10 / 172
Abstracts of the 22
nd
National Congress of Digestive Diseases / Digestive and Liver Disease 48S2 (2016) e67–e231
e71
BE, 53 with EAC and 107 controls, including 42 blood donors and
65 patients with gastroesophageal reflux but with no endoscopic/
histologic diagnosis of BE (GERD). The cut-off levels for the
determination of SCCA-IgM for BE/EAC versus controls and for
Barrett “at risk” (long and/or dysplastic BE) versus BE “at low risk”
(short non-dysplastic BE) were calculated by ROC curves. Statistics
also included Kolmogorov-Smirnov, Kruskall-Wallis, Mann-Whitney
and chi square tests. Immuno-staining for SCCA-IgMwas obtained in
a subgroup of 75 patients (Hepa-Ab, Xeptagen).
Results:
Median SCCA-IgM values were significantly higher in BE
and EAC patients than in GERD (p<0.0001). Patients with SCCA-IgM
levels higher than the cut-off calculated on the basis of a ROC curve
(56.6 AU/mL, 91.5% sensitivity, 75.4% specificity, positive predictive
value [[PPV] 85.8%, negative predictive value [NPV] 84.4%, AUC of
0.799) had a 33 times higher Relative Risk (RR) of harboring BE or
EAC (p=0.0001). BE patients “at risk” (long and/or dysplastic BE)
had SCCA-IgM levels significantly higher than those with short
non-dysplastic BE (p=0.035) and patients with SCCA-IgM above
the calculated cut-off (78.5 AU/mL, sensitivity 85%, specificity
54%, PPV=68%, NPV=76%, AUC=0.67) had a 15 times higher RR of
having Barrett “at risk”. SCCA was expressed in the proliferative
compartment of BE mucosa in 66% of the cases and in cardiac-type
gastric metaplasia only in15% (p=0.003).
Conclusions:
Serum SCCA-IgM determination allows the
identification of patients at risk for Barrett’s esophagus and
esophageal adenocarcinoma and the stratification of Barrett patients
in subgroups with increasing cancer risk. Large, prospective studies
are required to confirm this evidence in stage IV biomarker studies.