A multicentre prospective comparison of magnet-controlled upper GI capsule endoscopy and oesophagogastroduodenoscopy in the detection of lesions with bleeding potential in patients with iron deficiency anaemia

Aims Preliminary studies suggest that MAgnet-controlled upper GI Capsule Endoscopy (MACE), a novel, non-invasive technology, approaches the diagnostic sensitivity of OGD [1]. This study aimed to compare the two modalities in the diagnosis of lesions with bleeding potential.

Methods Patients with IDA from Sheffield (UK), Székesfehérvár (Hungary) and Hong Kong underwent MACE [2] followed by OGD within a week, performed by an endoscopist blinded to the results. MACE reporters recorded if each landmark was seen adequately (sufficient to identify significant pathology). Images of all lesions were described from the same pre-defined diagnostic register and according to a modified Saurin classification (P0, P1 or P2: unlikely, possible or likely bleeding lesions respectively) by a three member consensus panel.

Results Mean MACE examination time in 158 patients was 19.8 minutes (SD+/-7.46 min). Adequate visualisation rate of oesophagus, stomach and duodenum was 66.9% (95% CI 62.6-71), 94.1% (95% CI 92.5%- 95.5%), 78.5% (95% CI 92.5- 95.5). In 55 (34.8%) patients inadequate views were because of rapid transit or inadequate control and in 13 (8.2%) patients, views were obscured by luminal content.

In the upper GI tract 92 (58.2%) patients had 139 P1/P2 lesions identified by the consensus panel from MACE and/or OGD examinations. MACE detected 119 lesions versus 57 lesions by OGD in total (95% CI 0.0- 0.08; P<0.001) and more patients with pathology compared to OGD (89 vs 44; 95% CI 6.2- 64.0; P<0.0001). MACE identified 37 (92.5%) P2 lesions and OGD identified 12 (30.0%; 95% CI 0.46-0.78, P<0.0001) P2 lesions in the upper GI tract.

For oesophageal lesion detection, OGD and MACE had sensitivities of 50% and 62.5%, specificities of 100% and 100%, positive predictive values of 100% and 100% and negative predictive values of 90.6% and 92.7% respectively. For gastric pathologies, OGD and MACE had sensitivities of 39.1% and 96.5%, specificities of 100% and 100%, positive predictive values of 100% and 100% and negative predictive values of 49.5% respectively. For duodenal bulb lesions OGD and MACE had sensitivities of 39.3% and 71.4%%, specificities of 100% and 100%, positive predictive values of 100% and negative predictive values of 86.7% and 93.3% respectively.

Conclusions MACE identified more total and P2 lesions than OGD. Although visualisation of oesophagus and duodenum by MACE was less often adequate than in the stomach, it was better than OGD. Although currently more costly, [3] it is safer, preferred by patients, [2] an automated examination reviewed using AI is already feasible⁴ and it can be performed in the community.

Publication History

Article published online:
27 March 2025

• References

• 1 Tai FWD. et al. Cost comparison of oral, transnasal and magnet assisted capsule endoscopy in the examination of the upper gastrointestinal tract of patients with dyspepsia. Frontline Gastroenterol 2022; 14: 300-305
• 2 Xiao Y-F. et al. Fully automated magnetically controlled capsule endoscopy for examination of the stomach and small bowel: a prospective, feasibility, two centre study. Lancet Gastroenterol Hepatol 2021; 6: P914-921
• 3 Tai FWD. et al. Comparison of patient tolerability and acceptability of magnet-controlled capsule endoscopy and flexible endoscopy in the investigation of dyspepsia. Endosc Int Open 2022; 10: E735-E744