AI (Artificial Intelligence), Breast Cancer, Cancer Screening, Mass Screening, Ultrasound Imaging
Conditions
Keywords
Remote breast cancer screening, AB US, AI, Scanning and reading patterns
Brief summary
A multi-center, prospective, cohort study to evaluate the efficiency of breast cancer screening based on Automated Breast Ultrasound (AB US) with remote reading mode.
Detailed description
The burden of breast cancer incidence and mortality is rapidly growing during the past two decades in China. Screening has been proven to be effective in detecting early-stage disease and reducing mortality of 10% to 39% due to breast cancer. US is used as a primary screening method among Chinese women because Asian women characteristically have higher-density breasts than other ethnic groups. However, US is dependent on operator experience, lack of standardized scanning protocols, limited ultrasound physician and heavy workloads. In order to improve the efficiency and quality of breast cancer screening among Chinese women, a new remote screening pattern based on AB US was proposed and studied for breast cancer in China. This multi-center, prospective, cohort study aims to evaluate the efficiency of breast cancer screening based on AB US with remote reading mode. In addition, we will also explore the practice value of AI on AB US screening and efficient image acquiring and reading modes of AB US.
Interventions
DIAGNOSTIC_TESTConventional-reading
Two radiologists interpret at least three views of each breast without AI independently.
DIAGNOSTIC_TESTSecond-reading
One radiologist reads AB US images without AI first, then combines the indications of AI marks to make the final decision.
DIAGNOSTIC_TESTConcurrent-reading
One radiologist identifies CAD marks first, then quickly browses the entire AB US examination to make the final decision.
DIAGNOSTIC_TESTTow views-reading
Two radiologists interpret only two views of each breast without AI independently.
DIAGNOSTIC_TESTHandheld US-screening
One radiologist screens the breast cancer using Handheld US and interprets images immediately.
Sponsors
The First Affiliated Hospital of the Fourth Military Medical University
Study design
Observational model
CASE_ONLY
Time perspective
PROSPECTIVE
Eligibility
Sex/Gender
FEMALE
Age
35 Years to No maximum
Healthy volunteers
No
Inclusion criteria
1. Any age at 35 and over 2. Screening breast cancer with AB US 3. Obtaining written informed consent
Exclusion criteria
1. symptoms of breast cancer 2. surgical of breast within 12 months prior to the study 3. known diagnosis breast cancer 4. poor image quality 5. follow-up was less than 12 months
Design outcomes
Primary
| Measure | Time frame | Description |
|---|---|---|
| The performance of breast cancer screening based on AB US with remote reading mode | 4 years | The performance of breast cancer screening based on AB US with remote reading mode will be evaluated by cancer detection rate, recall rate, sensitivity, specificity, positive predictive value, early detection of breast cancer, interval cancers and time for image acquisition and interpretation. |
Secondary
| Measure | Time frame | Description |
|---|---|---|
| Evaluate the performance of AI on breast cancer screening based on AB US with remote reading mode | 4 years | Evaluate the performance of AI on breast cancer screening based on AB US with remote reading mode (Conventional-reading vs. Second-reading / Concurrent-reading) |
| Evaluate the performance of the image acquisition and interpretation modes in breast cancer screening based on AB US with remote reading mode | 4 years | Evaluate the performance of the image acquisition and interpretation modes in breast cancer screening based on AB US with remote reading mode (two views vs. three views) |
| Evaluate the performance of breast cancer screening according to the breast US device | 4 years | Evaluate the performance of breast cancer screening according to the breast US device (Hand-held US vs. AB US) |
Countries
China
Outcome results
None listed