Back

Accelerated cancer registration from the National Disease Registration Service to support the NHS-Galleri trial

Eversfield, C.; Petersen, N.; Smittenaar, R.; Liang, W.; Rocha, C.; Harrop, L.; Graham, K.; Dilling, N.; Tulloch, O.; Lloyd, D.; Sasieni, P.; Rous, B.; Bomb, M.; McPhail, S.

2026-03-25 epidemiology
10.64898/2026.03.24.26345471 medRxiv
Show abstract

Purpose: High-quality population-based cancer registration data is crucial for cancer research. We describe the methodology that the National Disease Registration Service (NDRS) established to accelerate provision of cancer registration data in England to support timely analysis of the NHS-Galleri trial (NCT05611632), assessing the clinical utility of a multi-cancer early detection test for population screening. Methods: NDRS established two accelerated cancer registration products for trial participants: Expedited Core (providing complete stage data 6 months after diagnosis) and Expedited Comprehensive (providing [≥]6 months of follow-up clinical pathway data 11 months after diagnosis). NDRS used pre-existing data sources and methodology, plus enhanced dedicated data liaison support. Timeliness and concordance of accelerated data was assessed 6, 11, and [≥]19 months after diagnosis. Results: At 6-10 months after diagnosis, most (98.0%) registrations achieved Expedited Core status; at >10 months, all (100%) registrations were Expedited Comprehensive. For fact of malignant cancer and stage III/IV diagnoses, concordance between registrations at 6 and [≥]19 months was 96.7% and 95.6%, respectively. Conclusions: NDRS delivered staged cancer registration data 6 months after diagnosis. High concordance at 6, 11, and 19 months provides confidence in the accelerated data. This supports reporting NHS-Galleri findings more promptly than possible via routine data.

Matching journals

The top 7 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.