Image Cytometry for Identification of High-Risk Patients with Barrett's Esophagus: A Prospective Multicenter Study

Abstract

Background and Aims: Flowcytometry on biopsy specimens has been previously shown to identify patients with Barrett's esophagus at high risk for progression to cancer. Patients with aneuploidy or increased 4N have a 5-yr cumulative cancer incidence of 28% (Am J Gastroenterol 2000;95:1669-1676). Using endoscopic brush cytology we studied the additional diagnostic benefit of Image Cytometry (ICM) of specialised intestinal metaplasia (SIM) compared to the histological diagnosis obtained from two separate endoscopies. Primary endpoints were detection of early cancer (EC) and high-grade dysplasia (HGD), secondary endpoint was low-grade dysplasia (LGD). Methods: 200 patients (73% male; mean age 67 yrs) with Barrett's esophagus (BE, mean segment length: 4.6 cm) were included, of whom 61% underwent two endoscopies (total of 325 examinations). In total, ICMs were obtained in 240 endoscopies from 162 patients. ICM was performed with a CCD 3-chip color videocamera Sony and AutoCyte QUIC DNA workstation on Feulgen restained cytology smears. DNA patterns were defined as diploid (2c), intermediate (diploid stemline and >10% cells below 4.5c threshhold) and aneuploid (>15% aneuploid cells or >3 cells > 9c). Results: 20 (8.3%) of 240 ICM examinations showed aneuploidy with the corresponding histological diagnosis of 3 EC (15%) , 8 HGD (40%), 5 LGD (30%), 1 indefinite for dysplasia (5%) and 3 SIM (15%) without dysplasia. 21 (8.8%) intermediate DNA-patterns at ICM were recorded at 2 endoscopies with EC, 3 endoscopies with HGD and 16 with indefinite for dysplasia. 199 (82.9%) ICMs showed diploidy in patients with EC (n = 1, 0.5%), HGD (n = 3, 1.5%), LGD (n = 10, 5%), indefinite for dysplasia (n = 20, 10%) and SIM (n = 164, 82.9%). Thus, ICM detected 25 additional high risk constellations in endoscopies with histology of LGD, indefinite for dysplasia (ID) and SIM or 1 in every 10 endoscopies with low-risk histology from modified gold standard four-quadrant biopsies. Conclusions: The data suggest that in this population of patients referred to a tertiary center, combined brush cytology and image cytometry in Barrett's esophagus identifies 1 additional high risk patients with aneuploid and intermediate DNA-pattern in every 10 endoscopies. Prospective studies need to demonstrate that this additional information affects patient management in a way ultimately improving outcomes. Background and Aims: Flowcytometry on biopsy specimens has been previously shown to identify patients with Barrett's esophagus at high risk for progression to cancer. Patients with aneuploidy or increased 4N have a 5-yr cumulative cancer incidence of 28% (Am J Gastroenterol 2000;95:1669-1676). Using endoscopic brush cytology we studied the additional diagnostic benefit of Image Cytometry (ICM) of specialised intestinal metaplasia (SIM) compared to the histological diagnosis obtained from two separate endoscopies. Primary endpoints were detection of early cancer (EC) and high-grade dysplasia (HGD), secondary endpoint was low-grade dysplasia (LGD). Methods: 200 patients (73% male; mean age 67 yrs) with Barrett's esophagus (BE, mean segment length: 4.6 cm) were included, of whom 61% underwent two endoscopies (total of 325 examinations). In total, ICMs were obtained in 240 endoscopies from 162 patients. ICM was performed with a CCD 3-chip color videocamera Sony and AutoCyte QUIC DNA workstation on Feulgen restained cytology smears. DNA patterns were defined as diploid (2c), intermediate (diploid stemline and >10% cells below 4.5c threshhold) and aneuploid (>15% aneuploid cells or >3 cells > 9c). Results: 20 (8.3%) of 240 ICM examinations showed aneuploidy with the corresponding histological diagnosis of 3 EC (15%) , 8 HGD (40%), 5 LGD (30%), 1 indefinite for dysplasia (5%) and 3 SIM (15%) without dysplasia. 21 (8.8%) intermediate DNA-patterns at ICM were recorded at 2 endoscopies with EC, 3 endoscopies with HGD and 16 with indefinite for dysplasia. 199 (82.9%) ICMs showed diploidy in patients with EC (n = 1, 0.5%), HGD (n = 3, 1.5%), LGD (n = 10, 5%), indefinite for dysplasia (n = 20, 10%) and SIM (n = 164, 82.9%). Thus, ICM detected 25 additional high risk constellations in endoscopies with histology of LGD, indefinite for dysplasia (ID) and SIM or 1 in every 10 endoscopies with low-risk histology from modified gold standard four-quadrant biopsies. Conclusions: The data suggest that in this population of patients referred to a tertiary center, combined brush cytology and image cytometry in Barrett's esophagus identifies 1 additional high risk patients with aneuploid and intermediate DNA-pattern in every 10 endoscopies. Prospective studies need to demonstrate that this additional information affects patient management in a way ultimately improving outcomes.