Purpose: Early detection of pulmonary nodules is an effective way to improve patients' chances of survival. In this work, we propose a novel and efficient way to build a computer-aided detection (CAD) system for pulmonary nodules based on computed tomography (CT) scans. Methods: The system can be roughly divided into two steps: nodule candidate detection and false positive reduction. Considering the three-dimensional (3D) nature of nodules, the CAD system adopts 3D convolutional neural networks (CNNs) in both stages. Specifically, in the first stage, a segmentation-based 3D CNN with a hybrid loss is designed to segment nodules. According to the probability maps produced by the segmentation network, a threshold method and connected component analysis are applied to generate nodule candidates. In the second stage, we employ three classification-based 3D CNNs with different types of inputs to reduce false positives. In addition to simple raw data input, we also introduce hybrid inputs to make better use of the output of the previous segmentation network.In experiments, we use data augmentation and batch normalization to avoid overfitting. Results: We evaluate the system on 888 CT scans from the publicly available LIDC-IDRI dataset, and our method achieves the best performance by comparing with the state-of-the-art methods, which has a high detection sensitivity of 97.5% with an average of only one false positive per scan. An additional evaluation on 115 CT scans from local hospitals is also performed. Conclusions: Experimental results demonstrate that our method is highly suited for the detection of pulmonary nodules.

Background: Prenatal evaluation of fetal lung maturity (FLM) is a challenge, and an effective non-invasive method for prenatal assessment of FLM is needed. The study aimed to establish a normal fetal lung gestational age (GA) grading model based on deep learning (DL) algorithms, validate the effectiveness of the model, and explore the potential value of DL algorithms in assessing FLM. Methods: A total of 7013 ultrasound images obtained from 1023 normal pregnancies between 20 and 41 + 6 weeks were analyzed in this study. There were no pregnancy-related complications that affected fetal lung development, and all infants were born without neonatal respiratory diseases. The images were divided into three classes based on the gestational week: class I: 20 to 29 + 6 weeks, class II: 30 to 36 + 6 weeks, and class III: 37 to 41 + 6 weeks. There were 3323, 2142, and 1548 images in each class, respectively. First, we performed a pre-processing algorithm to remove irrelevant information from each image. Then, a convolutional neural network was designed to identify different categories of fetal lung ultrasound images. Finally, we used ten-fold cross-validation to validate the performance of our model. This new machine learning algorithm automatically extracted and classified lung ultrasound image information related to GA. This was used to establish a grading model. The performance of the grading model was assessed using accuracy, sensitivity, specificity, and receiver operating characteristic curves. Results: A normal fetal lung GA grading model was established and validated. The sensitivity of each class in the independent test set was 91.7%, 69.8%, and 86.4%, respectively. The specificity of each class in the independent test set was 76.8%, 90.0%, and 83.1%, respectively. The total accuracy was 83.8%. The area under the curve (AUC) of each class was 0.982, 0.907, and 0.960, respectively. The micro-average AUC was 0.957, and the macro-average AUC was 0.949. Conclusions: The normal fetal lung GA grading model could accurately identify ultrasound images of the fetal lung at different GAs, which can be used to identify cases of abnormal lung development due to gestational diseases and evaluate lung maturity after antenatal corticosteroid therapy. The results indicate that DL algorithms can be used as a non-invasive method to predict FLM.

Automatic liver segmentation from abdominal computed tomography (CT) images is a fundamental task in computer-assisted liver surgery programs. Many liver segmentation algorithms are very sensitive to fuzzy boundaries and heterogeneous pathologies, especially when the data are scarce. To solve these problems, we propose an automatic liver segmentation framework based on three-dimensional (3D) convolutional neural networks with a hybrid loss function. Methods: Two networks are incorporated in our method with the first being a liver shape autoencoder that is trained to obtain compressed codes of liver shapes, and the second being a liver segmentation network that is trained with a hybrid loss function. The design of the hybrid loss function is comprised of three parts. The first part is an adaptively weighted cross-entropy loss, which pays more attention to misclassified pixels. The second part is an edge-preserving smoothness loss, which guarantees that the adjacent pixels with the same label have similar outputs, while dissimilar for pixels with different labels. The third part of the loss is a shape constraint to model high-level structural differences based on the learned shape codes. Both networks use 3D operations for data processing. In our experiments, data augmentation is performed at both the training and the test stage. Results: We extensively evaluated our method on two datasets: the Segmentation of the Liver Competition 2007 (Sliver07), and the Combined (CT-MR) Healthy Abdominal Organ Segmentation (CHAOS) Challenge. Finally, with only 20 training scans, we achieved the best score of 82.55 on the Sliver07 challenge, and a score of 83.02 on the CHAOS challenge. Conclusions: In this study, we proposed a novel hybrid loss to overcome the difficulties in liver segmentation. The quantitative and qualitative results demonstrate that our method is highly suited for pathological liver segmentation, even when trained with a small dataset.

No abstract