Using an artificial intelligence platform to enhance cancer detection rates in primary care.

Abstract

1560 Background: Patients with cancer who are diagnosed in a primary care setting have the best survival outcomes compared to any other route (e.g. Emergency department, routine referrals, incidental findings etc). This is primarily attributed to detection of cancer at an early stage. In the UK, the Cancer Detection Rate (CDR) metric serves as a measure of this, representing the percentage of cancer diagnoses originating from primary care referral routes. However, detecting cancer presents significant challenges, primarily due to the non-specific nature of cancer presentations, their tendency to overlap with symptoms of other conditions, and their relatively low prevalence. Artificial intelligence (AI) platforms may be able to support this endeavor but are yet to be established. C the Signs is an AI-enabled clinical decision support tool integrated into primary care practices to help clinicians stratify the risk of cancer and advise on which cancer pathway is most appropriate (if at all) for the patient’s presentation and history. Methods: An observational cohort study was undertaken between 1st May 2021 and 31st March 2022 where 35 practices (covering a population of almost 420,000) within the same geographic region in the East of England were offered the use of C the Signs, with the practices opting out as controls. All practices had the same access to referral and cancer diagnostic pathways within the region. The primary end point was to identify if C the Signs had a statistical impact on cancer detection rates in primary care through nationally linked data via Public Health England. Results: In practices utilizing C the Signs the CDR significantly increased from 58.7% in 2020-21 (prior to the implementation of C the Signs) to 66.0% in 2021-22, reflecting a significant increase of 12.3% (p <0.05). In contrast, practices not using C the Signs maintained a stable CDR of 58.4% in both years. There was no statistically significant variance observed in the referral rate between the two groups, indicating no notable increase in diagnostic or referral activities. Conclusions: These findings underscore the importance of integrating cutting-edge technologies into primary care to improve cancer detection rates and facilitate early-stage diagnosis. Implementation of such advancements holds promise for reducing cancer-related mortality rates and enhancing overall patient outcomes.