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

• Published in: Medical physics (Lancaster) 2021-02, Vol.48 (2), p.733-744
• Main Authors: Zheng, Sunyi, Cornelissen, Ludo J., Cui, Xiaonan, Jing, Xueping, Veldhuis, Raymond N. J., Oudkerk, Matthijs, Ooijen, Peter M. A.
• Format: Article
• Language: English
• Subjects: computed tomography; computer‐aided detection; convolutional neural network; deep learning; Humans; Lung; Lung Neoplasms - diagnostic imaging; Neural Networks, Computer; pulmonary nodule detection; QUANTITATIVE IMAGING AND IMAGE PROCESSING; Radiographic Image Interpretation, Computer-Assisted; Solitary Pulmonary Nodule - diagnostic imaging; Tomography, X-Ray Computed
• ISSN: 0094-2405; 2473-4209
• DOI: 10.1002/mp.14648

Purpose Early detection of lung cancer is of importance since it can increase patients’ chances of survival. To detect nodules accurately during screening, radiologists would commonly take the axial, coronal, and sagittal planes into account, rather than solely the axial plane in clinical evaluation. Inspired by clinical work, the paper aims to develop an accurate deep learning framework for nodule detection by a combination of multiple planes. Methods The nodule detection system is designed in two stages, multiplanar nodule candidate detection, multiscale false positive (FP) reduction. At the first stage, a deeply supervised encoder–decoder network is trained by axial, coronal, and sagittal slices for the candidate detection task. All possible nodule candidates from the three different planes are merged. To further refine results, a three‐dimensional multiscale dense convolutional neural network that extracts multiscale contextual information is applied to remove non‐nodules. In the public LIDC‐IDRI dataset, 888 computed tomography scans with 1186 nodules accepted by at least three of four radiologists are selected to train and evaluate our proposed system via a tenfold cross‐validation scheme. The free‐response receiver operating characteristic curve is used for performance assessment. Results The proposed system achieves a sensitivity of 94.2% with 1.0 FP/scan and a sensitivity of 96.0% with 2.0 FPs/scan. Although it is difficult to detect small nodules (i.e.,