Performance of reduced- versus conventional-dose computed tomography in detection of parenchymal lung abnormalities: a systematic review protocol

OBJECTIVE: This review will examine how reduced-dose computed tomography (CT; ≤1.5mSv), compared with conventional-dose CT (>1.5mSv), performs in detecting parenchymal lung abnormalities in humans and phantoms undergoing chest CT imaging. INTRODUCTION: Deliberate exposure to ionizing radiation during medical investigations, such as chest CT, should be as low as reasonably practicable to reduce the risk of inducing cancer. Certain populations are at particular risk, such as those undergoing lung cancer screening and workers exposed to dust, and require multiple scans over time with high baseline risk. It is unclear whether reduced-dose CT scans provide sufficient image quality to accurately identify relevant lung abnormalities. ELIGIBILITY CRITERIA: Crossover studies including a non-contrast, reduced-dose CT scan (≤1.5mSv or 107mGy·cm) compared to conventional-dose CT, and reporting on at least 1 parenchymal lung abnormality will be considered. Comparison to only chest x-ray will be excluded. METHODS: The proposed systematic review will be conducted in accordance with Cochrane methodology for systematic reviews of diagnostic test accuracy, incorporating additional quantitative performance measures where applicable, and reported in line with Preferred Reporting Items for Systematic Reviews and Meta-Analysis for Diagnostic Test Accuracy (PRISMA-DTA). MEDLINE (Ovid), Embase (Ovid), and Scopus will be searched for articles published from January 1, 2014, to March 31, 2026. Prospective within-subject crossover studies will be considered, and systematic reviews will be screened for eligible primary studies. Risk of bias will be evaluated using the Quality Assessment of Diagnostic Accuracy Studies-2 (QUADAS-2) tool, with adaptation as required. Data on study and participant characteristics, CT acquisition, and performance outcomes (including diagnostic accuracy, agreement, and image quality metrics) will be extracted. Data will be synthesized using meta-analysis, where feasible, or narratively. REVIEW REGISTRATION: PROSPERO CRD42025631825.