2022 World Conference on Lung Cancer (ePosters)-EP01.06-002. Impact of Immediate AI Enabled Patient Triage to Chest CT on the Lung Cancer Pathway: LungIMPACT

Pdf Summary

The LungIMPACT study aims to assess the impact of an artificial intelligence (AI) triage pathway on the time it takes to diagnose lung cancer. The study will evaluate several outcomes, including the difference in time between intervention and non-intervention days for lung cancer diagnosis, the agreement between AI triage and human readers on the classification of normal/abnormal cases, and the difference in the number of urgent referrals to respiratory medicine with a non-cancer diagnosis.<br /><br />The intervention involves using an AI triage system and prioritizing worklists for immediate reporting. A class 2 CE approved algorithm called qXR is used for analyzing chest x-rays (CXR). The study design is a prospective multicenter randomized trial conducted at eight NHS sites. The intervention is implemented on randomly selected days, and all CXRs undergo analysis using the qXR clinical decision support system. The study spans nine months and involves 107,250 consecutive CXRs.<br /><br />The primary focus of the study is to reduce the time to diagnose lung cancer, as it is a critical factor in patient outcomes. Currently, the median time for the diagnosis of lung cancer is 13 days, while the recommended time according to the English National Optimal Lung Cancer Pathway is three days. This pathway has faced challenges in implementation due to limited resources, including radiology capacity.<br /><br />The study's inclusion criteria include all consecutive CXRs referred from primary care, with patients aged 18 years or above. The sample size is estimated based on the desired statistical power for the outcomes of interest, including a 10-day reduction in time to diagnosis for 312 lung cancers and an agreement between qXR and human readers for 90,482 CXRs. The expected prevalence of lung cancer is around 6 cases per 1,000 CXRs.<br /><br />The study aims to provide valuable insights into the impact of AI triage pathways on the time it takes to diagnose lung cancer, potentially improving patient outcomes and optimizing resource utilization in radiology.

Asset Subtitle

David Baldwin

Meta Tag

Speaker David Baldwin

Topic Early Detection and Screening - Risk Stratification

Keywords

LungIMPACT study

artificial intelligence triage pathway

time to diagnose lung cancer

AI triage

qXR algorithm

chest x-rays analysis

prospective multicenter randomized trial

English National Optimal Lung Cancer Pathway

radiology capacity

patient outcomes