EvoXRare AI-Assisted Interpretation of Emergency Chest Radiographs (CXR-AI-RCT)

EvoXRare-Assisted Chest Radiograph Interpretation in Emergency and Critically Ill Adults: A Nationwide Multicenter Randomized Controlled Trial

Chest radiography is widely used in emergency departments, resuscitation rooms, intensive care units, and other acute-care settings. Clinically important abnormalities may be missed because findings are subtle, uncommon, obscured by overlapping anatomy, or encountered under high workload. This pragmatic randomized trial evaluates whether EvoXRare, an artificial intelligence system that provides a draft chest radiograph report and high-risk alerts for uncommon or clinically important abnormalities, reduces clinically important diagnostic misses compared with routine physician interpretation alone. Approximately 70,000 adults will be randomized 1:1 across 15-20 hospitals in China. The primary endpoint is the patient-level rate of clinically important abnormalities present on the reference standard but not reported in the initial clinical interpretation. Key secondary outcomes include detection of rare or easily missed findings, time to recognition of critical abnormalities, report turnaround time, false-positive alerts, additional imaging, recall/communication events, and selected short-term clinical outcomes.

Study Overview

Detailed Description

This is a nationwide, multicenter, prospective, parallel-group, superiority randomized controlled trial with individual participant-level 1:1 allocation and blinded central endpoint adjudication. Eligible participants are adults undergoing clinically indicated chest radiography in emergency or critical-care settings. Only one index radiograph encounter per participant contributes to the primary randomized analysis.

In the control arm, physicians interpret the chest radiograph using the participating hospital's usual workflow without access to EvoXRare trial output. In the intervention arm, EvoXRare processes the same image on an independent AI workstation and provides a pre-generated report and/or high-risk prompts. The physician must review the original image, verify or edit the AI output, and remains responsible for the final report and all clinical communication. Both arms retain routine senior review and local critical-result procedures.

The AI model version, weights, preprocessing, thresholds, supported finding set, and report format are frozen before formal enrollment. No online learning or outcome-driven recalibration occurs during the randomized trial. If AI processing fails or images do not meet technical requirements, clinical care proceeds without delay and the participant remains in the assigned group for intention-to-treat analysis.

The reference standard is established by a blinded adjudication process using the index radiograph together with clinically available follow-up information, which may include subsequent CT, ultrasound, pathology, surgery, laboratory findings, clinical course, discharge diagnosis, or follow-up imaging. Adjudicators do not use the randomized assignment or EvoXRare output when determining whether a prespecified abnormality was present on the index radiograph.

The trial is designed primarily to test clinical diagnostic performance and workflow impact, not to retrain the AI system. It specifically emphasizes uncommon, high-consequence, and easily missed chest radiograph findings while also evaluating overall report quality and efficiency.

Study Type

Interventional

Enrollment (Estimated)

70000

Phase

  • Phase 1

Contacts and Locations

This section provides the contact details for those conducting the study, and information on where this study is being conducted.

Study Contact

Participation Criteria

Researchers look for people who fit a certain description, called eligibility criteria. Some examples of these criteria are a person's general health condition or prior treatments.

Eligibility Criteria

Ages Eligible for Study

18 years to 90 years (Adult, Older Adult)

Accepts Healthy Volunteers

No

Genders Eligible for Study

All

Description

Inclusion Criteria:

  • Age 18 years or older.
  • Undergoing a clinically indicated chest radiograph in a participating emergency department, resuscitation room, intensive care unit, or other prespecified acute/critical-care setting.
  • Chest radiograph is technically accessible to the study workflow before or during routine clinical interpretation.
  • The participant has not previously been randomized in this trial, unless the final protocol explicitly permits a separate recurrent-encounter analysis that does not enter the primary endpoint.
  • Consent or ethics-approved waiver/alteration of consent is satisfied according to the activated site protocol.

Exclusion Criteria:

  • Age younger than 18 years.
  • Image is missing, corrupted, or technically non-diagnostic to a degree that prevents clinical interpretation.
  • Radiograph was obtained solely for research rather than clinical care.
  • AI processing cannot be initiated because the image format or transfer route is outside the locked system specification; such cases may be recorded in screening logs but are not randomized unless the protocol allows post-randomization technical failure.
  • Prior randomization of the same participant in this trial for the primary participant-level analysis.
  • Any local legal, ethical, or institutional restriction that prevents study participation or required data use.

Study Plan

This section provides details of the study plan, including how the study is designed and what the study is measuring.

How is the study designed?

Design Details

  • Primary Purpose: Diagnostic
  • Allocation: Randomized
  • Interventional Model: Parallel Assignment
  • Masking: Quadruple

Arms and Interventions

Participant Group / Arm
Intervention / Treatment
Experimental: EvoXRare AI-Assisted Chest Radiograph Interpretation.

Usual chest radiograph interpretation plus access to EvoXRare-generated draft report and high-risk prompts. The physician reviews the original image, accepts/rejects/edits AI suggestions, submits the final report, and follows local critical-result communication procedures.

EvoXRare is an artificial intelligence software system that analyzes chest radiographs and provides draft reports and high-risk prompts to the interpreting physician. The physician reviews the original images and AI outputs and accepts, rejects, or edits the suggestions before finalizing the radiology report.

No Intervention: No Intervention / Active usual-care comparator: Standard Chest Radiograph Interpretation.

Routine physician interpretation according to the participating hospital's usual workflow, without access to EvoXRare trial output. Routine nontrial tools already in use may continue only if documented before site activation and applied consistently.

What is the study measuring?

Primary Outcome Measures

Outcome Measure
Measure Description
Time Frame
Patient-Level Clinically Important Abnormality Miss Rate on the Index Chest Radiograph
Time Frame: From index radiograph interpretation through completion of blinded endpoint adjudication, using clinically available information obtained during the index encounter and prespecified follow-up window (target: within 30 days when available).

Proportion of randomized participants with at least one prespecified clinically important abnormality (Category A or B) judged present on the reference standard but absent from the initial finalized chest radiograph report produced under the assigned strategy.

From index radiograph interpretation through completion of blinded endpoint adjudication, using clinically available information obtained during the index encounter and prespecified follow-up window (target: within 30 days when available).

Collaborators and Investigators

This is where you will find people and organizations involved with this study.

Study record dates

These dates track the progress of study record and summary results submissions to ClinicalTrials.gov. Study records and reported results are reviewed by the National Library of Medicine (NLM) to make sure they meet specific quality control standards before being posted on the public website.

Study Major Dates

Study Start (Estimated)

November 1, 2026

Primary Completion (Estimated)

November 1, 2027

Study Completion (Estimated)

November 1, 2028

Study Registration Dates

First Submitted

November 3, 2026

First Submitted That Met QC Criteria

November 3, 2026

First Posted (Actual)

November 9, 2026

Study Record Updates

Last Update Posted (Actual)

November 9, 2026

Last Update Submitted That Met QC Criteria

November 3, 2026

Last Verified

November 1, 2026

More Information

Terms related to this study

Other Study ID Numbers

  • WHUH-CXR-AI-RCT-2026-11

Plan for Individual participant data (IPD)

Plan to Share Individual Participant Data (IPD)?

YES

IPD Plan Description

Yes, through controlled access subject to compatible consent/ethics approval and institutional authorization.

IPD Sharing Supporting Information Type

  • STUDY_PROTOCOL
  • SAP
  • ICF
  • ANALYTIC_CODE
  • CSR

Drug and device information, study documents

Studies a U.S. FDA-regulated drug product

No

Studies a U.S. FDA-regulated device product

No

This information was retrieved directly from the website clinicaltrials.gov without any changes. If you have any requests to change, remove or update your study details, please contact register@clinicaltrials.gov. As soon as a change is implemented on clinicaltrials.gov, this will be updated automatically on our website as well.

Subscribe