Predicting the location and interval of lung nodule occurrence from low-dose CT
Our aim is to develop and test a novel artificial intelligence (AI) architecture that combines baseline LDCT studies with other patient information such as age, smoking status, and smoking history (patient demographics) to identify the most relevant data for personalized screening intervals for patients at risk of developing lung cancer.
Our hypothesis is that apparently "normal‟ baseline LDCT scans will have certain features, perhaps not visible to the human eye, that can predict time-to-lesion-appearance of lung tumours.
We aim to identify and model a pre-cancerous signature present on baseline LDCT lung scans using our technology. Using these data, we will train a novel end-to-end deep-learning network using state-of-the-art high-performance computing, and validate our model on an independent test set randomly sampled from the available data.
Professor Xujiong Ye - University of Lincoln
Dr Reyer Zwiggelaar - University of Aberystwyth
Dr Spencer Thomas - NPL Management Ltd
Dr Balaji Ganeshan - University College London
Dr Richard Lee - Royal Marsden Hospital
Dr Carolyn Horst - University College London