Abstract
BackgroundReidentification of prior nodules for temporal comparison is an important but time-consuming step in lung cancer screening. We develop and evaluate an automated nodule detector that utilizes the axial-slice number of nodules found in radiology reports to generate high precision nodule predictions.Methods888 CTs from Lung Nodule Analysis were used to train a 2-dimensional (2D) object detection neural network. A pipeline of 2D object detection, 3D unsupervised clustering, false positive reduction, and axial-slice numbers were used to generate nodule candidates. 47 CTs from the National Lung Cancer Screening Trial (NLST) were used for model evaluation.ResultsOur nodule detector achieved a precision of 0.962 at a recall of 0.573 on the NLST test set for any nodule. When adjusting for unintended nodule predictions, we achieved a precision of 0.931 at a recall 0.561, which corresponds to 0.06 false positives per CT. Error analysis revealed better detection of nodules with soft tissue attenuation compared to ground glass and undeterminable attenuation. Nodule margins, size, location, and patient demographics did not differ between correct and incorrect predictions.ConclusionsUtilization of axial-slice numbers from radiology reports allowed for development of a lung nodule detector with a low false positive rate compared to prior feature-engineering and machine learning approaches. This high precision nodule detector can reduce time spent on reidentification of prior nodules during lung cancer screening and can rapidly develop new institutional datasets to explore novel applications of computer vision in lung cancer imaging.
Highlights
The National Lung Screening Trial (NLST) demonstrated that low-dose computed tomographic (CT) screening of high risk patients can result in a 20% reduction in Chillakuru et al BMC Med Imaging (2021) 21:66With over 8.6 million individuals eligible for low-dose lung cancer screening each year, 575 screens must be performed per lung cancer death avoided [5, 6]
Utilization of axial-slice numbers from radiology reports allowed for development of a lung nodule detector with a low false positive rate compared to prior feature-engineering and machine learning approaches
This high precision nodule detector can reduce time spent on reidentification of prior nodules during lung cancer screening and can rapidly develop new institutional datasets to explore novel applications of computer vision in lung cancer imaging
Summary
The National Lung Screening Trial (NLST) demonstrated that low-dose computed tomographic (CT) screening of high risk patients can result in a 20% reduction in Chillakuru et al BMC Med Imaging (2021) 21:66With over 8.6 million individuals eligible for low-dose lung cancer screening each year, 575 screens must be performed per lung cancer death avoided [5, 6]. The workflow for nodule detection and evaluation can be time consuming for a radiologist. In addition to identifying nodules on a new CT, radiologists must identify old nodules from prior scans and determine if there has been any temporal change. Despite having the nodule axial-slice number available in prior radiology reports, the process of identifying old nodules to cross-reference on the new CT is labor intensive. While advances in deep learning and computer-aided nodule detection have shown promise in nodule identification [8, 9], they do not focus on augmenting this critical aspect of the lung nodule screening workflow—locating previously identified nodules to observe changes over time by utilizing prior knowledge available in radiology reports (i.e. axial-slice location). We develop and evaluate an automated nodule detector that utilizes the axial-slice number of nodules found in radiology reports to generate high precision nodule predictions
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
