Reducing false positives
Lung Cancer is the most common type of cancer, with more than 200,000 new cases every year in the US alone. Current assessment methods diagnose a lot of patients with lung cancer who don’t have cancerous lesions, leading to a lot of patient anxiety and unnecessary treatment. AI can greatly improve the accuracy of the diagnoses, which means less false positives.
Faster access to life-saving treatments
An estimated 20% of lung cancer deaths could be prevented with early detection (source). Faster and more accurate assessments with AI mean more confidence and earlier treatment.
More time for patients
Better automated assessments give radiologists more time for their patients – time previously spent on additional treatment and monitoring of patients who don’t have cancer.
How does it work?
Neural Network Algorithms (in particular Convolutional Neural Networks) can be trained on low-dose, high-resolution CT images. The algorithm learns to make predictions that are as close as possible to expert annotations of the images.
An integral part of building a good prediction algorithm is preprocessing – for example, identifying the lungs in the scan: