Lung Nodule Malignancy Classification Based ON NLSTx Data

978-1-5386-9330-8

1-5386-9330-5

IEEE. 2020; Pages pp. 1870-1874

Benjamin Veasey , Medical Imaging Lab, University of Lousiville, Louisville, KY, USA , M. Mehdi Farhangi , U.S. Food and Drug Administration, Silver Spring, MD, USA , Hichem Frigui , Multimedia Lab, University of Lousiville, Louisville, KY, USA , Justin Broadhead , University of Lousiville, Louisville, KY, USA , Michael Dahle , University of Lousiville, Louisville, KY, USA , Aria Pezeshk , U.S. Food and Drug Administration, Silver Spring, MD, USA , Albert Seow , University of Lousiville, Louisville, KY, USA , Amir A. Amini , Medical Imaging Lab, University of Lousiville, Louisville, KY, USA

While several datasets containing CT images of lung nodules exist, they do not contain definitive diagnoses and often rely on radiologists' visual assessment for malignancy rating. This is in spite of the fact that lung cancer is one of the top three most frequently misdiagnosed diseases based on visual assessment. In this paper, we propose a dataset of difficult-to-diagnose lung nodules based on data from the National Lung Screening Trial (NLST), which we refer to as NLSTx. In NLSTx, each malignant nodule has a definitive ground truth label from biopsy. Herein, we also propose a novel deep convolutional neural network (CNN) / recurrent neural network framework that allows for use of pre-trained 2-D convolutional feature extractors, similar to those developed in the ImageNet challenge. Our results show that the proposed framework achieves comparable performance to an equivalent 3-D CNN while requiring half the number of parameters.