Development and clinical application of deep learning model for lung nodules screening on CT images

Abstract

Lung cancer screening based on low-dose CT (LDCT) has now been widely applied because of its effectiveness and ease of performance. Radiologists who evaluate a large LDCT screening images face enormous challenges, including mechanical repetition and boring work, the easy omission of small nodules, lack of consistent criteria, etc. It requires an efficient method for helping radiologists improve nodule detection accuracy with efficiency and cost-effectiveness. Many novel deep neural network-based systems have demonstrated the potential for use in the proposed technique to detect lung nodules. However, the effectiveness of clinical practice has not been fully recognized or proven. Therefore, the aim of this study to develop and assess a deep learning (DL) algorithm in identifying pulmonary nodules (PNs) on LDCT and investigate the prevalence of the PNs in China. Radiologists and algorithm performance were assessed using the FROC score, ROC-AUC, and average time consumption. Agreement between the reference standard and the DL algorithm in detecting positive nodules was assessed per-study by Bland-Altman analysis. The Lung Nodule Analysis (LUNA) public database was used as the external test. The prevalence of NCPNs was investigated as well as other detailed information regarding the number of pulmonary nodules, their location, and characteristics, as interpreted by two radiologists. Lung cancer screening has now been widely applied because of its effectiveness and ease of performance. More than 10 million chest CT scans were performed in the United States alone in 2012, highlighting the potential for this clinical scenario 1. Radiologists who evaluate a large number of low-dose CT (LDCT) screening images face enormous challenges, including mechanical repetition and boring work, the easy omission of small nodules, lack of consistent criteria, etc. 2-5. Artificial intelligence (AI) is showing rapid advantages and exciting achievements in the fields of imaging diagnosis and/or evaluation 6-12. AI detection of lung nodules has long been expected to be an effective assistant in daily clinical practice, especially for LDCT lung nodule screening. Thus far, many novel deep neural networkbased systems have demonstrated the potential for use in the proposed technique for helping radiologists improve nodule detection accuracy with efficiency and cost-effectiveness 13-18. However, the majority of the proposed systems were trained on CT scans from the Lung Image Database Consortium/Image Database Resource Initiative (LIDC-IDRI), the LUNA16 (Lung Nodule Analysis 2016) database and the ANODE09 (Automatic Nodule Detection 2009) database 19-24. The effectiveness in clinical practice has not been fully recognized or proven. From the previous study, the prevalence of pulmonary nodules (PNs) varies greatly in different populations, ranging from 13 to 58% 25-32. Different demographic features, selection criteria of participants, and the referral pattern of the study centre may explain such differences 27. In China, the incidence of non-c...

Deep convolutional neural networks for multiplanar lung nodule detection: Improvement in small nodule identification

“…For instance, Cui et al developed a nodule detection system using 39 014 scans from multiple centers. 29 Although the system reached a sensitivity of 93.4% with 0.8 FP/scan, a number of true nodules were still missed by the system when the FP rate was smaller than 0.5. Nevertheless, the external validation results showed the potential use of the deep learning-based system in clinical practice.…”

“…Some systems did not report sensitivities at various FP rates. For instance, Cui et al developed a nodule detection system using 39 014 scans from multiple centers 29 . Although the system reached a sensitivity of 93.4% with 0.8 FP/scan, a number of true nodules were still missed by the system when the FP rate was smaller than 0.5.…”

“…27 Besides, a vector quantization algorithm was used by Tan et al 28 to detect potential nodules and knowledge of shape, texture was infused to the FP reduction model. Additionally, to develop an efficient nodule detection system, a multicenter study with 39 014 cases was conducted by Cui et al 29 deep learning techniques.…”

Deep learning for predicting subtype classification and survival of lung adenocarcinoma on computed tomography

“…In order to ensure the average distribution of the data of each category, 200 patients were randomly selected for each category except for the extremely rare AAH and pMAC. In our research, the region of interest (ROI) detection of lung nodules were built on a deep residual network, which is able to detect target ROI automatically [24] . Then two senior radiologists (with 10 years of work experience in thoracic imaging diagnosis) blindly and independently assessed the mask of all images, and any disagreements between two observers were resolved by discussion until a consensus was reached.…”

Pulmonary Hypertension in Association with Lung Disease: Quantitative CT and Artificial Intelligence to the Rescue? State-of-the-Art Review

“…Within image analysis, machine learning in chest CT broadly has two domains: nodule detection and radiomics. Multiple AI solutions exist for lung cancer nodule detection and have shown to have a high level of accuracy, sensitivity, and specificity [ 45 , 46 , 47 ]. Labeled public datasets for model development exist, the largest of which is the Lung Image Database Consortium Image Collection (LIDC-IDRI) [ 48 ].…”