Evaluating the Performance of AI in Evaluating Breast MRI Performed With Dose Reduction
- Conditions
- Breast Benign TumorBreast Malignant Tumor
- Interventions
- Drug: Standard of Care (SOC) gadolinium Breast MRIDrug: reduced 1/4 dose gadolinium Breast MRI with Artificial Intelligence (AI) to aid in evaluation.
- Registration Number
- NCT04340180
- Lead Sponsor
- University of Alabama at Birmingham
- Brief Summary
The purpose of this study is to evaluate the ability of AI to correctly aid in characterization of benign and malignant lesions even when a low dose of gadolinium is administered. This study is relevant for several reasons, most notably being the reduction of MRI dose and decreased gadolinium deposition in the brain. In addition, use of AI may provide increased sensitivity and specificity for the radiologist evaluating a breast MRI exam. Half of the population will have benign pathologies and the other half will have malignant pathologies.
- Detailed Description
The study involves each patient presenting for an initial MRI with a regular dose of gadolinium and then presenting at least 48 hours after (no less than 14 days later) for a ¼ dose (see below regarding dosing) gadolinium MRI exam. Both exams will be performed on a 1.5 Tesla magnet. Both exams will include a full protocol. The full dose contrast exam will be read as standard protocol. All images will be anonymized. Images from the reduced dose study will be collected and an AI algorithm applied. All three anonymized data sets (regular dose, low dose, and AI algorithm applied to low dose) will be provided to the readers.
Readers will be three attending radiologists specializing in breast imaging. Exams will be scored on quality, background parenchymal enhancement (BPE), and lesion conspicuity. Enhancing lesions will be identified and characterized by the radiologists in a document provided.
Recruitment & Eligibility
- Status
- ACTIVE_NOT_RECRUITING
- Sex
- Female
- Target Recruitment
- 20
Not provided
- Non-female patients
- Patients < 18 years old
- Women who are lactating or pregnant
- Patients with recurrent breast cancer
- Patients who have already received neoadjuvant chemotherapy
- Unable to lie still on the imaging table for one (1) hour
- Patients that are unable to undergo MRI evaluation for reasons specific to MRI
Study & Design
- Study Type
- INTERVENTIONAL
- Study Design
- SINGLE_GROUP
- Arm && Interventions
Group Intervention Description Subjects with enhancing breast lesions Standard of Care (SOC) gadolinium Breast MRI - Subjects with enhancing breast lesions reduced 1/4 dose gadolinium Breast MRI with Artificial Intelligence (AI) to aid in evaluation. -
- Primary Outcome Measures
Name Time Method Compare lesion size in millimeters from low-dose to SOC-dose imaging. From 1st SOC breast MRI to 2nd low-dose Breast MRI (48 hours to 14 days later). Compare lesion conspicuity on a scale of 1-4 from low-dose and SOC-dose imaging. From 1st SOC breast MRI to 2nd low-dose Breast MRI (48 hours to 14 days later). Compare lesion margins on a scale from 1-4 from low-dose and SOC-dose imaging. From 1st SOC breast MRI to 2nd low-dose Breast MRI (48 hours to 14 days later). Compare lesion internal enhancement pattern on a scale from 1-4 from low-dose and SOC- dose imaging. From 1st SOC breast MRI to 2nd low-dose Breast MRI (48 hours to 14 days later).
- Secondary Outcome Measures
Name Time Method Compare the overall background enhancement (scale per standard imaging guidelines. Minimal, mild, moderate, marked). From 1st SOC breast MRI to 2nd low-dose Breast MRI (48 hours to 14 days later).
Trial Locations
- Locations (1)
The Kirklin Clinic
🇺🇸Birmingham, Alabama, United States