Assess the Clinical Effectiveness in AI Prioritising CT Heads

NCT06027411 | Recruiting FDA Device

• 1 other identifier: 313507
• Study Type: observational
• Target: 16,800
• Locations: 1 country
• Sites: 4

Brief Summary

Non-Contrast Computed Tomography (NCCT) of the head is the most common imaging method used to assess patients attending the Emergency Department (ED) with a wide range of significant neurological presentations including trauma, stroke, seizure and reduced consciousness. Rapid review of the images supports clinical decision-making including treatment and onward referral. Radiologists, those reporting scans, often have significant backlogs and are unable to prioritise abnormal images of patients with time critical abnormalities. Similarly, identification of normal scans would support patient turnover in ED with significant waits and pressure on resources. To address this problem, Qure.AI has worked to develop the market approved qER algorithm, which is a software program that can analyse CT head to identify presence of abnormalities supporting workflow prioritisation. This study will trial the software in 4 NHS hospitals across the UK to evaluate the ability of the software to reduce the turnaround time of reporting scans with abnormalities that need to be prioritised.

Conditions

Trauma, Brain; Mass Lesion; Skull Fractures; Hemorrhage, Intracranial; Midline Deviation; Stroke, Ischemic; Encephalomalacia

Outcome Measures

Primary Outcomes (1)

• Reporting turnaround time with qER prioritisation

  Time taken to report NCCT head from acquisition for patients with prioritised findings in Emergency Department compared to standard of care. Measured as time in minutes from the scan acquisition to the final radiology report of prioritised scans.

  1 year

Secondary Outcomes (14)

• Reporting turnaround time with qER prioritisation for scans without prioritised findings in Emergency Department compared to standard of care.

  1 year

• Reporting turnaround time with qER prioritisation for scans with an absence of findings in Emergency Department compared to standard of care.

  1 year

• Assess the impact of qER on radiology reporting workflow on other requests for CT scans.

  1 year

• Impact of qER supported reporting on teleradiology.

  1 year

• Assess utility of qER to support clinical decision making of the patients from the emergency department requiring an NCCT.

  1 year

Study Arms (2)

Pre-implementation of qER

Baseline data: During the pre-implementation phase, we will be gathering data around the technical requirements for integrating qER into the radiology workflow. A random sample of 500 scans per site will be sent for the ground-truthing process for the purpose of technical evaluation. We will also be collecting data on the baseline status of all the endpoints including TAT. The reporting of NCCT scans will follow the same workflow as the current standard of care (i.e., the images/cases will appear in the RIS chronologically and the radiologist either follows this order or prioritises some cases based on communication from ED).

Post-implementation of qER

Post-implementation (Trial Intervention) In the post-implementation phase, there will be a notification (prioritised flag) in RIS. The order of the cases in RIS will not be altered. When the radiologist clicks a case in RIS, a secondary capture of qER along with the original images will be available in PACS. This secondary capture will have a contour showing the algorithm's attention point for a specific abnormality. The radiologist can then choose to agree with qER findings as it is or modify or ignore it according to their clinical judgement, writing and finally signing off the report. For scans which were not processed by qER the radiologist can prioritise and report as per the standard of care.

Device: qER (qER EU 2.0)

Interventions

qER (qER EU 2.0) DEVICE

Qure.ai's emergency room software solution qER (qER EU 2.0) is an AI medical device, developed by training a deep-learning algorithm using over 300,000 scans labelled by expert radiologists. qER has been shown to be accurate in identifying a range of abnormalities in NCCT head scans as well as prioritising them for urgent review and radiologist reporting. It is designated as a clinical support tool and, when used with original scans, can assist the clinician to improve efficiency, accuracy, and turnaround time in reading head CTs.

Post-implementation of qER

Eligibility Criteria

• Age: 18 Years+
• Sex: all
• Age Groups: Adult (18-64), Older Adult (65+)
• Sampling Method: Non-Probability Sample

Study Population

At each of the four participating sites, we will identify all patients referred through the Emergency Department NCCT requests.

You may qualify if:

• Individuals undergoing Head CT scan at the ED / A\&E (Accident and Emergency Services).
• Non-contrast axial CT scan series with consistently spaced axial slices.
• Soft reconstruction kernel covering the complete Brain.
• Maximum slice thickness of 6mm.

Sponsors & Collaborators

• Guy's and St Thomas' NHS Foundation Trust: lead
• Qure.ai: collaborator
• NHS Greater Glasgow and Clyde: collaborator
• Northumbria Healthcare NHS Foundation Trust: collaborator
• Oxford University Hospitals NHS Trust: collaborator

Study Sites (4)

NHS Greater Glasgow and Clyde

Glasgow, United Kingdom

RECRUITING

Guy's and St.Thomas Trusts

London, SE1 7EH, United Kingdom

RECRUITING

Northumbria Healthcare NHS Foundation Trust

Northumberland, NE27 0QJ, United Kingdom

RECRUITING

Oxford University Hospitals

Oxford, OX3 9DU, United Kingdom

RECRUITING

Related Publications (1)

• Vimalesvaran K, Robert D, Kumar S, Kumar A, Narbone M, Dharmadhikari R, Harrison M, Ather S, Novak A, Grzeda M, Gooch J, Woznitza N, Hall M, Shuaib H, Lowe DJ. Assessing the effectiveness of artificial intelligence (AI) in prioritising CT head interpretation: study protocol for a stepped-wedge cluster randomised trial (ACCEPT-AI). BMJ Open. 2024 Jun 16;14(6):e078227. doi: 10.1136/bmjopen-2023-078227.

  PMID: 38885990 DERIVED

MeSH Terms

Conditions

Brain Injuries, Traumatic; Skull Fractures; Intracranial Hemorrhages; Ischemic Stroke; Encephalomalacia

Condition Hierarchy (Ancestors)

Brain Injuries; Brain Diseases; Central Nervous System Diseases; Nervous System Diseases; Craniocerebral Trauma; Trauma, Nervous System; Wounds and Injuries; Fractures, Bone; Cerebrovascular Disorders; Vascular Diseases; Cardiovascular Diseases; Hemorrhage; Pathologic Processes; Pathological Conditions, Signs and Symptoms; Stroke

Study Officials

• Haris Shuaib, MSc

  Guy's and St.Thomas' Hospitals

  STUDY CHAIR

Central Study Contacts

Haris Shuaib, MSc

CONTACT

020 7188 7188; haris.shuaib@gstt.nhs.uk

Kavitha Vimalesvaran, MBBS MSc

CONTACT

020 7188 7188; kavitha.vimalesvaran@gstt.nhs.uk

Study Design

• Study Type: observational
• Observational Model: COHORT
• Time Perspective: PROSPECTIVE
• Target Duration: 28 Days
• Sponsor Type: OTHER
• Responsible Party: SPONSOR

Study Record Dates

• First Submitted: August 11, 2023
• First Posted: September 7, 2023
• Study Start: March 27, 2024
• Primary Completion: August 1, 2024
• Study Completion: August 1, 2024
• Last Updated: March 21, 2024

Record last verified: 2024-03

Data Sharing

• IPD Sharing: Will not share

Locations

GB (4)