Blood Based Risk Evaluation With AI for Targeted Primary Health Care in Early Lung Cancer Detection (BREATHE)

The study is a prospective, non-randomized feasibility study evaluating blood sample and machine learning-based risk stratification for lung cancer in patients with COPD (chronic obstructive pulmonary disease).

Patients with COPD will be recruited in general practice, where they will have a blood sample drawn. All data will be analyzed by the machine learning model, and patients with increased risk of lung cancer will be referred for a low-dose CT scan of the chest.

The primary objective of the study is to evaluate the feasibility of AI and DNA methylation-based risk stratification for lung cancer in patients with COPD in a primary care setting.

The secondary objectives are to evaluate the safety of the risk stratification approach, the potential effects on quality of life and wellbeing, to gain insight into the patient and physician perspectives, and to estimate the health economic consequences.

Study Overview

• Status: Recruiting
• Conditions: Lung Cancer (Diagnosis)
• Intervention / Treatment: Other: Risk stratification

Detailed Description

Lung cancer causes the highest number of cancer-related deaths. Around 5000 people are diagnosed with lung cancer annually in Denmark, and people with chronic obstructive pulmonary disease (COPD) have a higher risk compared to the general population. Screening with low-dose computed tomography (LDCT) can reduce the mortality from lung cancer, but patient adherence and LDCT capacity represent considerable challenges.

The selection criteria commonly applied to LDCT screening programs center around age and tobacco consumption resulting in a large number of individuals eligible for screening. A more personalized approach could reduce the resources required for a lung cancer screening program. Smoking is the single greatest risk factor for developing lung cancer, but the damaging effect can vary between individuals. The methylation-level of the AHRR gene was found to be related to the risk of developing lung cancer. Artificial intelligence (AI) is another promising approach to risk evaluation, and a machine learning model based on clinical data and standard blood tests developed by Danish researchers can be used to predict the risk of lung cancer.

The present project aims to investigate the feasibility of blood sample and AI-based risk stratification for lung cancer in patients with COPD treated and followed in general practice.

A thousand patients with COPD will be enrolled by general practitioners located in the general Vejle area in the Region of Southern Denmark. Consenting patients will fill out basic clinical data in an online REDCap database, and then they will have the blood sample collected by a healthcare professional at the general practice clinic. The sample will be transported to the laboratory at Lillebaelt Hospital, Vejle, for analysis.

A collaborative group at Lillebaelt Hospital Vejle will perform the risk stratification including analyzing DNA methylation and running the AI algorithm. Patients with a score indicating increased risk of lung cancer will be referred for LDCT.

The project will evaluate both feasibility, safety, economy and the experiences of the participants and health care professionals.

• Study Type: Interventional
• Enrollment (Estimated): 1000
• Phase: Not Applicable

Contacts and Locations

• Study Contact: Name: Sara Witting Christensen WC Wen, MD, PhD; Phone Number: +45 79406511; Email: sara.witting.christensen.wen@rsyd.dk
• Study Contact Backup: Name: Lene Horsted, Study nurse; Phone Number: +45 79409946; Email: breathe@rsyd.dk

Study Locations

• Denmark
  • Vejle, Denmark, 7100
    • Recruiting
    • Lillebaelt Hospital Vejle, University Hospital of Southern Denmark
    • Contact: Sara Witting Christensen WC Wen, MD, PhD; Phone Number: +45 79406511; Email: sara.witting.christensen.wen@rsyd.dk
    • Contact: Lene Horsted, Study nurse; Phone Number: +45 79409946; Email: breathe@rsyd.dk
  • Vejle, Denmark
    • Recruiting
    • General practices, Vejle area
    • Contact: Sara Witting Christensen Wen, MD, PhD; Phone Number: +45 79409946; Email: breathe@rsyd.dk

Participation Criteria

Eligibility Criteria

• Ages Eligible for Study: Adult; Older Adult
• Accepts Healthy Volunteers: No

Description

Inclusion Criteria:

• Diagnosed with COPD.
• => 50 years.
• Former or current smoker.
• Speaks and understands Danish.
• Able to give informed consent to participation.

Exclusion Criteria:

• Had a CT scan of the thorax within 6 months.
• Received active treatment for cancer within one year (except non-melanoma skin cancer and carcinoma in situ cervicis uteri).
• Diagnosed with cancer within one year (except non-melanoma skin cancer and carcinoma in situ cervicis uteri).
• Presents with symptoms giving suspicion of cancer (except non-melanoma skin cancer and carcinoma in situ cervicis uteri).
• In a condition not allowing diagnostic workup for or treatment of lung cancer.
• Does not have Eboks (electronic communication with Danish authorities).

Study Plan

How is the study designed?

Design Details

• Primary Purpose: Other
• Allocation: N/A
• Interventional Model: Single Group Assignment
• Masking: None (Open Label)

Number of Arms

1

Arms and Interventions

Participant Group / Arm

Intervention / Treatment

Experimental: Risk stratification; Risk stratification for lung cancer using standard blood tests, machine learning and a DNA methylation analysis.

Other: Risk stratification; Patients with COPD will have their risk of lung cancer evaluated using a machine learning model incorporating clinical data and standard blood tests as well as a DNA methylation biomarker. If the risk of lung cancer is above the cut-off, the patient will be referred for a low-dose CT scan of the chest. Currently smoking patients will be referred for a smoking cessation program.

What is the study measuring?

Primary Outcome Measures

Outcome Measure	Measure Description	Time Frame
The fraction of patients consenting to participate in the study.	The fraction of patients consenting to participate in the study.	2 years

Secondary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Number of low-dose CT scans performed	The total number of low-dose CT scans performed in the study	2 years
Number of correctly identified lung cancer cases	The number of correctly identified lung cancer cases when evaluated by the machine learning model, the DNA methylation biomarker, the PLCOm2012 model, and the USPSTF lung cancer screening criteria.	Up to 8 years
Number of lung cancer cases	The total number of lung cancer cases identified during the study and during 6 years of subsequent follow-up.	Up to 8 years
Stage distribution of lung cancer cases	The number of lung cancer cases identified within each stage from I-IV.	Up to 8 years
Number of patients with incidental findings on low-dose CT	The total number of patients with an incidental finding on the low-dose CT scan requiring treatment or further diagnostic procedures.	2 years
Number of patients without malignant disease who undergo invasive diagnostic procedures	The number of patients who undergo invasive diagnostic procedures who do not have a lung cancer diagnosis at one and two years of follow-up.	Up to 4 years
Number of adverse events	The number of adverse events in the form of pneumothorax, bleeding, infection and hospital admission.	2 years
Number of patients who initiate smoking cessation	The number and fraction of active smokers initiating and maintaining a smoking cessation program.	Up to 4 years
The fraction of participants who adhere to the study protocol	The fraction of participants who have the blood sample drawn, and when applicable, the fraction of referred participants who undergo low-dose CT.	2 years
Health economic consequences	The estimated health economic consequences of implementing AI and DNA methylation-based risk stratification in a primary healthcare setting including an estimation of the extra workload placed in the primary healthcare sector. A cost-utility analysis will calculate the incremental quality-adjusted life years (QALYs) gained by the program.	Up to 8 years
Differences in World Health Organization Five Well-being Index (WHO-5) score	Differences in World Health Organization Five Well-being Index (WHO-5) score between patients with and without increased risk of lung cancer after 1 month and 12 months. The scale minimum is 0 and the maximum is 100. A higher score indicates a better outcome.	Up to 3 years
Differences in Anxiety Symptom Scale 2 (ASS-2) score	Differences in Anxiety Symptom Scale 2 (ASS-2) score between patients with and without increased risk of lung cancer after 1 month and 12 months. The scale minimum is 0 and the maximum is 10. A lower score indicates a better outcome.	Up to 3 years
Differences in Major Depression Inventory 2 (MDI-2) score	Differences in Major Depression Inventory 2 (MDI-2) score between patients with and without increased risk of lung cancer after 1 month and 12 months. The scale minimum is 0 and the maximum is 10. A lower score indicates a better outcome.	Up to 3 years
Differences in EQ-5D-5L (quality of life) score	Differences in EQ-5D-5L score between patients with and without increased risk of lung cancer after 1 month and 12 months. Each of the five domains in the scale has a minimum of 1 and the maximum of 5. A lower score indicates a better outcome. The visual analog scale has a minimum of 0 and a maximum of 100. A higher score indicates a better outcome.	Up to 3 years

Other Outcome Measures

Outcome Measure	Measure Description	Time Frame
Qualitative analysis of interview data	Selected participants and general practitioner clinics will be invited to participate in interviews. They will inform about their experiences with the study, the intervention, the setting, the logistics and other thoughts, considerations and barriers they might have.	3 years.

Collaborators and Investigators

• Sponsor: Vejle Hospital
• Investigators: Study Chair: Ole Hilberg, MD, DMSc, Department of Medicine, Lillebaelt Hospital Vejle, University Hospital of Southern Denmark; Principal Investigator: Sara Witting Christensen Wen, MD, PhD, Department of Biochemistry and Immunology, Lillebaelt Hospital Vejle, University Hospital of Southern Denmark

Study record dates

Study Major Dates

• Study Start (Actual): May 26, 2026
• Primary Completion (Estimated): April 1, 2028
• Study Completion (Estimated): April 1, 2034

Study Registration Dates

• First Submitted: April 20, 2026
• First Submitted That Met QC Criteria: April 20, 2026
• First Posted (Actual): April 27, 2026

Study Record Updates

• Last Update Posted (Actual): June 3, 2026
• Last Update Submitted That Met QC Criteria: June 2, 2026
• Last Verified: June 1, 2026

More Information

Terms related to this study

• Keywords: COPD; Lung cancer; Biomarkers; Risk stratification; Machine learning
• Additional Relevant MeSH Terms: Pathologic Processes; Neoplasms by Site; Neoplasms; Chronic Disease; Disease Attributes; Respiratory Tract Diseases; Lung Diseases; Lung Diseases, Obstructive; Respiratory Tract Neoplasms; Thoracic Neoplasms; Pathological Conditions, Signs and Symptoms; Pulmonary Disease, Chronic Obstructive; Lung Neoplasms; Disease
• Other Study ID Numbers: BREATHE; 26/9896 (Other Identifier: Region of Southern Denmark)

Plan for Individual participant data (IPD)

• Plan to Share Individual Participant Data (IPD)?: NO

Drug and device information, study documents

• Studies a U.S. FDA-regulated drug product: No
• Studies a U.S. FDA-regulated device product: No