Feasibility of Predicting Incidental Gallbladder Cancer: An Interim Analysis of Feasibility Outcomes of a Trainee-Led, Multi-Centre Diagnostic Modelling Study

Aim: Incidental gallbladder cancer (iGBC) is identified in approximately 1 in 500 cholecystectomies. Although rare, early detection is essential to improve outcomes. Existing risk stratification tools have poor sensitivity or are not validated in the UK population. The aim of Feasibility of Predicting Incidental Gallbladder Cancer (fP-iGBC) is to assess the feasibility of collecting the data required to develop a diagnostic score to stratify iGBC risk.

Method: The trainee-led Predicting Incidental Gallbladder Cancer (P-iGBC) collaborative prospectively identified adult patients undergoing cholecystectomy for benign disease. Collaborators collected routinely recorded, de-identified data on demographics, surgical pathway, pre-operative investigations and interventions, operative approach and findings, and final histopathology. To assess internal validity, a double data entry exercise was undertaken for 10% of records. Primary feasibility outcomes included recruitment rates and data completeness.

Results: Across eight sites, 551 patients were screened and 508 recruited over 594 site-days (mean 74.3 per site). Despite delays to site activation, mean screening and recruitment rates were 17.4 and 16.0 patients per site per 30 days. Six sites met the minimum recruitment target of 10.5 patients per month. All case record forms were signed off as complete. Data completeness for core variables was 98.7%. Absolute inter-rater agreement for core variables was 91.4%, with reliability rated as good or excellent in most cases.

Conclusions: fP-iGBC demonstrates that high-quality, prospective data on adult cholecystectomy can be collected by a trainee-led collaborative. Although further evaluation is required, these findings are promising and suggest that progression to a full diagnostic modelling study may be feasible.