Background: Lung cancer is a leading cause of premature death. Early diagnosis is associated with improved outcomes. While chest X-ray remains the first line investigation, individual studies have indicated that a proportion of lung cancers are not identified (false negatives); this is one cause of delayed diagnoses.
Aim: To estimate the proportion of patients with lung cancer who have had a chest X-ray in the year prior to diagnosis which was not interpreted as having features suspicious for lung cancer.
Method: A broad search strategy was devised including several synonyms for chest X-ray and lung cancer. Screening and paediatric data, studies from before 1999 and studies based on retrospective interpretation of radiology were excluded. A 20% random sample screened by a second reviewer to ensure consistency.
Results: Analysis is almost complete: 8362 studies were identified. Twenty-nine studies met the inclusion criteria. Heterogeneity precluded meta-analysis, however the combined population was 1557 patients of whom 954 (61.27%) had a positive chest X-ray. In four studies with a lower risk of bias, comprising 1052 patients, sensitivity ranged from 73.5% to 79.3% (mean 76.19%).
Conclusion: This review suggests that chest X-ray does not identify approximately 25% of cases, although the findings may be affected by publication bias. This finding needs replicating in a high-quality study that could also demonstrate the potential consequences on diagnostic intervals and stage at diagnosis. The findings also suggest that studies to evaluate the effectiveness and cost-effectiveness of other modalities (e.g. low-dose CT scan) are needed.
CORE (COnnecting REpositories)
CORE (COnnecting REpositories)