AI Detection for Cardiovascular Disease Prevention

(AI INFORM Trial)

Age: 18+

Sex: Any

Trial Phase: Academic

Sponsor: Brigham and Women's Hospital

Disqualifiers: Coronary artery disease, Cancer, others

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

Approved in 1 Jurisdiction This treatment is already approved in other countries

Compensation: Varies

Number of Visits: Unknown

1500 Participants Needed

Trial Summary

What is the purpose of this trial?

AI INFORM is a multicenter randomized trial that will test the hypothesis that providing clinicians information on the presence and amount of coronary artery calcifications (CAC), will result in initiation or intensification of preventive therapies. The study will use a cloud-based artificial intelligence (AI) platform (Nanox.AI) that can analyze non contrast chest CT and estimate the amount of CAC.

Will I have to stop taking my current medications?

The trial information does not specify whether you need to stop taking your current medications.

What data supports the effectiveness of the treatment Nanox.AI Coronary Artery Calcification Assessment for cardiovascular disease prevention?

Research shows that AI-based systems can automatically and accurately measure coronary artery calcium, which is a strong predictor of heart-related events. This automated approach is efficient and correlates well with manual methods, suggesting it could help in early detection and prevention of cardiovascular diseases. ^{1 2 3 4 5}

Is the AI Detection for Cardiovascular Disease Prevention safe for humans?

The research articles focus on the accuracy and reliability of AI systems in detecting coronary calcifications using CT scans, but they do not provide specific safety data for human use of the AI technology itself. ^{4 6 7 8 9}

How does the AI-based treatment for cardiovascular disease prevention differ from other treatments?

This AI-based treatment is unique because it uses artificial intelligence to automatically detect and quantify coronary artery calcium from CT scans, which helps in assessing the risk of cardiovascular disease more efficiently and accurately compared to traditional methods that rely on manual analysis. ^{1 2 5 9 10}

Eligibility Criteria

This trial is for individuals aged 40-75 who've had a chest CT scan in the last 3 years and have no history of coronary artery disease, cancer, or any other life-limiting conditions. It's aimed at enhancing cardiovascular disease prevention.

Inclusion Criteria

I am between 40-75 years old and had a chest CT scan in the last 3 years.

Exclusion Criteria

I do not have any other condition that could shorten my life.

I have had heart disease related to my arteries.

I have had cancer before.

Timeline

Screening

Participants are screened for eligibility to participate in the trial

2-4 weeks

Intervention

Notification or non-notification of coronary artery calcification detected by AI, followed by recommendation or non-recommendation of preventive therapy

6 months

Follow-up

Participants are monitored for changes in LDL-C, initiation or intensification of preventive therapies, and occurrence of cardiovascular events

12 months

Treatment Details

Interventions

• Nanox.AI Coronary Artery Calcification Assessment

Trial Overview The AI INFORM study is testing if notifying clinicians about the presence and amount of calcium buildup in arteries using an AI tool (Nanox.AI) leads to better preventive care. This involves analyzing past chest CT scans without contrast.

Participant Groups

Group I: Notification of Coronary Artery Calcification Experimental Treatment 1 Intervention

Notification to providers of the presence of coronary artery calcification automatically detected by AI based device (software) on chest CT. Recommendation of preventive therapy.

Group II: Non-notification of Coronary Artery Calcification Active Control 1 Intervention

No notification to providers of the presence of coronary artery calcification on chest CT automatically detected by AI based device (software) on chest CT. No recommendation of preventive therapy.

Who Is Running the Clinical Trial?

Brigham and Women's Hospital

Lead Sponsor

Trials

1,694

Recruited

14,790,000+

Nano-X Imaging Limited

Collaborator

Trials

1

Recruited

1,500+

Findings from Research

The AI-based automatic coronary artery calcium (CAC) scoring software demonstrated excellent correlation and agreement with a semi-automatic method, with a Spearman correlation coefficient of 0.935 for Agatston score, indicating high reliability in measuring coronary artery health.

The automatic software was significantly faster, taking a median of 36 seconds compared to 59 seconds for the semi-automatic method, making it a more efficient option for clinical use in assessing cardiovascular risk.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Evaluation of an AI-based, automatic coronary artery calcium scoring software. Sandstedt, M., Henriksson, L., Janzon, M., et al.[2021]

The gradient boosting machine (GBM) model demonstrated a remarkable sensitivity of 100% for predicting obstructive coronary artery disease (CAD) in a study of 435 patients, indicating it can effectively rule out the disease.

While the model showed high negative predictive value (100%), its positive predictive value was lower at 38%, suggesting that while it is excellent at identifying patients without CAD, it may not be as reliable in confirming the presence of the disease.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Machine Learning-based Algorithm Enables the Exclusion of Obstructive Coronary Artery Disease in the Patients Who Underwent Coronary Artery Calcium Scoring. Głowacki, J., Krysiński, M., Czaja-Ziółkowska, M., et al.[2020]

An AI-based automatic coronary artery calcium (CAC) scoring system demonstrated excellent reliability when compared to manual scoring in low-dose chest CT scans, with a high intraclass correlation coefficient (0.989) across a study of 452 subjects from three institutions.

While the AI system showed good performance overall, the reliability of CAC severity categorization varied among different datasets, indicating that results may differ based on the institution's imaging quality and conditions.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Fully automatic coronary calcium scoring in non-ECG-gated low-dose chest CT: comparison with ECG-gated cardiac CT. Suh, YJ., Kim, C., Lee, JG., et al.[2023]

A deep learning system was developed to automatically quantify coronary artery calcium from routine CT scans, demonstrating strong predictive power for cardiovascular events in a study of 20,084 individuals across various cohorts.

The automated scoring system showed high correlation with manual quantification and robust reliability, suggesting it could effectively enhance clinical practice by streamlining the assessment of cardiovascular risk.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Deep convolutional neural networks to predict cardiovascular risk from computed tomography. Zeleznik, R., Foldyna, B., Eslami, P., et al.[2022]

An AI model was developed to automatically quantify coronary artery calcium (CAC) using deep learning on ECG-gated cardiac CT images, trained on 560 images and validated on 409 additional images, achieving excellent accuracy with a Cohen's kappa of 0.95 in the test cohort.

The AI model demonstrated minimal differences in CAC scores compared to manual evaluations, indicating its potential for reliable and efficient assessment of coronary artery health, which could enhance clinical decision-making.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Fully automated coronary artery calcium quantification on electrocardiogram-gated non-contrast cardiac computed tomography using deep-learning with novel Heart-labelling method. Takahashi, D., Fujimoto, S., Nozaki, YO., et al.[2023]

In a study involving 25 patients, the reliability of scoring coronary artery calcium burden using CT scans showed significant interobserver and intraobserver disagreements, particularly on small lesions.

Despite these disagreements, the impact on the overall measures of calcium burden (calcium score, number of calcified lesions, and calcified area) was negligible, suggesting that multiple observers scoring the same CT exam may not be necessary.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Detection of heart calcification with electron beam CT: interobserver and intraobserver reliability for scoring quantification. Kaufmann, RB., Sheedy, PF., Breen, JF., et al.[2016]

Coronary calcifications are a key indicator of coronary heart disease (CHD) and their presence can help diagnose atherosclerotic disease in asymptomatic individuals with risk factors, potentially before any symptoms appear.

Calcium scoring using advanced imaging techniques like electron beam computed tomography (EBCT) and multi-slice CT allows for precise quantification of coronary calcifications, but the long-term prognostic value of these scores in asymptomatic high-risk patients is still uncertain.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

[Detection and quantification of coronary calcification: an update]. Fischbach, R., Heindel, W.[2016]

A new method using two convolutional neural networks can automatically detect calcifications in low-dose chest CT scans, which helps identify cardiovascular risks in heavy smokers, similar to lung cancer risks.

The method showed high reliability in detecting calcifications, achieving F1 scores up to 0.89 for coronary artery calcifications, indicating it can effectively assess cardiovascular risk in patients undergoing lung cancer screening.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Automatic Calcium Scoring in Low-Dose Chest CT Using Deep Neural Networks With Dilated Convolutions. Lessmann, N., van Ginneken, B., Zreik, M., et al.[2019]

A fully automated method for detecting coronary calcifications from non-contrast-enhanced CT scans achieved a detection rate of 73.8% for coronary calcifications with only 0.1 false positives per scan, demonstrating its efficacy in identifying potential cardiovascular risks.

The method successfully categorized 93.4% of scans into the correct risk categories based on computed calcium scores, indicating its potential for accurate risk assessment in patients.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Detection of coronary calcifications from computed tomography scans for automated risk assessment of coronary artery disease. Isgum, I., Rutten, A., Prokop, M., et al.[2019]

A fully automated AI model for scoring coronary artery calcium (CAC) demonstrated high accuracy, correlating strongly with manual assessments, with a Spearman's correlation of 0.90 and an intraclass correlation coefficient of 0.98 based on 1849 cardiac CT scans.

The AI model was able to classify 89% of scans into the same risk categories as human experts, while significantly reducing analysis time to an average of just 13.1 seconds per scan, suggesting it could enhance clinical workflow efficiency.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Evaluation of an artificial intelligence coronary artery calcium scoring model from computed tomography. Ihdayhid, AR., Lan, NSR., Williams, M., et al.[2023]

References

1. Germany pubmed.ncbi.nlm.nih.gov

Evaluation of an AI-based, automatic coronary artery calcium scoring software. [2021]

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

Machine Learning-based Algorithm Enables the Exclusion of Obstructive Coronary Artery Disease in the Patients Who Underwent Coronary Artery Calcium Scoring. [2020]

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

Fully automatic coronary calcium scoring in non-ECG-gated low-dose chest CT: comparison with ECG-gated cardiac CT. [2023]

4. England pubmed.ncbi.nlm.nih.gov

Deep convolutional neural networks to predict cardiovascular risk from computed tomography. [2022]

5. England pubmed.ncbi.nlm.nih.gov

Fully automated coronary artery calcium quantification on electrocardiogram-gated non-contrast cardiac computed tomography using deep-learning with novel Heart-labelling method. [2023]

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

Detection of heart calcification with electron beam CT: interobserver and intraobserver reliability for scoring quantification. [2016]

7. Germany pubmed.ncbi.nlm.nih.gov

[Detection and quantification of coronary calcification: an update]. [2016]

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

Automatic Calcium Scoring in Low-Dose Chest CT Using Deep Neural Networks With Dilated Convolutions. [2019]

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

Detection of coronary calcifications from computed tomography scans for automated risk assessment of coronary artery disease. [2019]

10. Germany pubmed.ncbi.nlm.nih.gov

Evaluation of an artificial intelligence coronary artery calcium scoring model from computed tomography. [2023]