Abstract
The COVID-19 pandemic caused a significant backlog in the UK National Health Service (NHS), with 7.6 million patients awaiting consultant appointments as of August 2024. A major consequence of this was a surge in urgent suspected cancer (USC) referrals, which rose by 27% compared to pre-pandemic levels without a proportional increase in cancer detection rates. General Practitioners (GPs) often default to USC referrals, bypassing initial investigations, which strains secondary care services.
C the Signs, an AI-led cancer prediction platform, supports GPs to improve cancer triage by identifying high-risk patients for appropriate pathways (such as direct access diagnostics or non-urgent referrals) without increasing referral volumes. This system optimizes USC referrals, reserving them for patients truly requiring intensive diagnostic resources, reducing pressure on secondary care.
From January 2020 to September 2024, data from 1, 084 GP practices using C the Signs were analyzed, covering 235, 662 risk assessments. The platform stratified patients into categories: urgent admission, USC referral, non-urgent referral, and direct access diagnostics (e.g. blood tests, FIT, X-ray). Patients not at risk were safely excluded from cancer pathways.
C the Signs redirected 21.7% of patients away from the USC pathway, safely excluding 21, 449 (9.1%), triaging 18, 063 (7.7%) to tests and diagnostics, and 11, 587 (4.9%) to non-urgent referrals—reducing USC pathway pressure by 61, 099 patients. Meanwhile, 184, 487 patients (78.3%) were appropriately referred for USC, ensuring focus on high-risk cases. Only 1.5% of assessments resulted in multiple USC referrals, and just 76 patients (0.0%) required urgent admission.
A total of 13, 585 cancers were diagnosed, yielding a 7.3% conversion rate—20.9% higher than the NHS England average of 6.0%.
C the Signs effectively stratifies suspected cancer patients, excluding low-risk individuals from USC pathways and directing others to appropriate care. This enhances NHS resource efficiency and supports elective recovery while maintaining diagnostic safety. If implemented nationally, C the Signs could reduce USC referrals by 505, 196 (17.5%) without compromising safety or outcomes. Longitudinal studies could further assess the platform’s impact on healthcare outcomes and efficiency.
Miles Payling, Seema Dadhania, Michael Moss, Judith Gordon, Bea Bakshi. Implementing an AI triage platform for identification and management of patients suspected of cancer risk to improve elective recovery post pandemic [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 2518.
RSS Feeds