Bashar J Qumseya1, Michael J Bartel2, Sherif Gendy3, Paul Bain4, Amira Qumseya5, Herbert Wolfsen6. 1. Division of Gastroenterology and Hepatology, Archbold Medical Group/Florida State University, United States. Electronic address: bqumseya@archbold.org. 2. Division of Gastroenterology, Fox Chase Cancer Center, Philadelphia, PA, United States. 3. Florida A&M University, Tallahassee, FL, United States. 4. Harvard Medical School, Boston, MA, United States. 5. Evidence-Based Research Institute, Tallahassee, FL, United States. 6. Division of Gastroenterology and Hepatology, Mayo Clinic, 4500 San Pablo Road, Jacksonville, FL, United States.
Abstract
BACKGROUND: The use of endoscopic ultrasound (EUS) to stage patients with Barrett's esophagus (BE) with suspected neoplasia is controversial due to high rates of over-staging. However, this rate of over-staging has not been adequately investigated or quantified. AIM: To determine the rate of over-staging related EUS in this population. METHODS: Search included Medline, Embase, Web of Science, and Cochrane Central ending on 9/30/2016. The primary effect-estimate of interest was the false positive rate of advanced disease on EUS at the tumor level (T1a vs. T1b). Secondary outcomes included false detection rate, false negative rate, accuracy, sensitivity, and specificity. Study heterogeneity was assessed using the I2 and Cochrane's Q. RESULTS: Of 1872 studies, 11 met our inclusion criteria totaling 895 patients. Based on random effects models, the pooled FPR for advanced disease was 9.1% ([6.5-12.5%], p<0.001). Tests of heterogeneity showed no significant heterogeneity for this outcome. The pooled false negative rate was 9.2% [95%CI: 4.7-17.3%], p<0.01. Overall, the pooled accuracy of EUS results in BE neoplasia patients was low at 74.6% [58.7-85.8%], p=0.004. CONCLUSIONS: The use of EUS in BE patients with dysplasia and early neoplasia results in a large proportion of patients falsely over-staged and under-staged.
BACKGROUND: The use of endoscopic ultrasound (EUS) to stage patients with Barrett's esophagus (BE) with suspected neoplasia is controversial due to high rates of over-staging. However, this rate of over-staging has not been adequately investigated or quantified. AIM: To determine the rate of over-staging related EUS in this population. METHODS: Search included Medline, Embase, Web of Science, and Cochrane Central ending on 9/30/2016. The primary effect-estimate of interest was the false positive rate of advanced disease on EUS at the tumor level (T1a vs. T1b). Secondary outcomes included false detection rate, false negative rate, accuracy, sensitivity, and specificity. Study heterogeneity was assessed using the I2 and Cochrane's Q. RESULTS: Of 1872 studies, 11 met our inclusion criteria totaling 895 patients. Based on random effects models, the pooled FPR for advanced disease was 9.1% ([6.5-12.5%], p<0.001). Tests of heterogeneity showed no significant heterogeneity for this outcome. The pooled false negative rate was 9.2% [95%CI: 4.7-17.3%], p<0.01. Overall, the pooled accuracy of EUS results in BE neoplasiapatients was low at 74.6% [58.7-85.8%], p=0.004. CONCLUSIONS: The use of EUS in BE patients with dysplasia and early neoplasia results in a large proportion of patients falsely over-staged and under-staged.