read, saving dictation time and allowing for AI to fit seamlessly within the workflow.
Prefer to see it live? Schedule an end-to-end workflow demo.