Table 2

Diagnostic performance of 10-fold cross-validated VOC analyses of breath samples of Barrett’s oesophagus, GORD patients and controls

Comparisons (samples)AUC (95% CI)Sensitivity (%)Specificity (%)
Total cohort (n=402)
 BO (129) vs GORD+control (273)0.91 (0.87 to 0.94)91 (84–95)74 (69–79)
 BO (129) vs GORD (141)0.73 (0.67 to 0.79)64 (55–72)74 (66–81)
 BO (129) vs control (132)0.70 (0.64 to 0.76)57 (48–66)67 (59–75)
PPI users (n=300)*
 BO (122) vs GORD+control (178)0.77 (0.71 to 0.82)68 (59–76)71 (64–78)
 BO (122) vs GORD (95)0.78 (0.72 to 0.84)75 (67–83)68 (58–77)
 BO (102) vs control (77)†0.80 (0.74 to 0.87)71 (61–79)81 (70–88)
Hiatal hernia (n=205)
 BO (102) vs GORD+control (103)0.76 (0.70 to 0.83)80 (71–87)60 (50–70)
 BO (87) vs GORD (56)‡0.83 (0.75 to 0.90)85 (75–91)75 (61–85)
 BO (81) vs control (36)§0.96 (0.91 to 1.00)96 (89–99)86 (70–95)
BO segment ≥M3 cm (n=370)
 BO (98) vs GORD+control (273)0.94 (0.92 to 0.97)96 (90–99)72 (66–76)
 BO (98) vs GORD (141)0.79 (0.74 to 0.85)66 (56–75)76 (68–83)
 BO (98) vs control (132)0.77 (0.70 to 0.83)68 (58–77)75 (67–82)
  • *Daily use of PPIs for at least 1 month before study enrolment.

  • †20 BO patients and 6 controls were excluded in the model due to <5 measurements per Aeonose device.

  • ‡15 BO patients and 7 GORD patients were excluded in the model due to <5 measurements per Aeonose device.

  • §21 BO patients and 4 controls were excluded in the model due to <5 measurements per Aeonose device.

  • AUC, area under the curve; BO, Barrett’s oesophagus; GORD, gastro-oesophageal reflux disease; PPI, proton pump inhibitor; VOC, volatile organic compound.