18F-fluorodeoxyglucose positron emission tomography (18FDG-PET) for patients with biliary tract cancer: Systematic review and meta-analysis

•Role of 18FDG-PET in diagnosis (T), staging (N/M) and relapse of BTC was assessed. •18FDG-PET is not recommended for diagnosis (T) in the absence of cytology/histology. •18FDG-PET should be incorporated into current guidelines for staging (N/M) and relapse. •18FDG-PET should be used for staging (N/M) if identification of occult sites of disease will alter management. •18FDG-PET should be used to identify relapse if suspicion remains following standard imaging. Background & Aims The role of 18F-fluorodeoxyglucose positron emission tomography (18FDG-PET) in the diagnosis and staging of patients with biliary tract cancers (BTCs) remains controversial, so we aimed to provide robust information on the utility of 18FDG-PET in the diagnosis and management of BTC. Methods This systematic review and meta-analysis explored the diagnostic test accuracy of 18FDG-PET as a diagnostic tool for diagnosis of primary tumour, lymph node invasion, distant metastases and relapsed disease. Subgroup analysis by study quality and BTC subtype were performed. Changes in management based on 18FDG-PET and impact of maximum standardised uptake values (SUVmax) on prognosis were also assessed. A random effects model was used for meta-analyses. Results A total of 2,125 patients were included from 47 eligible studies. The sensitivity (Se) and specificity (Sp) of 18FDG-PET for the diagnosis of primary tumour were 91.7% (95% CI 89.8–93.2) and 51.3% (95% CI 46.4–56.2), respectively, with an area under the curve (AUC) of 0.8668. For lymph node invasion, Se was 88.4% (95% CI 82.6–92.8) and Sp was 69.1% (95% CI 63.8–74.1); AUC 0.8519. For distant metastases, Se was 85.4% (95% CI 79.5–90.2) and Sp was 89.7% (95% CI 86.0–92.7); AUC 0.9253. For relapse, Se was 90.1% (95% CI 84.4–94.3) and Sp was 83.5% (95% CI 74.4–90.4); AUC 0.9592. No diagnostic threshold effect was identified. Meta-regression did not identify significant sources of heterogeneity. Sensitivity analysis revealed no change in results when analyses were limited to studies with low risk of bias/concern. The pooled proportion of change in management was 15% (95% CI 11–20); the majority (78%) due to disease upstaging. Baseline high SUVmax was associated with worse survival (pooled hazard ratio of 1.79; 95% CI 1.37–2.33; p <0.001). Conclusions There is evidence to support the incorporation of 18FDG-PET into the current standard of care for the staging (lymph node and distant metastases) and identification of relapse in patients with BTC to guide treatment selection; especially if the identification of occult sites of disease would change management, or if diagnosis of relapse remains unclear following standard of care imaging. The role for diagnosis of the primary tumour remains controversial due to low sensitivity and 18FDG-PET should not be considered as a replacement for pathological confirmation in this setting. Lay summary A positron emission tomography (PET scan), using 18F-fluorodeoxyglucose (18FDG), can help doctors identify areas of cancer in the body by highlighting “hot spots”. These hotspots may be cancerous (true positive) but may also be non-cancerous, like inflammation (false positive). We show that PET scans are useful to assess how far advanced the cancer is (by assessing spread to lymph glands and to other organs) and also to identify if the cancer has recurred (for example after surgery), thus helping doctors to make treatment decisions. However, a biopsy is still needed for the initial diagnosis of a biliary tract cancer, because of the high chance of a “false positive” with PET scans. The role of 18F-fluorodeoxyglucose positron emission tomography (18FDG-PET) in the diagnosis and staging of patients with biliary tract cancers (BTCs) remains controversial, so we aimed to provide robust information on the utility of 18FDG-PET in the diagnosis and management of BTC. This systematic review and meta-analysis explored the diagnostic test accuracy of 18FDG-PET as a diagnostic tool for diagnosis of primary tumour, lymph node invasion, distant metastases and relapsed disease. Subgroup analysis by study quality and BTC subtype were performed. Changes in management based on 18FDG-PET and impact of maximum standardised uptake values (SUVmax) on prognosis were also assessed. A random effects model was used for meta-analyses. A total of 2,125 patients were included from 47 eligible studies. The sensitivity (Se) and specificity (Sp) of 18FDG-PET for the diagnosis of primary tumour were 91.7% (95% CI 89.8–93.2) and 51.3% (95% CI 46.4–56.2), respectively, with an area under the curve (AUC) of 0.8668. For lymph node invasion, Se was 88.4% (95% CI 82.6–92.8) and Sp was 69.1% (95% CI 63.8–74.1); AUC 0.8519. For distant metastases, Se was 85.4% (95% CI 79.5–90.2) and Sp was 89.7% (95% CI 86.0–92.7); AUC 0.9253. For relapse, Se was 90.1% (95% CI 84.4–94.3) and Sp was 83.5% (95% CI 74.4–90.4); AUC 0.9592. No diagnostic threshold effect was identified. Meta-regression did not identify significant sources of heterogeneity. Sensitivity analysis revealed no change in results when analyses were limited to studies with low risk of bias/concern. The pooled proportion of change in management was 15% (95% CI 11–20); the majority (78%) due to disease upstaging. Baseline high SUVmax was associated with worse survival (pooled hazard ratio of 1.79; 95% CI 1.37–2.33; p <0.001). There is evidence to support the incorporation of 18FDG-PET into the current standard of care for the staging (lymph node and distant metastases) and identification of relapse in patients with BTC to guide treatment selection; especially if the identification of occult sites of disease would change management, or if diagnosis of relapse remains unclear following standard of care imaging. The role for diagnosis of the primary tumour remains controversial due to low sensitivity and 18FDG-PET should not be considered as a replacement for pathological confirmation in this setting.