Non-Invasive Identification of Endometrial Cancer/Endometrial Atypical Hyperplasia with an AI-Based Classifier Applied to Transvaginal Ultrasound in Patients with Post-Menopausal Bleeding

Recruiting

• Conditions: Endometrial Carcinoma
• Interventions: Other: Non-Interventional Study

• Registration Number: NCT06365905
• Lead Sponsor: Mayo Clinic

Brief Summary

This study evaluates if AI can be used with transvaginal ultrasound images for early detection of endometrial cancer or premalignant lesions.

Detailed Description

PRIMARY OBJECTIVE:

I. Prospective validation of an artificial intelligence (AI) algorithm applied to transvaginal ultrasound (TVUS) to identify patients with premalignant/malignant in a population of women with postmenopausal bleeding (PMB).

OUTLINE: This is an observational study.

Patients undergo a transvaginal ultrasound examination and endometrial sampling per standard of care and have their medical records reviewed on study.

Study Timeline

• Study Start: 2024-03-27
• Study First Submitted: 2024-04-10
• Study First Submitted QC: 2024-04-10
• Study First Posted: 2024-04-15
• Status Verified: 2024-10
• Last Update Submitted: 2024-10-03
• Last Update Posted: 2024-10-10
• Primary Completion: 2027-06-30
• Study Completion: 2027-06-30

Recruitment & Eligibility

• Status: RECRUITING
• Sex: Female
• Target Recruitment: 300

Inclusion Criteria

• Age ≥ 55 years
• At least one episode of postmenopausal bleeding (PMB)
• Endometrial biopsy available from at Mayo Clinic (MN, AZ, FL), Mayo Clinic Health System (MCHS), or an external institution
• Acceptance to participate in the study

Exclusion Criteria

• Diagnostic tests performed more than 90 days apart
• Physical impediment/refusal to undergo transvaginal ultrasound (TVUS)

Study & Design

• Study Type: OBSERVATIONAL
• Study Design: Not specified

Arm && Interventions

Group	Intervention	Description
Observational	Non-Interventional Study	Patients undergo a transvaginal ultrasound examination and endometrial sampling per standard of care and have their medical records reviewed on study.

Primary Outcome Measures

Name	Time	Method
Accuracy of artificial intelligence (AI) algorithm applied to transvaginal ultrasound (TVUS)	Baseline	Two static TVUS images (one longitudinal, one transversal) will be independently validated and compared to results of endometrial sampling. Outcomes from these will be compared with predictions made by the AI models for accuracy of assessing premalignant/malignant disease.

Trial Locations

Locations (3)

Mayo Clinic in Arizona; 🇺🇸 Scottsdale, Arizona, United States

Mayo Clinic in Florida; 🇺🇸 Jacksonville, Florida, United States

Mayo Clinic in Rochester; 🇺🇸 Rochester, Minnesota, United States