Evaluation of AI-assisted LDCT Screening in Lung Cancer

Taipei Veterans General Hospital

Status: Not yet enrolling

Conditions: Lung Cancer

Treatments: Diagnostic Test: AI-assisted Low-Dose CT Interpretation

Study type: Interventional

Funder types: Other

Identifiers

NCT07280559

2025-11-002AU-3AI

Details and patient eligibility

About

This multicenter pragmatic randomized controlled trial evaluates whether AI-assisted interpretation of low-dose CT (LDCT) improves lung cancer screening performance compared with standard reading. Eligible participants are randomized to AI-assisted or conventional interpretation. The study assesses diagnostic accuracy, efficiency, lung cancer incidence, mortality, recurrence, and smoking cessation outcomes. Results will inform the clinical utility and potential implementation of AI-assisted LDCT in routine screening practice.

Full description

Lung cancer is a leading cause of cancer-related mortality worldwide, and early detection is essential for improving survival. Low-dose computed tomography (LDCT) has been shown to reduce lung cancer mortality in high-risk populations, but image interpretation is time-consuming and may lead to overdiagnosis. Artificial intelligence (AI)-assisted diagnostic tools offer the potential to improve accuracy and efficiency in LDCT-based lung cancer screening, though challenges related to model adaptability, data heterogeneity, user trust, and regulatory compliance remain.

This multicenter pragmatic randomized controlled trial evaluates the effectiveness of AI-assisted LDCT interpretation compared with standard interpretation. Eligible participants will be randomized to an AI-assisted arm or a standard-reading arm. Outcomes include diagnostic accuracy, efficiency, lung cancer incidence, lung cancer mortality, recurrence, and smoking cessation.

The findings will provide evidence on the clinical utility of AI-assisted LDCT screening and support future implementation in routine practice and policy development.

Enrollment: 1,120 estimated patients

Sex: All

Ages: 40 to 74 years old

Volunteers: Accepts Healthy Volunteers

Inclusion criteria

1. Family History of Lung Cancer: Males aged 45-74 or females aged 40-74 with first-degree relatives (parents, siblings, or children) diagnosed with lung cancer.
2. Heavy Smoking History: Ages 50-74 with ≥20 pack-years smoking history, currently smoking or quit <15 years ago.
3. General Screening Participants: Adults aged 40 years or older attending routine health check-ups.

Exclusion criteria

1. Pregnant women.
2. Chest CT or other higher-radiation chest imaging within the past 12 months.
3. Individuals holding a major illness certificate for lung cancer.
4. Inability to undergo thoracic puncture or surgery.
5. Inability to hold breath or otherwise complete the scanning procedure.
6. Hemoptysis of unknown cause within the past month.
7. Chest X-ray within the past month showing suspicious lung lesions.
8. Unexplained weight loss >6 kg within the past year.
9. History of lung cancer within the past three years.
10. Presence of other severe diseases with an expected life expectancy <5 years.

Trial design

Primary purpose: Screening

Allocation: Randomized

Interventional model: Parallel Assignment

Masking: Single Blind

---

1,120 participants in 2 patient groups

AI-assisted LDCT Interpretation

Experimental group

Treatment:

Diagnostic Test: AI-assisted Low-Dose CT Interpretation

Standard Radiologist LDCT Interpretation

No Intervention group

Description:

Participants undergo low-dose computed tomography (LDCT) lung cancer screening. Images are interpreted solely by radiologists without AI assistance.