AI Tool for Breast Cancer Screening

(PRISM Trial)

Recruiting at 1 trial location

Overseen By Michelle L'Hommedieu, PhD

Age: 18+

Sex: Any

Trial Phase: Academic

Sponsor: Jonsson Comprehensive Cancer Center

Disqualifiers: Non-adult, Non-patient, Non-radiologist, others

No Placebo Group All trial participants will receive the active study treatment (no placebo)

Compensation: Varies

Number of Visits: Unknown

154474 Participants Needed

Trial Summary

Will I have to stop taking my current medications?

The trial information does not specify whether participants need to stop taking their current medications.

What data supports the effectiveness of the AI decision-support tool for breast cancer screening?

Research shows that using an AI tool as an additional reader in breast cancer screening can detect 0.7-1.6 more cancers per 1,000 cases compared to standard methods, with minimal unnecessary recalls. Another study found that the AI system reduced false positives and false negatives, outperforming human experts and maintaining high accuracy while reducing the workload for radiologists. ^{1 2 3 4 5}

Is the AI tool for breast cancer screening safe for humans?

AI tools have been used to improve patient safety by predicting and preventing adverse events (unwanted side effects) in healthcare, such as drug reactions and diagnostic errors. While specific safety data for breast cancer screening isn't mentioned, AI has shown potential to enhance safety in various medical areas. ^{6 7 8 9 10}

How does the AI decision-support tool for breast cancer screening differ from other treatments?

The AI decision-support tool for breast cancer screening is unique because it uses artificial intelligence to assist radiologists in interpreting mammograms, reducing false positives and negatives, and improving accuracy compared to human readers alone. Unlike traditional methods, this tool enhances the efficiency of the screening process and can adapt to different clinical settings. ^{2 3 11 12 13}

What is the purpose of this trial?

The goal of this clinical trial is to compare patient-centered outcomes when 3D screening mammograms are interpreted with versus without a leading FDA-cleared AI decision-support tool in real-world U.S. settings and to assess patient's perspectives on AI in medicine. The main questions it aims to answer are: 1. Will AI use be associated with an increase in cancer detection and an initially higher recall rate as radiologists start using AI, followed by a recall rate comparable to that without AI (no more than 1.5 percentage-points higher) after a learning curve period? Will AI use will be associated with lower rates of missed breast cancers and similar rates of false alarms after a learning curve period? 2. Will improved patient outcomes with AI be most pronounced for exams on women who are White, older, and have less dense breasts, and on baseline exams? Will AI aid patient outcomes when the interpretation is by radiologists with less clinical experience, lower annual interpretive volume, and less tolerance of ambiguity? Yet, will there be greater automation bias (the tendency for humans to defer to a computer algorithms' results) noted among these radiologists? 3. What are patients' perspectives on AI in mammography, including their confidence in breast cancer screening when interpreted with vs. without AI? What are patients' perspectives on the importance of the study results? Researchers will compare patient-centered outcomes when 3D screening mammograms are interpreted with versus without a leading FDA-cleared AI decision-support tool in real-world U.S. settings. This trial will include all adult patients undergoing 3D mammography breast cancer screening at imaging facilities across University of California at Los Angeles and University of Washington health systems and all radiologists interpreting breast cancer screening. All screening mammograms at these facilities will be randomized to either intervention (radiologist with AI support) versus usual care (radiologist alone) to see if interpreting these mammograms with the AI tool's assistance improves patient outcomes.

Research Team

Hannah S Milch, MD

Principal Investigator

University of California, Los Angeles

Eligibility Criteria

This trial is for all adult patients getting 3D mammography breast cancer screenings at UCLA and UW health systems. It includes radiologists interpreting these screenings. There are no specific inclusion or exclusion criteria provided, suggesting it's a broad study.

Inclusion Criteria

I am an adult getting a 3D mammogram at UCLA or UW.

I am a radiologist who interprets breast cancer screenings at a specified facility.

Exclusion Criteria

Is neither an adult patient undergoing 3D mammography breast cancer screening at any of the 16 imaging facilities across UCLA and UW health systems

I am a radiologist who interprets breast cancer screenings at a specified facility.

Timeline

Screening

Participants are screened for eligibility to participate in the trial

2-4 weeks

Randomized Controlled Trial

3D screening mammograms are interpreted with versus without AI to assess immediate performance measures and outcomes

2 years

Regular visits as per mammography schedule

Follow-up

Participants are monitored for longer-term performance measures and clinical patient outcomes

1 year

Subgroup Analysis

Analysis of patient-, exam-, and radiologist-level characteristics associated with improved screening performance with AI

Concurrent with main trial

Treatment Details

Interventions

• Artificial intelligence (AI) decision-support tool

Trial Overview The trial tests if using an FDA-cleared AI tool to help read 3D mammograms improves detection of breast cancer and affects the recall rate after initial use. It also examines if outcomes differ based on patient demographics or radiologist experience.

Participant Groups

Group I: Intervention (radiologist assisted by AI decision-support tool) Active Control 1 Intervention

3D screening exams randomized to this arm will be interpreted by the radiologist assisted by the AI decision-support tool (i.e., intervention).

Group II: Standard care (radiologist alone) Active Control 1 Intervention

3D screening exams randomized to this arm will be interpreted in accordance with standard care (i.e., interpreted by the radiologist alone, without an AI decision-support tool's assistance).

Who Is Running the Clinical Trial?

Jonsson Comprehensive Cancer Center

Lead Sponsor

Trials

373

Recruited

35,200+

National Cancer Institute (NCI)

Collaborator

Trials

14,080

Recruited

41,180,000+

University of California, Los Angeles

Collaborator

Trials

1,594

Recruited

10,430,000+

National Institutes of Health (NIH)

Collaborator

Trials

2,896

Recruited

8,053,000+

Findings from Research

AI has the potential to enhance the processing and evaluation of Individual Case Safety Reports (ICSRs) in pharmacovigilance, as explored by the FDA, but current algorithms still require human oversight to ensure quality.

To fully leverage AI in pharmacovigilance, several challenges must be addressed, including the need for quality assurance, large training datasets, and the development of best practices and regulatory frameworks for implementation.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

"Artificial Intelligence" for Pharmacovigilance: Ready for Prime Time? Ball, R., Dal Pan, G.[2023]

A multidisciplinary expert group identified key signals that a new AI system should emit to enhance the management of adverse drug events (ADEs) in patients taking oral antineoplastic agents, including educational reminders for patients and alerts for necessary tests and examinations.

In a 6-month observational study at a university hospital, a total of 3,641 ADEs were reported, highlighting the importance of patient education and clinical monitoring in improving patient safety and quality of life.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Analysis of adverse drug events as a way to improve cancer patient care. Vicente-Oliveros, N., Gramage-Caro, T., Corral de la Fuente, E., et al.[2022]

A machine-learning model was developed using data from 4638 patients across 16 FDA-approved small molecule kinase inhibitors (SMKIs) to analyze the relationship between kinase targets and adverse events (AEs), providing a new tool for predicting safety risks in cancer treatments.

The model not only helps identify potential kinase-inhibitor adverse event pairs but also serves as a precision medicine tool to enhance patient safety by forecasting clinical safety signals and aiding in the development of safer SMKI therapies.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Decoding kinase-adverse event associations for small molecule kinase inhibitors. Gong, X., Hu, M., Liu, J., et al.[2022]

References

1. United States pubmed.ncbi.nlm.nih.gov

Prospective implementation of AI-assisted screen reading to improve early detection of breast cancer. [2023]

2. England pubmed.ncbi.nlm.nih.gov

International evaluation of an AI system for breast cancer screening. [2022]

3. Germany pubmed.ncbi.nlm.nih.gov

Impact of real-life use of artificial intelligence as support for human reading in a population-based breast cancer screening program with mammography and tomosynthesis. [2023]

4. United States pubmed.ncbi.nlm.nih.gov

Detection of Breast Cancer with Mammography: Effect of an Artificial Intelligence Support System. [2019]

5. United States pubmed.ncbi.nlm.nih.gov

A Deep Learning Decision Support Tool to Improve Risk Stratification and Reduce Unnecessary Biopsies in BI-RADS 4 Mammograms. [2023]

6. New Zealand pubmed.ncbi.nlm.nih.gov

"Artificial Intelligence" for Pharmacovigilance: Ready for Prime Time? [2023]

7. England pubmed.ncbi.nlm.nih.gov

Root causes of adverse drug events in hospitals and artificial intelligence capabilities for prevention. [2021]

8. England pubmed.ncbi.nlm.nih.gov

The potential of artificial intelligence to improve patient safety: a scoping review. [2021]

9. England pubmed.ncbi.nlm.nih.gov

Analysis of adverse drug events as a way to improve cancer patient care. [2022]

10. England pubmed.ncbi.nlm.nih.gov

Decoding kinase-adverse event associations for small molecule kinase inhibitors. [2022]

11. United States pubmed.ncbi.nlm.nih.gov

Artificial Intelligence (AI) for Screening Mammography, From the AJR Special Series on AI Applications. [2022]

12. Saudi Arabia pubmed.ncbi.nlm.nih.gov

The utilization of artificial intelligence applications to improve breast cancer detection and prognosis. [2023]

13. United States pubmed.ncbi.nlm.nih.gov

Impact of Different Mammography Systems on Artificial Intelligence Performance in Breast Cancer Screening. [2023]