Unmet Need
Lung cancer is a type of cancer that affects lung tissues and is categorized by the presence of abnormal cell division and rapid growth of lung tissue. Globally, lung cancer is a leading cause of cancer death and poses a major therapeutic burden with poor survival rates. According to estimates from the American Cancer Society, lung cancer is the second-most commonly diagnosed cancer type and has the highest mortality rate of any cancer in the United States. According to the American Lung Association, an estimated 234,030 new lung cancer cases will be diagnosed in the United States in 2018. There is a need in the field for diagnostic imaging that can predict the onset of lung cancer where clinicians can make more informed decisions for appropriate treatment.
Technology Overview
This is a machine learning algorithm to calculate yearly personalized probabilities of lung cancer at 1-3 years after the most recent CT scan. It also provides a personalized Kaplan-Meier
curve of lung cancer free probability within 3 years from the most recent CT scan date and its point-wise 95% confidence intervals. Clinicians can use these results to determine
whether a patient needs an invasive procedure immediately (such as biopsy or surgery), or whether a conservative approach such as monitoring without any cancer treatment is appropriate. It also helps clinicians determine the optimal time interval for next clinical visit.
Stage of Development
The algorithm has been validated in a retrospective study. A simplified version of DeepLR is available for testing at https://www.caced.jhu.edu.
Publications
P. Huang, et al. "Added Value of Computer-aided CT Image Features for Early Lung Cancer Diagnosis with Small Pulmonary Nodules: A Matched Case-Control Study." Radiology 2018 286:1, 286-295.
P. Huang, et al. “Prediction of lung cancer risk at follow-up screening with low-dose CT: a training and validation study of a deep learning method.” The Lancet Digital Health (Oct. 17, 2019).
