Screening for lung cancer using sub-millisievert chest CT with iterative reconstruction algorithm: image quality and...

Area of Science:

• Radiology
• Medical Imaging
• Oncology

Background:

• Lung cancer screening aims to detect nodules early for improved patient outcomes.
• Routine low-dose (LD) computed tomography (CT) is effective but involves radiation exposure.
• Optimizing imaging protocols to reduce dose while maintaining or improving image quality is crucial.

Purpose of the Study:

• To evaluate image quality and nodule detectability using an ultra-low dose (ULD) protocol with iterative model reconstruction (IMR) compared to standard LD chest CT.
• To assess the impact of different reconstruction algorithms on image noise and nodule detection in lung cancer screening.

Main Methods:

• Chest CT scans were acquired from 300 subjects using ULD (120 kVp/17 mAs) and LD (120 kVp/30 mAs) protocols.
• Images were reconstructed using filtered back projection (FBP), hybrid iterative reconstruction (HIR), and IMR algorithms.
• Image quality was assessed by radiologists, and objective image noise was measured; nodule detection rates were compared.

Main Results:

• The ULD group had a 44% lower effective dose (0.67 mSv) than the LD group (1.20 mSv).
• IMR significantly improved image quality and reduced noise compared to HIR and FBP in both ULD and LD groups.
• IMR demonstrated higher nodule detection rates than FBP and HIR, particularly for small solid nodules (<4 mm).

Conclusions:

• Iterative model reconstruction (IMR) enhances diagnostic accuracy in ultra-low dose (ULD) CT for lung cancer screening.
• IMR offers potential for improved nodule detectability, even at significantly reduced radiation doses.
• This approach represents an advancement in low-radiation imaging for lung cancer screening.