RESUMEN

Objective.Multi-phase computed tomography (CT) has become a leading modality for identifying hepatic tumors. Nevertheless, the presence of misalignment in the images of different phases poses a challenge in accurately identifying and analyzing the patient's anatomy. Conventional registration methods typically concentrate on either intensity-based features or landmark-based features in isolation, so imposing limitations on the accuracy of the registration process.Method.We establish a nonrigid cycle-registration network that leverages semi-supervised learning techniques, wherein a point distance term based on Euclidean distance between registered landmark points is introduced into the loss function. Additionally, a cross-distillation strategy is proposed in network training to further improve registration performance which incorporates response-based knowledge concerning the distances between feature points.Results.We conducted experiments using multi-centered liver CT datasets to evaluate the performance of the proposed method. The results demonstrate that our method outperforms baseline methods in terms of target registration error. Additionally, Dice scores of the warped tumor masks were calculated. Our method consistently achieved the highest scores among all the comparing methods. Specifically, it achieved scores of 82.9% and 82.5% in the hepatocellular carcinoma and the intrahepatic cholangiocarcinoma dataset, respectively.Significance.The superior registration performance indicates its potential to serve as an important tool in hepatic tumor identification and analysis.

Asunto(s)

RESUMEN

Accurate liver tumor segmentation is a prerequisite for data-driven tumor analysis. Multiphase computed tomography (CT) with extensive liver tumor characteristics is typically used as the most crucial diagnostic basis. However, the large variations in contrast, texture, and tumor structure between CT phases limit the generalization capabilities of the associated segmentation algorithms. Inadequate feature integration across phases might also lead to a performance decrease. To address these issues, we present a domain-adversarial transformer (DA-Tran) network for segmenting liver tumors from multiphase CT images. A DA module is designed to generate domain-adapted feature maps from the non-contrast-enhanced (NC) phase, arterial (ART) phase, portal venous (PV) phase, and delay phase (DP) images. These domain-adapted feature maps are then combined with 3D transformer blocks to capture patch-structured similarity and global context attention. The experimental findings show that DA-Tran produces cutting-edge tumor segmentation outcomes, making it an ideal candidate for this co-segmentation challenge.

Asunto(s)

RESUMEN

Liver cancer has been one of the top causes of cancer-related death. For developing an accurate treatment strategy and raising the survival rate, the differentiation of liver cancers is essential. Multiphase CT recently acts as the primary examination method for clinical diagnosis. Deep learning techniques based on multiphase CT have been proposed to distinguish hepatic cancers. However, due to the recurrent mechanism, RNN-based approaches require expensive calculations whereas CNN-based models fail to explicitly establish temporal correlations among phases. In this paper, we proposed a phase difference network, termed as Phase Difference Network (PDN), to identify two liver cancer, hepatocellular carcinoma and intrahepatic cholangiocarcinoma, from four-phase CT. Specifically, the phase difference was used as interphase temporal information in a differential attention module, which enhanced the feature representation. Additionally, utilizing a multihead self-attention module, a transformer-based classification module was employed to explore the long-term context and capture the temporal relation between phases. Clinical datasets are used in experiments to compare the performance of the proposed strategy versus conventional approaches. The results indicate that the proposed method outperforms the traditional deep learning based methods.

Asunto(s)

RESUMEN

BACKGROUND: Accurate segmentation of pulmonary nodules is important for image-driven nodule analysis and nodule malignancy risk prediction. However, due to interobserver variability caused by manual segmentation, an accurate and robust automatic segmentation method has become an essential task. Therefore, the aim of the present study was to construct an accurate segmentation and malignant risk prediction algorithm for pulmonary nodules. METHODS: In the present study, we proposed a coarse-to-fine 2-stage framework consisting of the following 2 convolutional neural networks: a 3D multiscale U-Net used for localization and a 2.5D multiscale separable U-Net (MSU-Net) used for segmentation refinement. A multitask framework was proposed for nodules' malignancy risk prediction. Features from encoding and decoding paths of MSU-Net were integrated for pathology or morphology characteristic classification. RESULTS: Experimental results showed that our method achieved state-of-art results on the Lung Image Database Consortium and Image Database Resource Initiative dataset. The proposed method achieved a Dice similarity coefficient (DSC) of 83.04% and an overlapping error of 27.47% on the dataset. Our method achieved accuracy of 77.8% and area under the receiver-operating characteristic curve of 84.3% for malignancy risk prediction. Moreover, we compared our method with the inter-radiologist agreement, and the average DSC difference was only 0.39%. CONCLUSIONS: The results showed the effectiveness of the multitask end-to-end framework. The coarse-to-fine 2.5D strategy increased the accuracy and efficiency of pulmonary nodule segmentation and malignancy risk prediction of the computer-aided diagnosis system. In clinical practice, doctors can obtain accurate morphological characteristics and quantitative information of nodules by using the proposed method, so as to make future treatment plan.

RESUMEN

The early stage lung cancer often appears as ground-glass nodules (GGNs). The diagnosis of GGN as preinvasive lesion (PIL) or invasive adenocarcinoma (IA) is very important for further treatment planning. This paper proposes an automatic GGNs' invasiveness classification algorithm for the adenocarcinoma. 1431 clinical cases and a total of 1624 GGNs (3-30 mm) were collected from Shanghai Cancer Center for the study. The data is in high-resolution computed tomography (HRCT) format. Firstly, the automatic GGN detector which is composed by a 3D U-Net and a 3D multi-receptive field (multi-RF) network detects the location of GGNs. Then, a deep 3D convolutional neural network (3D-CNN) called Attention-v1 is used to identify the GGNs' invasiveness. The attention mechanism was introduced to the 3D-CNN. This paper conducted a contract experiment to compare the performance of Attention-v1, ResNet, and random forest algorithm. ResNet is one of the most advanced convolutional neural network structures. The competition performance metrics (CPM) of automatic GGN detector reached 0.896. The accuracy, sensitivity, specificity, and area under curve (AUC) value of Attention-v1 structure are 85.2%, 83.7%, 86.3%, and 92.6%. The algorithm proposed in this paper outperforms ResNet and random forest in sensitivity, accuracy, and AUC value. The deep 3D-CNN's classification result is better than traditional machine learning method. Attention mechanism improves 3D-CNN's performance compared with the residual block. The automatic GGN detector with the addition of Attention-v1 can be used to construct the GGN invasiveness classification algorithm to help the patients and doctors in treatment.

Asunto(s)