Fig 1.
Prisma flow chart of database search.
Table 1.
Bias risk in selected studies.
Table 2.
Summary of selected studies.
Table 3.
Likelihood ratios for each presentation where indicted in selected studies.
Fig 2.
Forest plots with pooled diagnostic odds ratio (95% confidence interval) and weights calculated using a random effects model for haemoptysis in the diagnosis of lung cancer.
Fig 3.
Forest plots with pooled diagnostic odds ratio (95% confidence interval) and weights calculated using a random effects model for Dyspnoea in the diagnosis of lung cancer.
Fig 4.
Forest plots with pooled diagnostic odds ratio (95% confidence interval) and weights calculated using a random effects model for Cough in the diagnosis of lung cancer.
Fig 5.
Forest plots with pooled diagnostic odds ratio (95% confidence interval) and weights calculated using a random effects model for Chest Pain in the diagnosis of lung cancer.
Fig 6.
Summary Receiver Operator Curve for Haemoptysis as a diagnostic symptom in lung cancer.
Fig 7.
Summary Receiver Operator Curve for Dyspnoea as a diagnostic symptom in lung cancer.
Fig 8.
Summary Receiver Operator Curve for Cough as a diagnostic symptom in lung cancer.
Fig 9.
Summary Receiver Operator Curve for Chest pain as a diagnostic symptom in lung cancer.
Table 4.
Positive Likelihood Ratios (LR) for symptoms in Lung Cancer patients prior to diagnosis.
Table 5.
Clinical case analysis using prior risk assessment and Bayesian incorporation of clinical symptoms to determine posterior risk.