Abstract

Selection of patients for ongoing assessment, investigation and referral (or not) from primary care, for suspected lung cancer, has to date been predicated upon as assessment of underlying risk of undiagnosed cancer as determined by presentation of particular symptoms, usually to a GP.

A lot is known about the epidemiology of lung cancer symptoms, and their presentation, and the risk of individual and multiple symptoms. There is more to understand though, for example how AI approaches my able to combine data about symptoms and changing symptoms with other data in health records (including for example test results, weight, medication and attendance patterns) to estimate cancer risk more accurately. More work is also needed to ensure that an understanding of symptom-based risk is fully implemented in practice (i.e. implementation of current NICE guidelines in England and Wales). However, there is a limit to how much symptoms can actually help in cancer diagnosis.

Diagnosing lung cancer is also complicated by the fact that chest-x-ray, the standard first-line investigation, has a relatively low specificity, meaning that up to about 20% of patients who are subsequently diagnosed with lung cancer will have had a normal chest x-ray reported in the preceding months. Normal x-rays reassures doctors and patients and are likely to delay the diagnosis.

This presentation will make the case that GPs need help in identifying those people at risk of having undiagnosed lung cancer, and equally importantly, those who are not at risk, and present some of the possibilities for the use of new technologies that may help GPs. These technologies will include: biomarkers and machine learning developed algorithms based upon multiple biomarkers (using for example, residual testing of samples), and VOCs; and determining the place and value of imaging modalities.

The presentation will build upon the work of the NIHR In-Vitro Diagnostics Cooperative in Leeds, the development of a primary care ‘testbed’ for such technologies, and the Yorkshire Lung Screening Trial that allows a platform for biomarker collection for patients at higher risk of lung cancer who are undergoing screening in community settings.