ROTEM Interpretation AI vs Experts

NCT07502950 | Not Yet Recruiting

• 1 other identifier: ROTEMAI
• Study Type: observational
• Target: 400
• Locations: 0 countries
• Sites: N/A

Brief Summary

This prospective multicenter observational study aims to evaluate the agreement between artificial intelligence (AI)-based interpretation and expert interpretation of rotational thromboelastometry (ROTEM) findings in clinically relevant settings. ROTEM is widely used to guide hemostatic therapy in perioperative and critically ill patients, but its interpretation is complex and subject to interobserver variability. The primary objective is to determine whether AI-based interpretation achieves agreement comparable to variability between expert clinicians. Secondary objectives include comparison of interpretation time, assessment of consistency of AI outputs, and evaluation of potential differences in clinical decision-making. ROTEM datasets will be independently assessed by multiple expert anesthesiologists and by an AI-based model using standardized input. Agreement between methods and variability of interpretation will be analyzed. The study aims to determine whether AI-assisted interpretation could serve as a reliable decision-support tool and reduce variability in ROTEM-guided clinical practice.

Conditions

Coagulopathy; Thromboelastography; Agreement

Keywords

ROTEM; thromboelastography; coagulopathy

Outcome Measures

Primary Outcomes (1)

• Agreement between AI-based and expert interpretation of ROTEM findings

  Agreement between artificial intelligence (AI)-based interpretation and expert interpretation of ROTEM findings will be assessed and compared to inter-expert agreement. Agreement will be evaluated for predefined clinical decision outputs, including: * presence of coagulation disorder (yes/no), * identification of dominant coagulation abnormality (categorical), * selection of therapeutic intervention (categorical), * and recommended treatment dose (continuous). Agreement will be quantified using appropriate statistical methods for categorical and continuous data (e.g., Cohen's/Fleiss' kappa and intraclass correlation coefficients).

  Up to 24 hours after ROTEM measurement (time required for interpretation and data recording).

Secondary Outcomes (4)

• Interpretation time

  Up to 24 hours after ROTEM measurement.

• Consistency of interpretation (intra-method variability)

  Up to 24 hours after ROTEM measurement

• Proportion of clinically discordant decisions

  Up to 24 hours after ROTEM measurement

• Inter-expert agreement

  Up to 24 hours after ROTEM measurement

Other Outcomes (1)

• Potential treatment change

  Up to 24 hours after ROTEM measurement

Study Arms (1)

Patients indicated for thromboelastography.

Intensive care unit adult patients that are indicated for thromboelastography.

Other: Large Language Model (LLM) artificial intelligence assesment

Interventions

Large Language Model (LLM) artificial intelligence assesment OTHER

The thromboelastography record will be assessed by LLM based artificial intelligence.

Patients indicated for thromboelastography.

Eligibility Criteria

• Age: 18 Years+
• Sex: all
• Healthy Volunteers: No
• Age Groups: Adult (18-64), Older Adult (65+)
• Sampling Method: Non-Probability Sample

Study Population

The study population will consist of adult patients undergoing rotational thromboelastometry (ROTEM) as part of routine clinical care in participating centers. Patients will be recruited from perioperative and intensive care settings, including major surgery, trauma, and critical illness, where assessment of coagulation status is clinically indicated. This study reflects real-world practice, with no modification of standard care. The unit of analysis will be individual ROTEM measurements rather than individual patients, as repeated measurements may occur during clinical management. Only ROTEM datasets with complete graphical and numerical outputs and meeting predefined quality criteria will be included.

You may qualify if:

• Adult patients (age ≥18 years)
• ROTEM analysis performed using a ROTEM Sigma device as part of routine clinical care
• Availability of complete ROTEM output (graphical and numerical data)
• ROTEM measurement obtained in a clinical context where assessment of coagulation status is indicated (e.g., perioperative setting, trauma, or critical illness)

You may not qualify if:

• ROTEM measurements performed using the HEPTEM channel
• Incomplete or missing ROTEM data preventing standardized evaluation
• ROTEM measurements obtained under non-standardized or technically unreliable conditions

Sponsors & Collaborators

• Masarykova Nemocnice v Usti nad Labem, Krajska Zdravotni a.s.: lead

MeSH Terms

Conditions

Hemostatic Disorders

Interventions

Large Language Models

Condition Hierarchy (Ancestors)

Vascular Diseases; Cardiovascular Diseases; Hemorrhagic Disorders; Hematologic Diseases; Hemic and Lymphatic Diseases

Intervention Hierarchy (Ancestors)

Deep Learning; Machine Learning; Artificial Intelligence; Algorithms; Mathematical Concepts; Neural Networks, Computer

Study Design

• Study Type: observational
• Observational Model: OTHER
• Time Perspective: PROSPECTIVE
• Sponsor Type: OTHER
• Responsible Party: PRINCIPAL INVESTIGATOR
• PI Title: deputy for science and research

Study Record Dates

• First Submitted: March 25, 2026
• First Posted: March 31, 2026
• Study Start (Estimated): September 1, 2026
• Primary Completion (Estimated): June 1, 2027
• Study Completion (Estimated): July 1, 2027
• Last Updated: March 31, 2026

Record last verified: 2026-03