Finding H. pylori in the Fine Print: Evidence-Linked Multi-Agent Case Finding from Gastric Biopsy Reports

2026-07-07

AI Summary

This study addresses the challenge of accurately identifying Helicobacter pylori infection and associated gastritis in gastric biopsy reports by precisely interpreting affirmative and negated contexts within unstructured clinical text—a task where traditional methods struggle to scale efficiently. The authors propose a field-driven Nimblemind multi-agent system (nMAS) that leverages an evidence-linking mechanism to automate information extraction and classification. Evaluated on 54 de-identified reports comprising 216 clinical assertions, the system achieved an accuracy of 98.61%. By innovatively integrating a multi-agent architecture with traceable evidence provenance, the approach delivers unified report-level outputs along with source sentences supporting each prediction, substantially enhancing clinical verifiability and workflow integration. Simulations demonstrated that reviewing 1,000 reports would require only 1.4 hours using nMAS, compared to 83.3 hours manually, yielding a reduction of approximately 81.9 labor hours and $6,100 in costs.

Abstract

Data from Singapore indicated that about 31% of the population had evidence of Helicobacter pylori infection. Persistent H. pylori infection is associated with chronic active gastritis and peptic ulcer disease, and its eradication is key to gastric cancer prevention. However, evidence supporting \textit{H. pylori} positivity and H. pylori-associated gastritis may be distributed across heterogeneous coded and free-text report fields and may require contextual interpretation of assertion and negation, limiting keyword search, and making manual review difficult to scale. We conducted a retrospective pilot evaluation of the Nimblemind Multi-Agent System (nMAS), a field-name-driven, evidence-linked extraction workflow, using 54 de-identified gastric biopsy pathology reports from a large healthcare system in Singapore. Four clinician-scoped binary fields were evaluated: gastric/stomach biopsy, biopsy status, H. pylori positivity, and H. pylori-associated gastritis. Across 216 feature-case decisions, nMAS correctly classified 213, corresponding to 98.61% overall accuracy. A separately implemented UMA-style MiniMax M2.5 comparator produced similar aggregate and per-field classification metrics. Although predictive performance was similar, nMAS maintained unified report-level outputs with supporting source sentences; the demonstrated contribution is therefore workflow integration and traceability rather than predictive superiority. Under an illustrative, unmeasured scenario, reviewing 1,000 reports at five minutes per manual review versus five seconds per evidence-linked verification would reduce review time from 83.3 to 1.4 staff-hours, corresponding to 81.9 staff-hours and about USD~6,100 in potential staff-time value. Larger multi-institutional studies should evaluate evidence-span correctness, clinician verification time, and generalizability.

Problem

• Helicobacter pylori
• gastric biopsy
• natural language processing
• clinical text mining
• evidence extraction

Innovation

• multi-agent system
• evidence-linked extraction
• H. pylori detection
• clinical text mining
• workflow traceability