Critical Test Results - Acute Intracranial Hemorrhage

Value Proposition

Many tests performed in an imaging department end up revealing a result that is life-threatening, also known as a critical finding. Acute intracranial hemorrhages are one of these potentially life-threatening emergencies in which the prognosis is highly dependent on timely detection and communication.  AI implementations that  meet this use case would help in rapidly screening, detecting, prioritizing, and alerting both the technologist and the radiologist to quickly assess the case and alert clinicians to implement life-saving care.  Additionally, the CT technologists are the first to see the patient images after the scan, and in some cases, will be the ones to alert the radiologists of a potential critical finding.  In clinical settings where radiologists are available, flagged cases for potential intracranial blood products could be escalated and be given an urgent designation for quicker radiologist interpretation.  

Narrative(s)

A 78-year-old patient presents to an outpatient imaging center  after being referred by their primary care physician for concern of altered mental status.  The patient has a significant past medical history of atrial fibrillation and long-term warfarin therapy.  A noncontrast head CT is obtained and the study is routed  to a PACS for interpretation.  Concurrently, an algorithm evaluates the cross-sectional imaging to screen for acute intracranial blood products, indicating present, absent, or undetermined.  When the algorithm returns a present or undetermined result, the exam is prioritized in the radiologist worklist and the technologist is notified of a possible critical result.  The technologist can not only alert the radiologist to look at the imaging sooner, but the technologist can also keep the patient at the imaging center until a management decision is made. 

Workflow Description

Computed tomography cross-sectional imaging is obtained from modality and sent to the AI engine first, followed by the RIS, Reporting Software platforms,  and  PACS.  The imaging is analyzed by the engine.  The system detects and reports acute intracranial blood products.  An alert message is sent to the worklist from the engine with the information regarding whether acute intracranial blood products are present, absent, or undetermined. A visual queue or designation identifying the images containing the finding is also generated. The worklist is modified in order to present the study in a means that increases the priority for the radiologist’s interpretation.

Input

DICOM Study

Primary Outputs

Acute intracranial blood products detection

Secondary Outputs

Side of acute intracranial blood products

Size of acute intracranial blood products

• The algorithm could be used to compare prior images and track the increase or resolution of the acute intracranial blood products.  

• Any midline shift can be tracked to assess for worsening herniation or as an indirect sign of acute intracranial hematoma growth.